CN101587657A - Learning assistance system, program and learning assistance method - Google Patents

Abstract

The invention relate to a learning assistance system, program and learning assistance method, wherein the learning assistance system (200) comprises an answer text string management unit (222) that stores answer text string that input to an answer frame into an answer text string storage area; an answer text string read unit (242) that reads the answer text string from the answer text string storage area; a text string similarity calculating unit (248) that calculates the text string similarity between the answer text string and a plurality of correct text string against the answer frame; a correct text string determining unit (250) corresponding to answer, which selects more than one correct text string corresponding to the answer text string from the plurality of answer text strings according to similarity between the answer text string and the plurality of correct text strings, and determines it as correct text string with more than one corresponding answer; and a correctness judgment unit (254) that checks the answer text string with the corresponding correct text string to judge if the answer text string correct or not.

Description

Study assistance system, program and study assisted method

Technical field

The present invention relates to learn assistance system, relate in particular to identification as the text strings of answer input and the study assistance system that this answer is marked.

Background technology

In Japanese kokai publication hei 7-199794 communique, put down in writing image processing apparatus.In this image processing apparatus, read the image of answer by image fetching unit, and its image information is stored in the image storage unit with paper.According to this image information, by the problem that the identification of problem identification unit proposes, the identification of answer recognition unit is at literal, mark or the symbol of the answer of this problem.Utilize the scoring unit at each problem, relatively this answer that identifies and normal solution/partition information preserve the unit that preserve with the normal solution at problem that identify, judge whether each answer correct, and calculate mark.The answer that the appraisal result output unit utilizes the image output unit that its appraisal result is stored with image storage unit forms image with the paper image on recording chart, and makes its output.Thus, utilize digital copier can carry out the automatic scoring of answer simply, can also easily confirm this appraisal result with paper.

In TOHKEMY 2001-282090 communique, putting down in writing the study program.This study is carried out following processing with program: prompting problem literal in window, before the input of seeking the exercise problem answers, in order to find answer, according to the rules thinking processes is pointed out a plurality of enquirements successively in window, and these a plurality of enquirements are used to verify a plurality of understandings that must items of learner to should be appreciated that.When the answer at a certain enquirement is non-normal solution, behind the prompting normal solution, point out next enquirement.When the answer of puing question to is normal solution, give corresponding process branch.After having answered all enquirements, accept the answer of exercise problem.Divide and, decide learner's problem mark by process at the mark of the answer of exercise problem.Above-mentioned a plurality of enquirement comprises: with a plurality of must the item relevant enquirement shown in the essence in the problem literal and with the problem literal in do not illustrate a plurality of must the relevant enquirement of item.Thus, can help learning and be used to find to practise the thinking processes of problem answers.

In TOHKEMY 2004-29649 communique, put down in writing learning device.This learning device has Data Control portion, and this Data Control portion carries out the control of this learning device.In this Data Control portion, at the shown problem of display part, check normal solution by the problem of listening the speaker to store via the answer and the data store of input part input, mark thus.And, when being correct, on display part, show normal solution in answer at problem, at the answer mistake of problem the time, the explanation that the video data storage part is stored on display part.Thus, in the mode of learning that utilizes personal computer etc., provide and to have improved the understanding of learning content and the learning device of learning efficiently.

In TOHKEMY 2006-308925 communique, put down in writing the study assistance system.This study assistance system has problem display part and answer input part, and this problem display part shows the problem of reading from issue database on picture.In this study assistance system, have: the answer detection unit, it is checked by the answer of answer input part input and the normal solution of issue database, carries out answer and judges; Answer result of determination display part, its answer result of determination that will obtain by the answer detection unit with show accordingly as the answer of judging object; And answer result of determination storer, it is preserved answer result of determination and problem accordingly, the answer result of determination existing problem of problem display part during at former answer, the corresponding form of answer result of determination (for example distinguishing the color of answer frame) the demonstration problem with former answer the time.Thus, when showing same problem later on the 2nd time, show problem with the form that can know former correctness situation.

Summary of the invention

Feature of the present invention is that a kind of study assistance system has: storage unit, and it comprises answer text strings storage area and normal solution text strings storage area; Answer text strings administrative unit, it will store in the above-mentioned answer text strings storage area an answer text strings of an answer frame input; Answer text strings sensing element, it reads the answer text strings from above-mentioned answer text strings storage area; The text strings similarity is calculated the unit, its calculate in above-mentioned answer text strings of reading and the above-mentioned normal solution text strings storage area at the text strings similarity between a plurality of normal solution text strings of an above-mentioned answer frame; The corresponding normal solution text strings decision of answer unit, it is according to each text strings similarity between above-mentioned answer text strings and the above-mentioned a plurality of normal solution text strings, selects the more than one normal solution text strings corresponding with above-mentioned answer text strings and determine to be the corresponding normal solution text strings of more than one answer from above-mentioned a plurality of answer text strings; And the correctness identifying unit, it is checked above-mentioned answer text strings with above-mentioned corresponding normal solution text strings, carry out the judgement of the correctness of above-mentioned answer text strings.

The invention still further relates to the program and the method that are used to realize above-mentioned study assistance system.

Description of drawings

Fig. 1 illustrates signal conditioning package and the study assist server that can be used for embodiment of the present invention.

Fig. 2 illustrates the structure of the study assistance system of being realized by the signal conditioning package of Fig. 1 and/or study assist server.

Fig. 3 illustrate by user input and the answer text strings of reading by the answer text strings portion of reading, with this answer corresponding normal solution text strings, answer text strings and normal solution text strings between consistent literal number and the example of the similarity of calculating between answer text strings and the normal solution text strings.

Fig. 4 A and 4B are illustrated in the answer text strings when being non-normal solution, by answer text strings display part, correctness judge answer text strings in display part and the normal solution text strings display part four jiaos of answer frames that show, big, at result of determination in the answer frame of this answer text strings and the example that is positioned at the corresponding normal solution text strings of answer of answer frame upper side position.

Fig. 5 illustrates other structures of the study assistance system of being realized by the signal conditioning package of Fig. 1 and/or study assist server.

Fig. 6 A and 6B illustrate to be compared, corresponding normal solution text strings of answer and the answer text strings that offers wrong text point extraction unit from the corresponding normal solution text strings of the answer of the study assistance system of Fig. 5 determination section, the wrong text point in the answer text strings in the corresponding normal solution text strings of being extracted by wrong text point extraction unit of answer and the kind of wrong literal.

Fig. 7 is illustrated in the answer text strings when being non-normal solution, by answer text strings display part, correctness judge that display part and normal solution text strings display part show, at the example of the result of determination and the corresponding normal solution text strings of answer of the answer text strings in the answer frame.

Fig. 8 A illustrate by user input and other answer text strings of reading by the answer text strings portion of reading, with the example of its corresponding normal solution text strings, consistent literal number and the similarity calculated.Fig. 8 B and 8C illustrate the corresponding normal solution text strings of answer to be compared and answer text strings and the example of answer text strings with respect to the position of the consistent text point of the text point of the corresponding normal solution text strings of answer, inconsistent, scarce word.Fig. 8 D illustrates the wrong text point in the corresponding normal solution text strings of answer and the example of error type thereof.

Fig. 9 is illustrated in the answer text strings when being non-normal solution, by answer text strings display part, correctness judge that display part and normal solution text strings display part show, at the example of the result of determination and the corresponding normal solution text strings of answer of the answer text strings in the answer frame.

Figure 10 A illustrate by user input and another answer text strings of reading by the answer text strings portion of reading, with the example of its corresponding normal solution text strings, consistent literal number and the similarity calculated.Figure 10 B and 10C illustrate the corresponding normal solution text strings of answer to be compared and answer text strings and answer text strings with respect to the consistent text point of the text point of the corresponding normal solution text strings of answer, the position of inconsistent, scarce word.Figure 10 D illustrates the wrong text point in the corresponding normal solution text strings of answer and the example of error type.

Figure 11 is illustrated in the answer text strings when being non-normal solution, by answer text strings display part, correctness judge answer text strings in that display part and normal solution text strings display part show, the answer frame, at the example of the result of determination and the corresponding normal solution text strings of answer of the answer text strings in the answer frame.

Figure 12 illustrates another structure of the study assistance system of being realized by the signal conditioning package of Fig. 1 and/or study assist server.

Figure 13 A and 13B illustrate from the corresponding normal solution text strings of the answer of the study assistance system of Figure 12 determination section offer the corresponding normal solution text strings of answer to be compared of wrong text point extraction unit and answer text strings, the wrong text point the corresponding normal solution text strings of answer extracted by wrong text point extraction unit, the wrong text point in the answer text strings, the error type of mistake literal.

Figure 14 is illustrated in the answer text strings when being non-normal solution, by answer text strings display part, correctness judge answer text strings in that display part and normal solution text strings display part show, the answer frame, at the example of the result of determination and the corresponding normal solution text strings of answer of the answer text strings in the answer frame.

Figure 15 A～15C illustrate from the corresponding normal solution text strings of the answer of the study assistance system of Figure 12 determination section offer the corresponding normal solution text strings of answer to be compared of wrong text point extraction unit and answer text strings, the example of the error type of wrong text point the corresponding normal solution text strings of answer extracted by wrong text point extraction unit and the wrong text point in the answer text strings.

Figure 16 is illustrated in the answer text strings when being non-normal solution, by answer text strings display part, correctness judge answer text strings in that display part and normal solution text strings display part show, the answer frame, at the example of the result of determination and the corresponding normal solution text strings of answer of the answer text strings in the answer frame.

Figure 17 A～17C illustrate from the corresponding normal solution text strings of the answer of the study assistance system of Figure 12 determination section offer the corresponding normal solution text strings of answer to be compared of wrong text point extraction unit and answer text strings, the example of the error type of wrong text point the answer text strings extracted by wrong text point extraction unit and the wrong literal in the answer text strings.

Figure 18 is illustrated in the answer text strings when being non-normal solution, by answer text strings display part, correctness judge answer text strings in that display part and normal solution text strings display part show, the answer frame, at the example of the result of determination and the corresponding normal solution text strings of answer of the answer text strings in the answer frame.

Figure 19 A illustrates the example by user input and three answer text strings of being read by the answer text strings portion of reading, corresponding three normal solution text strings and the similarity calculated.Figure 19 B illustrates according to the similarity between the corresponding normal solution text strings of answer text strings and answer to come each answer text strings is distributed or the example of the corresponding normal solution text strings of corresponding answer.

When Figure 20 is illustrated in the answer text strings and comprises normal solution and non-normal solution, by answer text strings display part, correctness judge answer text strings in that display part and normal solution text strings display part show, the answer frame, at the example of the result of determination and the corresponding normal solution text strings of answer of the answer text strings in the answer frame.

Figure 21 A illustrates the example of two answer text strings, possible four normal solution text strings of reading by user input and by the answer text strings portion of reading and the similarity of calculating.Figure 21 B illustrates and checks on one's answers according to the similarity between the corresponding normal solution text strings of answer text strings and answer that text strings is distributed or the example of the corresponding normal solution text strings of corresponding answer.

When Figure 22 is illustrated in the answer text strings and comprises normal solution and non-normal solution, by answer text strings display part, correctness judge answer text strings in that display part and normal solution text strings display part show, the answer frame, at the example of the result of determination and the corresponding normal solution text strings of answer of the answer text strings in the answer frame.

Figure 23 A illustrates three shown answer frame identiflication numbers of display device 112 and the display position (coordinate) of the corresponding normal solution text strings of answer text strings that obtained by normal solution text strings display position obtaining section.Figure 23 B illustrates the corresponding normal solution text strings of answer of utilizing the corresponding text strings determination section 250 of answer to distribute according to the text strings similarity of Figure 19 A.

When Figure 24 is illustrated in the answer text strings and comprises normal solution and non-normal solution, by answer text strings display part, correctness judge that display part and normal solution text strings display part show, at the example of the result of determination and the corresponding normal solution text strings of answer of the answer text strings in the answer frame.

Figure 25 illustrates the structure of another study assistance system of being realized by the signal conditioning package of Fig. 1 and/or study assist server.

Figure 26 A illustrates the example of the person's handwriting of the hand-written answer text strings that is taken into by user input and by the handwriting information portion of reading.Figure 26 B illustrates the example that is intercepted the literal intercepting candidate of regional candidate's generating unit generation at the person's handwriting coordinate data of Figure 26 A by literal.Figure 26 C illustrate obtain by indivedual character recognition portion, at the literal shown in Figure 26 B intercept regional candidate the identification candidate character and identification mark an example.

Figure 27 A～27D illustrates literal verify and the text strings verify of obtaining the identification candidate character of Figure 26 C by maximum likelihood text strings determination section, and decision is at the example of the flow sequence of the literal intercepting candidate's of Figure 26 B of one of normal solution text strings maximum likelihood text strings.

Figure 28 A～28D illustrates literal verify and the text strings verify of obtaining the identification candidate character of Figure 26 C by maximum likelihood text strings determination section, and decision is at another literal intercepting candidate's the example of flow sequence of maximum likelihood text strings of Figure 26 B of normal solution text strings.

Figure 29 is used to illustrate by the corresponding normal solution text strings of answer determination section and decides method at the corresponding normal solution text strings of answer of maximum likelihood text strings.

The example of the correctness result of determination of the answer text strings when Figure 30 illustrates the non-normal solution that the study assistance system by Figure 25 carries out.

Figure 31 A and 31B illustrate the process flow diagram of being judged the processing of normal solution/non-normal solution by the study assistance system of Fig. 2 being used to of carrying out.

Figure 32 A～32C is illustrated in the step of Figure 31 B the process flow diagram of processing that is extracted the wrong text point of the corresponding normal solution text strings of answer by the study assistance system of Fig. 5 being used for of carrying out.

Figure 33 A～33C is illustrated in the step 546 of Figure 31 B the process flow diagram of processing that is extracted the wrong text point of answer text strings by the study assistance system of Figure 12 being used for of carrying out.

Figure 34 A～34D illustrates by what the study assistance system of Figure 23 was carried out and is used to judge normal solution/non-normal solution, and decision is at the process flow diagram of the processing of the distribution of the corresponding normal solution text strings of answer of answer frame.

Figure 35 illustrates the part of other process flow diagrams of being judged the processing of normal solution/non-normal solution by the study assistance system of Figure 25 being used to of carrying out.

Figure 36 A～36E illustrates other process flow diagrams of the details of step among the Figure 35 that is carried out by maximum likelihood text strings determination section.

Embodiment

In the system of TOHKEMY 2004-29649 communique, judge the correctness of the answer of having imported, also show the normal solution of storing in pairs with problem.But, the inventor thinks, when existence can be suitable for a plurality of possible normal solution text strings group (group) of an answer frame, check answer text strings and the normal solution text strings imported, judge its correctness, when just showing and problem in pairs during a plurality of normal solution text strings of storage,, be difficult to grasp the corresponding normal solution text strings of answer with learner's input for the learner with enumerating.

The inventor thinks, when the different situation of a part that shows the answer text strings, shows that the part corresponding with wrong literal in the normal solution text strings that has shown is effective.The inventor thinks, can show the errors present in this normal solution text strings by determining from a plurality of normal solution text strings and the corresponding normal solution text strings of answer of learner's input.

The inventor thinks, at the problem that a plurality of possible answers are arranged in an answer frame, show that the answer text strings and the normal solution text strings group that check input carry out the result that correctness is judged, at this moment, except showing the correctness result of determination, also be shown as the corresponding normal solution text strings of answer text strings that to grasp and import, in addition, when showing the part corresponding with wrong literal in the normal solution text strings, can obtain results of learning preferably, wherein, described normal solution text strings is corresponding with the answer text strings.

The objective of the invention is and in the study assistance system, to select and to show and the corresponding normal solution of importing of answer.

Other purposes of the present invention are the wrong counterparts that can show in the study assistance system in the normal solution corresponding with the answer of input.

According to the present invention, can in the study assistance system, select and show the corresponding normal solution of answer with input, and can show the wrong counterpart in the normal solution corresponding with the answer of importing.

With reference to accompanying drawing embodiments of the present invention are described.In the accompanying drawings, identical reference is numbered to same textural element mark.

Fig. 1 illustrates signal conditioning package 10 and the study assist server 20 that can be used for embodiment of the present invention.Signal conditioning package 10 possesses: the tablet 114 of processor (CPU) 12, the memory storage 14 that contains various databases, display device (display) 112, writable input, the input media 116 that comprises keyboard (KY) and positioning equipment and network interface (N/W IF) 18.Signal conditioning package 10 has the study assist in functions of being carried out by processor 12 16.

Signal conditioning package 10 also can be individually promptly provides the study assistance services as the device of self to the user.Processor 12 also can move according to 14 program stored of memory storage, and this program is used for realizing that study assist in functions 16 etc. is all multi-functional.This all multi-functional at least a portion also can be installed on the processor 12 with the example, in hardware as integrated circuit.Signal conditioning package 10 can be tablet PC.

Structure instead, signal conditioning package 10 can be connected with the study assist server 20 with study assist in functions 26 via network 5.At this moment, signal conditioning package 10 is cooperated with study assist server 20 as terminal performance function, provides the study assistance services to the user.Study assist server 20 has processor (CPU) 22, comprises the memory storage 24 and the network interface (N/W IF) 28 of database.

Study assist server 20 has the study assist in functions of being carried out by processor 22 26.Processor 22 can move according to the program that is stored in the memory storage 24, and this program is used for realizing that study assist in functions 26 etc. is all multi-functional.This all multi-functional at least a portion also can be installed on the processor 22 with the example, in hardware as integrated circuit.Study assist in functions 16 as the terminal of signal conditioning package 10 is installed on the processor 12 with the form of hardware or software as mentioned above.

Fig. 2 illustrates the structure of the study assistance system of realizing by the signal conditioning package 10 of Fig. 1 and/or study assist server 20 200.

Study assistance system 200 comprise be arranged in the signal conditioning package 10 as the input handling part, answer input part 210 and correctness judge instruction unit 212, comprise be arranged in the signal conditioning package 10 as display process portion, answer text strings display part 214, normal solution text strings display part 216 and correctness judge display part 218.

Study assistance system 200 also comprise be arranged in the signal conditioning package 10 as information treatment part, answer text strings management department 222, answer text strings read portion 242, the text strings similarity is calculated portion 248, the corresponding normal solution text strings determination section 250 of answer and correctness detection unit 254.Study assistance system 200 also comprises answer text strings storage part 252 and normal solution text strings group storage part 258, as the zone in the memory storage 14.

Structure instead, when with signal conditioning package 10 as terminal, when mainly providing the study assistance services by study assist server 20, study assistance system 200 comprise be arranged in the study assist server 20 as information treatment part, answer text strings management department 222, answer text strings read portion 242, the text strings similarity is calculated portion 248, the corresponding text strings determination section 250 of answer and correctness detection unit 254.Study assistance system 200 also comprise be arranged in the study assist server 20 as the zone in the memory storage 24, answer text strings storage part 252 and normal solution text strings group storage part 258.

Among Fig. 1, when the user uses the key of input media 116 or mouse etc. to be disposed at cursor in answer frame or the answer hurdle on display device 112 and uses input media 116 in this answer frame successively during input (word processing input) answer text strings, answer input part 210 is taken into the data of answer text strings successively and offers answer text strings management department 222.The answer text strings management department 222 text strings storage part 252 that checks on one's answers is stored these answer text strings data successively and is offered answer text strings display part 214 successively.Answer text strings display part 214 is shown to the answer text strings of received data in the answer frame on the display device 112.

Fig. 3 illustrates by user input and by the answer text strings and reads the answer text strings (" Dry state ") that portion 242 reads, the normal solution text strings (“ South-Korea state corresponding with this answer " and “ Da South-Korea the Republic of China "), the consistent literal number between answer text strings and the normal solution text strings and the example of the similarity of calculating between answer text strings and the normal solution text strings.

When using positioning equipment or key to wait, the user operates the key (soft key) of " correctness is judged indication " usefulness shown on (click or press) display device 112, when operating correctness judgement instruction unit 212 thus, the answer text strings is read portion 242 and is read the answer text strings data of being stored in the answer text strings storage part 252, and offers the text strings similarity and calculate unit 248.

The text strings similarity is calculated portion 248 and is calculated the answer text strings " Dry state " that received and a plurality of possible normal solution text strings " South-Korea state " corresponding with this answer in the normal solution text strings group storage part 258 and " big South-Korea the Republic of China " the text strings similarity between separately.The text strings similarity for example can be represented by following formula.

The literal number of text strings similarity=unanimity/benchmark literal number

Here, any one big number in benchmark literal number=" the literal number of normal solution text strings " and " the literal number of answer text strings ".

At this moment, about normal solution text strings “ South-Korea state " and “ Da South-Korea the Republic of China ", consistent literal number is 1, and benchmark literal number is 2 and 4, and the text strings similarity is respectively " 0.5 " and " 0.25 ".

The corresponding text strings determination section 250 of answer is according to the text strings similarity of having calculated (" 0.5 ", " 0.25 "), from a plurality of possible normal solution text strings, select normal solution text strings “ South-Korea state with maximum text strings similarity ", and with its decision for the normal solution text strings corresponding with the answer text strings, be the corresponding normal solution text strings of answer.Exist under a plurality of situations in text strings, these a plurality of normal solution text strings decisions are the corresponding normal solution text strings of answer with maximum text strings similarity.

Correctness detection unit 254 is checked corresponding normal solution text strings of determined answer and answer text strings, if both unanimities then are judged to be normal solution, if both are inconsistent, then is judged to be non-normal solution.This result of determination is offered correctness judge display part 218.Correctness judge display part 218 when result of determination is normal solution for example in the answer frame with red display zero, when result of determination is non-normal solution for example in the answer frame with red display *.

The corresponding normal solution text strings of answer determination section 250 offers normal solution text strings display part 216 with the corresponding normal solution text strings of determined answer.Normal solution text strings display part 216 shows corresponding normal solution text strings on the display position, for example answer frame of the normal solution text strings of regulation.The corresponding normal solution text strings of answer that normal solution text strings display part 216 receives from the corresponding normal solution text strings of answer determination section 250 with different form demonstrations is so that this normal solution text strings can be distinguished with the normal solution text strings beyond it.

Fig. 4 A and 4B are illustrated in the answer text strings when being non-normal solution, by answer text strings display part 214, correctness judge that display part 218 and normal solution text strings display part are 216 that show, answer text strings (" Dry state ") in the big cubic answer frame, at result of determination in the answer frame of this answer text strings (for example " * ") and the corresponding normal solution text strings of the answer (“ South-Korea state that is in the upper side position of answer frame ") example.Among the figure, thick dashed line " zero " reaches normal solution (zero) or the non-normal solution (*) that " * " symbolic representation is used to judge correctness, two text strings (“ South-Korea states of answer frame upside " “ Da South-Korea the Republic of China ") be possible normal solution text strings, surround a normal solution text strings “ South-Korea state " cubic solid line represent the corresponding normal solution text strings of determined answer.

Among Fig. 4 A, show whole possible normal solution text strings, surround the corresponding normal solution text strings of answer with cubic frame at answer frame upside.Among Fig. 4 B, only show the corresponding normal solution text strings of answer at answer frame upside.

Fig. 5 illustrates other structures of the study assistance system of being realized by the signal conditioning package 10 of Fig. 1 and/or study assist server 20 200.

Among Fig. 5, study assistance system 200 on the basis of the structure of Fig. 2, also comprise be arranged at signal conditioning package 10 or study assist server 20 in as information treatment part, mistake text point extraction unit 256.Other structures of study assistance system 200 are identical with the structure of Fig. 2.

Mistake text point extraction unit 256 receives corresponding normal solution text strings of answer and answer text strings from the corresponding normal solution text strings of answer determination section 250, relatively check corresponding normal solution text strings of this answer and answer text strings, extract the wrong text point (with error section corresponding character position) in the corresponding normal solution text strings of determined answer " Dry state ".The wrong text point that mistake text point extraction unit 256 will be extracted offers normal solution text strings display part 218.Normal solution text strings display part 218 is with the wrong text point in the different form demonstration normal solution text strings, so that should can distinguish with other parts by the mistake text point.

Fig. 6 A and 6B illustrate to be compared, corresponding normal solution text strings of answer and the answer text strings that offers wrong text point extraction unit 256 from the corresponding normal solution text strings of the answer of the study assistance system 200 of Fig. 5 determination section 250, the corresponding normal solution text strings of the answer (“ South-Korea state that extracts by wrong text point extraction unit 256 ") in the answer text strings in wrong text point (for example, the 1st literal) and the kind of wrong literal.

For example, mistake text point extraction unit 256 check on one's answers the 1st literal " Dry " of text strings " Dry state " and the corresponding normal solution text strings of answer “ South-Korea state " the 1st literal “ South-Korea " compare, detect inconsistent, the corresponding normal solution text strings of the 2nd literal " state " of the text strings that then checks on one's answers " Dry state " and answer “ South-Korea state " the 2nd literal " state " compare, detect unanimity.Thus, mistake text point extraction unit 256 is judged to be the 1st literal with wrong text point.Mistake text point extraction unit 256 is extracted the 1st literal of the corresponding normal solution text strings of answer as wrong text point, the error type of the 1st literal " Dry " is judged to be " different word ".

Fig. 7 is illustrated in the answer text strings when being non-normal solution, by answer text strings display part 214, correctness judge display part 218 and normal solution text strings display part 216 that show, at the result of determination (for example, non-normal solution " * ") and the corresponding normal solution text strings of the answer (“ South-Korea state of the answer text strings in the answer frame (" Dry state ") ") example.At this moment, adopt the color different to emphasize demonstration with wrong literal or different word corresponding character in the corresponding normal solution text strings of answer with other literal.Among the figure, the part that surrounds with cubic dotted line represent by for example with other literal (for example, black) different color (for example, redness) emphasizes to show the part (not needing to show dotted line itself) of the literal of the wrong text point in the corresponding normal solution text strings of answer.

Fig. 8 A illustrate by user input and by the answer text strings read other answer text strings (" middle richness ") that portion 242 reads, with the example of its corresponding normal solution text strings (" Zhong Chen Sickle foot " reaches " Teng Yuan Sickle enough "), consistent literal number and the similarity calculated.At this moment, " the Zhong Chen Sickle foot " with highest similarity (0.25) extracts as the corresponding normal solution text strings of answer.

Fig. 8 B and 8C illustrate the corresponding normal solution text strings of answer to be compared and answer text strings and answer text strings with respect to the consistent text point of the text point of the corresponding normal solution text strings of answer (" Zhong Chen Sickle foot "), the example of inconsistent, scarce word location.Fig. 8 D illustrates the wrong text point (the 2nd literal～the 4th literal) in the corresponding normal solution text strings of answer (" Zhong Chen Sickle foot ") and the example of error type thereof.

When reference Fig. 8 C, mistake text point extraction unit 256 check on one's answers corresponding normal solution text strings " Zhong Chen Sickle foot " the 1st literal " in " and the 1st literal of answer text strings " middle richness " " in " compare, detect unanimity, obtain representing the value A[m of answer text strings with respect to the position of the consistent literal of the m literal of the corresponding normal solution text strings of answer]=1, as the checked result of the literal of the text point m=1 of the 1st literal of the corresponding normal solution text strings of expression answer.

Then, check on one's answers the 2nd literal " minister " of corresponding normal solution text strings " Zhong Chen Sickle foot " and the 2nd literal " richness " of answer text strings " middle richness " of mistake text point extraction unit 256 compares, detect inconsistent, obtain representing the value A[m of answer text strings with respect to the literal of the m literal of the corresponding normal solution text strings of answer inconsistent (or not having consistent text point)]=0, as the checked result of the literal of the text point m=2 of the 2nd literal of the corresponding normal solution text strings of expression answer.

Then, the 3rd literal " Sickle " of 256 pairs of normal solution text strings of mistake text point extraction unit " middle minister Sickle foot " and each literal of answer text strings " middle richness " compare; detect the literal disappearance in the corresponding normal solution text strings of answer; obtain representing the value A[m of answer text strings literal with respect to the scarce word (or not having text point) of the m literal of the corresponding normal solution text strings of answer]=-1, as the checked result of the literal of the text point m=3 of the 3rd literal of representing answer correspondence normal solution text strings.Then, check on one's answers the 4th literal " foot " of corresponding normal solution text strings " Zhong Chen Sickle foot " and each literal of answer text strings " middle richness " of mistake text point extraction unit 256 compares, detect the literal disappearance (or not having text point) in the answer text strings, obtain representing the value A[m of answer text strings with respect to the scarce word of the m literal of the corresponding normal solution text strings of answer]=-1, as the checked result of the text point m=4 of the 4th literal of the corresponding normal solution text strings of expression answer.

In addition, shown in Fig. 8 D, mistake text point extraction unit 256 is also carried out following processing: because the 2nd literal is inconsistent, so the 2nd literal of the corresponding normal solution text strings of answer is extracted as wrong text point, and the error type of the 2nd literal " minister " is judged to be different words, because the 3rd and the 4th literal is for lacking word, thus the 3rd and the 4th literal of the corresponding normal solution text strings of answer is extracted as wrong text point, and with the 3rd and the 4th literal “ Sickle enough " error type be judged to be scarce word.

Fig. 9 is illustrated in the answer text strings when being non-normal solution, by answer text strings display part 214, correctness judge display part 218 and normal solution text strings display part 216 that show, at the example of the result of determination (for example non-normal solution " * ") and the corresponding normal solution text strings of answer (" Zhong Chen Sickle enough ") of the answer text strings in the answer frame (" middle richness ").Among the figure, the part that surrounds with cubic dotted line represent by for example with other literal (for example, black) different colors (for example, red, blueness) emphasize to show the wrong text point corresponding character (not needing to show dotted line itself) in the corresponding normal solution text strings with answer.At this moment, " minister " is by red display ， “ Sickle foot " show by blueness.

Figure 10 A illustrate by user input and by the answer text strings read another answer text strings (" middle Chen Shang Sickle foot becomes ") that portion 242 reads, with the example of its corresponding normal solution text strings (" Zhong Chen Sickle foot " reaches " Teng Yuan Sickle is sufficient "), consistent literal number and the similarity calculated.At this moment, " the Zhong Chen Sickle foot " with highest similarity (0.4) extracts as the corresponding normal solution text strings of answer.

Figure 10 B and 10C illustrate the corresponding normal solution text strings of answer to be compared and answer text strings and consistent text point, inconsistent, the scarce word location of answer text strings with respect to the text point of the corresponding normal solution text strings of answer (" Zhong Chen Sickle foot ").Figure 10 D illustrates the wrong text point (the 2nd literal, the 4th literal) in the corresponding normal solution text strings of answer (" Zhong Chen Sickle foot ") and the example of error type.

When reference Figure 10 C, mistake text point extraction unit 256 check on one's answers corresponding normal solution text strings " Zhong Chen Sickle foot " the 1st literal " in " and the 1st literal of answer text strings " middle Chen Shang Sickle becomes enough " " in " compare, detect unanimity, obtain representing the value A[m of the consistent text point of the answer text strings consistent with the m literal of the corresponding normal solution text strings of answer]=1, as the checked result of the text point m=1 of the 1st literal of the corresponding normal solution text strings of expression answer.Then, check on one's answers the 2nd literal " minister " of corresponding normal solution text strings " Zhong Chen Sickle foot " and the 2nd literal " minister " of answer text strings " middle Chen Shang Sickle becomes enough " of mistake text point extraction unit 256 compares, detect unanimity, obtain representing the value A[m of the consistent text point of the answer text strings consistent with the m literal of the corresponding normal solution text strings of answer]=2, as the checked result of the text point m=2 of the 2nd literal of the corresponding normal solution text strings of expression answer.

Then, check on one's answers the 3rd literal “ Sickle of corresponding normal solution text strings " Zhong Chen Sickle foot " of mistake text point extraction unit 256 " and each literal (the 3rd literal; the 4th literal) of answer text strings " middle Chen Shang Sickle foot becomes " compare; detect unanimity; obtain representing the value A[m of the consistent text point of the answer text strings consistent with the m literal of the corresponding normal solution text strings of answer]=4, as the checked result of the text point m=3 that represents the 3rd literal.Then, check on one's answers the 4th literal " foot " of corresponding normal solution text strings " Zhong Chen Sickle foot " and each literal (the 5th literal) of answer text strings " middle Chen Shang Sickle becomes enough " of mistake text point extraction unit 256 compares, detect unanimity, obtain representing the value A[m of the consistent text point of the answer text strings consistent with the m literal of the corresponding normal solution text strings of answer]=5, as the checked result of the text point m=4 of expression the 4th literal.

Shown in Figure 10 D, mistake text point extraction unit 256 is also carried out following processing: the 3rd literal of " middle Chen Shang Sickle foot becomes " because the answer text strings " on " and the 6th literal " one-tenth " be unnecessary word, so with the text point before the unnecessary word in the corresponding normal solution text strings of each answer, promptly the 2nd literal and the 4th literal extract as wrong text point, and be judged to be " the unnecessary word " that is added respectively as the 2nd literal " minister " of answer text strings and the error type of the 4th literal " foot ".

Figure 11 is illustrated in the answer text strings when being non-normal solution, by answer text strings display part 214, correctness judge that display part 218 and normal solution text strings display part are 216 that show, answer text strings (" middle Chen Shang Sickle foot becomes ") in the answer frame, at the example of the result of determination (for example non-normal solution " * ") and the corresponding normal solution text strings of answer (" Zhong Chen Sickle foot ") of the answer text strings in the answer frame.Among the figure, the underscore in the normal solution text strings is represented the corresponding position of unnecessary text point of the answer text strings in the corresponding normal solution text strings with answer.This underscore can be emphasized to show with not homochromy (for example, redness).

Figure 12 illustrates the another structure of the study assistance system of being realized by the signal conditioning package 10 of Fig. 1 and/or study assist server 20 200.

In Figure 12, except the structure of Fig. 5, study assistance system 200 also comprise be arranged at signal conditioning package 10 or study assist server 20 in as information treatment part, answer text strings changing unit 259.Study assistance system 200 also comprises and is arranged in signal conditioning package 10 or the study assist server 20 as display position information storage part 262 information treatment part, in normal solution text strings display position obtaining section 260 and the memory storage 14.Other structures of study assistance system 200 are identical with the structure of Fig. 5.

Mistake text point extraction unit 256 receives corresponding normal solution text strings of answer and answer text strings from the corresponding normal solution text strings of answer determination section 250, relatively check corresponding normal solution text strings of this answer and answer text strings, extract the corresponding normal solution text strings of determined answer (“ South-Korea state ") in wrong text point (the 1st literal), also extract the wrong text point (the 1st literal) in the answer text strings (" Dry state ").Mistake text point extraction unit 256 should the mistake text point information offer answer text strings changing unit 259.Answer text strings changing unit 259 is according to the information of the wrong text point (the 1st literal) that receives from wrong text point extraction unit 256, change the demonstration of the wrong literal in the answer text strings shown in the answer frame, and use the color different to emphasize to show with other literal.

Figure 13 A and 13B illustrate to be compared, corresponding normal solution text strings of answer and the answer text strings of supplying with to wrong text point extraction unit 256 from the corresponding normal solution text strings of the answer of the study assistance system 200 of Figure 12 determination section 250, the wrong text point (the 1st literal) in the corresponding normal solution text strings of being extracted by wrong text point extraction unit 256 of answer (" South-Korea state "), the wrong text point (the 1st literal) in the answer text strings (" Dry state ") and the error type of wrong literal.

When reference Figure 13 B, mistake text point extraction unit 256 check on one's answers the 1st literal " Dry " of text strings " Dry state " and the corresponding normal solution text strings of answer “ South-Korea state " the 1st literal “ South-Korea " compare, detect inconsistent.Mistake text point extraction unit 256 is also carried out following processing: because the 1st literal of the corresponding normal solution text strings of answer is inconsistent, so each the 1st literal of corresponding normal solution text strings of answer and answer text strings is extracted as wrong text point, is judged to be " different word " as the error type of each the 1st literal " Dry ".

Figure 14 is illustrated in the answer text strings when being non-normal solution, by answer text strings display part 214, correctness judge that display part 218 and normal solution text strings display part are 216 that show, answer text strings (" Dry state ") in the answer frame, at the result of determination (for example " * ") and the corresponding normal solution text strings of the answer (“ South-Korea state of the answer text strings in the answer frame ") example.At this moment, with different color (for example, redness) emphasize to show with the corresponding normal solution text strings of answer in wrong literal corresponding character and wrong literal or the different word in the answer text strings.

Figure 15 A～15C illustrates to be compared, corresponding normal solution text strings of answer and the answer text strings that offers wrong text point extraction unit 256 from the corresponding normal solution text strings of the answer of the study assistance system 200 of Figure 12 determination section 250, the example of the wrong text point (the 2nd literal～the 4th literal) in the corresponding normal solution text strings of being extracted by wrong text point extraction unit 256 of answer (" Zhong Chen Sickle foot ") and the error type of the wrong text point (the 2nd literal～the 4th literal) in the answer text strings (" middle richness ").

Among Figure 15 B, n represents the text point in the answer text strings, A[n] the consistent text point of the expression answer corresponding normal solution text strings consistent with the n literal of answer text strings.Mistake text point extraction unit 256 check on one's answers text strings " middle richness " the 1st literal " in " and the 1st literal of normal solution text strings " Zhong Chen Sickle foot " " in " compare, detect unanimity, obtain representing the value A[n of the consistent text point of the answer corresponding normal solution text strings consistent with the n literal of answer text strings]=1, as the checked result of the text point n=1 of the 1st literal of expression answer text strings.

Then, mistake text point extraction unit 256 check on one's answers the 2nd literal " minister ", the 3rd literal “ Sickle of the 2nd literal " richness " of text strings " middle richness " and normal solution text strings " Zhong Chen Sickle foot " " and the 4th literal " foot " compare; detect inconsistent; obtain representing the value A[n of the literal of the corresponding normal solution text strings of answer with respect to the n literal of answer text strings inconsistent (or not having consistent text point)]=0, as the checked result of the text point n=2 of the 2nd literal of representing the answer text strings.

Shown in Figure 15 C, mistake text point extraction unit 256 is also carried out following processing: because the 3rd literal “ Sickle of the corresponding text strings of answer " Zhong Chen Sickle foot " " and the 4th literal " foot " for lacking word; so with the text point before the scarce word of institute in the answer text strings promptly the 2nd literal extract as wrong text point, and add the judgement of " scarce word " as the error type of the 2nd literal " richness " of answer text strings.

Figure 16 is illustrated in the answer text strings when being non-normal solution, by answer text strings display part 214, correctness judge that display part 218 and normal solution text strings display part are 216 that show, answer text strings (" middle richness ") in the answer frame, at the example of the result of determination (for example non-normal solution " * ") and the corresponding normal solution text strings of answer (" Zhong Chen Sickle foot ") of the answer text strings in the answer frame.Among the figure, emphasize to show the erroneous words (not needing to show dotted line itself) that surrounds by the dotted line in the answer frame by for example different color (for example, redness, blueness) with other literal (for example, black).Underscore in the answer frame is represented the scarce word location in the answer text strings.

Figure 17 A～17C illustrates to be compared, corresponding normal solution text strings of answer and the answer text strings that offers wrong text point extraction unit 256 from the corresponding normal solution text strings of the answer of the study assistance system 200 of Figure 12 determination section 250, the example of the wrong text point (the 3rd literal, the 6th literal) in the answer text strings of extracting by wrong text point extraction unit 256 (" middle Chen Shang Sickle foot becomes ") and the error type of the wrong literal in the answer text strings.

Check on one's answers respectively the 1st literal of text strings " middle Chen Shang Sickle foot become " and the 2nd literal " middle minister " of mistake text point extraction unit 256 compares separately with the 1st literal and the 2nd literal " middle minister " of the corresponding normal solution text strings of answer " Zhong Chen Sickle foot " separately, detect unanimity, obtain representing the consistent text point A[n of the answer corresponding normal solution text strings consistent with the n literal of answer text strings]=1, A[n]=2, as the 1st literal and the text point n=1 of the 2nd literal and 2 the checked result of expression answer text strings.Then, mistake text point extraction unit 256 check on one's answers text strings " middle Chen Shang Sickle foot become " the 3rd literal " on " and the 3rd literal and each literal of the corresponding normal solution text strings of answer " Zhong Chen Sickle enough " compare, detect inconsistent, obtain representing value A[n with respect to the literal inconsistent (or not having consistent text point) of the n literal of answer text strings]=0, as the checked result of the text point n=3 of the 3rd literal in the expression answer text strings.

Then, mistake text point extraction unit 256 check on one's answers respectively the 4th literal and the 5th literal “ Sickle foot of text strings " middle Chen Shang Sickle foot become " " separately with the 3rd literal and the 4th literal “ Sickle foot of the corresponding normal solution text strings of answer " Zhong Chen Sickle foot " " compare separately, detect unanimity, obtain representing the value A[n of the consistent text point of the answer corresponding normal solution text strings consistent with the n literal of answer text strings]=3, A[n]=4, as the 4th literal and the text point n=4 of the 5th literal and 5 the checked result of expression answer text strings.Then, because the text point of the corresponding normal solution text strings of answer " Zhong Chen Sickle foot " that compares with the 6th literal " one-tenths " of answer text strings " middle Chen Shang Sickle foot becomes " not, do not have the value A[n of text point so wrong text point extraction unit 256 obtains representing the n literal with respect to the answer text strings]=-1, as the checked result of the text point n=6 of the 6th literal in the expression answer text strings.Like this, mistake text point extraction unit 256 with the 3rd literal of answer text strings " middle Chen Shang Sickle foot become " and the 6th literal " on " reach " one-tenth " and extract as unnecessary literal.

Figure 18 is illustrated in the answer text strings when being non-normal solution, by answer text strings display part 214, correctness judge answer text strings (" middle Chen Shang Sickle foot becomes ") in the answer frame that display part 218 and normal solution text strings display part 216 show, at the example of the result of determination (for example non-normal solution " * ") and the corresponding normal solution text strings of answer (" Zhong Chen Sickle foot ") of the answer text strings in the answer frame.Emphasize to show unnecessary word (not needing to show dotted line itself) in the answer frame that surrounds by dotted line by for example color (for example, blueness) different among the figure with other literal (for example, black).

The decision and the display packing of the corresponding normal solution text strings of answer when having the above a plurality of possible normal solution text strings of h at h answer frame then, are described.

Figure 19 A illustrates by user input and by the answer text strings and reads three answer text strings (" people master master Righteousness ", " basic people Goodwill respects " reach " state people master Righteousness ") that portion 242 reads, corresponding three normal solution text strings (" basic people Goodwill respect ", " state democracy Goodwill " reach " War strives and puts Abandoned ") and the example of the similarity calculated.Figure 19 B illustrates distribution or the corresponding example that carries out the corresponding normal solution text strings of answer based on the similarity between answer text strings and the corresponding normal solution text strings of answer at each answer text strings.

The text strings similarity is calculated the whole combinations of portion 248 at answer text strings and normal solution text strings, calculates the text strings similarity (" 0 ", " 1 ", " 0.5 ", " 0.125 ", " 0.75 ") shown in Figure 19 A.The corresponding normal solution text strings of answer determination section 150 decides the distribution of the answer corresponding normal solution text strings corresponding with each answer frame according to this text strings similarity.The normal solution text strings that the corresponding normal solution text strings of answer determination section 150 is tried one's best the text strings similarity big is dispensed to the answer frame as the corresponding normal solution text strings of answer, so that the normal solution text strings that is dispensed in each answer frame does not repeat mutually.

Shown in Figure 19 B, the answer frame that 150 pairs of the corresponding normal solution text strings of answer determination sections have the answer text strings " basic people Goodwill respects " of highest similarity (" 1 ") distributes the normal solution text strings corresponding with it " basic people Goodwill respects ".Then, the text strings similarity is calculated the answer text strings " state people master Righteousness " that 248 pairs in portion has the second high similarity (" 0.75 ") and is distributed the normal solution text strings " state democracy Goodwill " corresponding with it.Then, 150 pairs of the corresponding normal solution text strings of answer determination sections have the 3rd high answer text strings " Min Zhu Zhu Righteousness " similarity (" 0 ") or remaining and distribute corresponding with it or remaining normal solution text strings " War strives and puts Abandoned " in unduplicated mode.

When Figure 20 is illustrated in the answer text strings and comprises normal solution and non-normal solution, by answer text strings display part 214, correctness judge that display part 218 and normal solution text strings display part are 216 that show, answer text strings (" Min Zhu Zhu Righteousness ", " basic people Goodwill respects " reach " Guo Min Zhu Righteousness ") in the answer frame, at the result of determination (for example " zero " of the answer text strings in the answer frame, " * ") and the example of the corresponding normal solution text strings of answer (" basic people Goodwill respects ", " state democracy Goodwill " reaches " War strives and puts Abandoned ")." zero " of thick dashed line reaches correct (zero) or the non-normal solution (*) that " * " symbolic representation is used to judge correctness in the drawings.

Figure 21 A illustrates by user input and by the answer text strings and reads two answer text strings that portion 242 reads (" Gao Zhi Ken " and “ Fu Island Ken "), possible four normal solution text strings (" Gao Zhi Ken ", " Love Yuan Ken ", " Xiang Chuan Ken " and “ De Island Ken ") and the example of the similarity calculated.Figure 21 B illustrates the distribution or the correspondence of the corresponding normal solution text strings of answer of carrying out at the answer text strings based on the similarity between the corresponding normal solution text strings of answer text strings and answer.

Shown in Figure 21 B, the text strings similarity is calculated answer text strings " height is known Ken " the distribution normal solution text strings " height know Ken " corresponding with it that 248 pairs in portion has highest similarity (" 1 ").Then, the text strings similarity is calculated the answer text strings " good fortune Island Ken " that 248 pairs in portion has the second high similarity (" 0.5 ") and is distributed normal solution text strings " moral Island Ken ".

When Figure 22 is illustrated in the answer text strings and comprises normal solution and non-normal solution, by answer text strings display part 214, correctness judge answer text strings in the answer frame that display part 218 and normal solution text strings display part 216 show (" Gao Zhi Ken " and “ Fu Island Ken "), at the example of the judgement Jie Ken (for example " zero ", " * ") of the answer text strings in the answer frame and the corresponding normal solution text strings of answer (" Gao Zhi Ken " and “ De Island Ken ").At this moment, only show and immediate two the normal solution text strings of answer text strings that remaining possible normal solution text strings " Love Yuan Ken " reaches " Xiang Chuan Ken " and do not show as the normal solution text strings.Therefore, only show with learner's the immediate answer of wish or with the immediate normal solution text strings of the text strings that will answer.But, also can append remaining possible normal solution text strings " Love Yuan Ken " and reach " Xiang Chuan Ken ", distinguish, show as possible answer text strings.

Among Figure 12, the answer text strings that normal solution text strings display position obtaining section 260 obtains in each answer frame, obtain expression and store the coordinate (X of the corresponding normal solution text strings of the answer display position of each answer frame upside display position information storage part 262, in the display device 112 into, Y), with the position is corresponding separately, normal solution text strings display part 216 is showing the corresponding normal solution text strings of answer on the position separately with the corresponding normal solution text strings of answer.

Figure 23 A illustrates the display position (coordinate) of the answer text strings correspondence normal solution text strings that obtains at shown three answer frame identiflication numbers (ID) the answer text strings 1,2,3, that imported of display device 112 (" Min Zhu Zhu Righteousness ", " basic people Goodwill respects " reach " Guo Min Zhu Righteousness ") and by normal solution text strings display position obtaining section 260.The corresponding normal solution text strings of answer (" basic people Goodwill respects ", " state democracy Goodwill " reach " War strives and puts Abandoned ") that Figure 23 B illustrates text strings similarity according to Figure 19 A, distributed by the corresponding text strings determination section 250 of answer.

When Figure 24 is illustrated in the answer text strings and comprises normal solution and non-normal solution, judge that by answer text strings display part 214, correctness display part 218 and normal solution text strings display part are 216 that show, at the result of determination of the answer text strings in the answer frame (" Min Zhu Zhu Righteousness ", " basic people Goodwill respects " reach " Guo Min Zhu Righteousness ") (for example, " zero ", " * ") and the example of the corresponding normal solution text strings of answer (" basic people Goodwill respects ", " state democracy Goodwill " reach " War strives and puts Abandoned ").Among the figure, emphasize to show wrong literal or different word (not needing to show dotted line itself) in the answer frame by for example different color (for example, red, blueness) with other literal (for example, black).

Figure 25 illustrates the structure of another study assistance system 200 of being realized by the signal conditioning package 10 of Fig. 1 and/or study assist server 20.

Study assistance system 200 is identical with the situation of Fig. 2, comprise be arranged in the signal conditioning package 10 as the input handling part, answer input part 210 and correctness judge instruction unit 212, comprise be arranged in the signal conditioning package 10 as display process portion, answer text strings display part 214, normal solution text strings display part 216 and correctness judge display part 218.

Study assistance system 200 also comprises to be arranged in signal conditioning package 10 or the study assist server 20 to be read portion 242, literal as answer text strings management department 222 information treatment part, the management handwriting information, the answer text strings of reading handwriting information and intercepts regional candidate's generating unit 244, indivedual character recognition portion 246, maximum likelihood text strings determination section 247, text strings similarity and calculate portion 248, the corresponding text strings determination section 250 of answer and correctness detection unit 254.Study assistance system 200 also comprises in memory storage 14 or 24 as answer text strings storage part 252, the literal zone, storage answer handwriting information and intercepts regional candidate's storage part 254, literal identification dictionaries store portion 256 and normal solution text strings group storage part 258.

Among Fig. 1, when the user uses when writing out the answer text strings successively in the answer hurdle of writing pencil on tablet 114, answer input part 210 is taken into successively that (X, Y) the person's handwriting coordinate data of Zu Chenging offers answer text strings management department 222 by continuous a plurality of coordinates corresponding with the answer text strings.Answer text strings management department 222 is stored to answer text strings storage part 252.Answer text strings management department 222 stores this person's handwriting coordinate data in the answer text strings storage part 252 successively, and offers answer text strings display part 214 successively.Answer text strings display part 214 shows the person's handwriting of the handwriting string of being represented by this person's handwriting coordinate data successively in real time on display device 112.

When user's operation display device 112 shown correctness are judged indication key, when operating correctness judgement instruction unit 212 thus, the answer text strings is read portion 242 and is read the person's handwriting coordinate data that answer text strings storage part 252 is stored, and offers literal and intercepts regional candidate's generating unit 244.Literal intercepts regional candidate's generating unit 244 and intercept the character area candidate who estimates from the person's handwriting coordinate data of reading, a plurality of groups that output is made up of a succession of character area candidate respectively.Employing is the regional candidate that a known method such as character area is estimated a literal with each stroke in the center of gravity zone intercepting within the specific limits of X-direction.

Figure 26 A illustrates by user input and is read the example of the person's handwriting of the hand-written answer text strings (" Dry state ") that portion 242 is taken into by the answer text strings.Figure 26 B illustrates person's handwriting coordinate data at Figure 26 A, is intercepted the example of literal intercepting candidate a～d that regional candidate's generating unit 244 generates by literal.Figure 26 C illustrates and utilizes the literal shown in 246 couples of Figure 26 B of indivedual character recognition portion to intercept an identification candidate character that regional candidate a～d identifies and an example of identification mark.

Among Figure 26 B, a character area candidate's having intercepted pattern by the upper left side with each zone of surrounding of cubic frame of letter, arrow → be illustrated in connection (link) relation that might connect between character area candidate (node).This example has been pointed out the character area candidate's of the annexation that is in group " a-b-d " and group " c-d " (Dry-state) two groups or path (paths).Among Figure 26 C, the identification mark of each identification candidate character of numeral that the right side of respectively discerning candidate character in each character area candidate (node) is shown, wherein, the numeral of mark is big more, and expressing possibility property or maximum likelihood degree are high more.

Indivedual character recognition portion 246 decide identification candidate character and identification mark at each mould case a～d of the character area candidate of Figure 26 B that is intercepted regional candidate's generating unit 244 interceptings by literal with reference to literal identification dictionaries store portion 256.Can use generally the OCR identification of adopting to draw sincere, ONLINE RECOGNITION in the extraction of identification candidate character draws identification arbitrarily such as sincere and draws sincere.The identification mark can use the pattern that calculates the character area candidate (a～d) and the distance between the authentic language pattern in the literal identification dictionaries store portion 256 or known the whole bag of tricks such as calculate according to unique point.

Indivedual character recognition portion 246 offer maximum likelihood text strings determination section 247 with a plurality of identification candidate characters among each character area candidate and each identification mark.Maximum likelihood text strings determination section 247 determines the maximum likelihood text strings with reference to the information of the normal solution text strings in the normal solution text strings group storage part 258 corresponding with the identifying information of each answer frame or identiflication number (h) from the identification candidate character.The processing of decision maximum likelihood text strings is as follows: with each group of forming by a succession of character area candidate or node or each path and each character area candidate (node) accordingly at a plurality of identification candidate characters from indivedual character recognition portion 246, check each literal in (coupling) itself and normal solution text strings, select literal, determine the maximum likelihood text strings thus.In this is checked, if exist with the checking of normal solution text strings in the consistent identification candidate character of literal of text point, then at the identification candidate character of this unanimity, add for example certain additive score (award mark), be set at the normal solution text strings thus and be selected as the maximum likelihood text strings easily.

With reference to Figure 26 C, the example of decision maximum likelihood text strings is shown.In this example, exist and select two paths " a-b-d " of candidate character string to reach " c-d ".Maximum likelihood text strings determination section 247 is at each candidate character in the node of each selection path, and the literal of each text point in itself and the normal solution text strings is checked.247 pairs of maximum likelihood text strings determination sections and the inconsistent identification candidate character of literal in the normal solution text strings are given the mark (for example, 650) of minimum, as the literal verify.In addition, maximum likelihood text strings determination section 247 will by to the identification candidate character consistent with the literal in the normal solution text strings, in the identification mark of this identification candidate character, add additive score (for example, 50) and the value that obtains, as the literal verify.

Like this, maximum likelihood text strings determination section 247 is selected the highest text strings of text strings verify.The text strings verify is represented to the mean value of the literal verify of current text point by the initial text point from the path, promptly, obtain for will be from initial text point to the average text strings verify of preceding text point * the literal verify of literal number+current text point from initial text point to preceding text point, divided by alpha-numeric value from initial text point to current text point.Therefore, the text strings verify in the final node in each path represent the text strings in each path the text strings verify, be average literal verify.Maximum likelihood text strings determination section 247 selects to be in the text strings in path of text strings verify maximum of last text point of the text strings after checking as the maximum likelihood text strings.

Figure 27 A～27D illustrates the literal verify of the identification candidate character that utilizes maximum likelihood text strings determination section 247 to obtain Figure 26 C and text strings verify, a decision (“ South-Korea state at the normal solution text strings ") the example of order of maximum likelihood text strings of literal intercepting candidate a～d of Figure 26 B.This example is as follows: when literal being intercepted regional candidate a～d and normal solution text strings “ South-Korea state " when checking, comprise that 3 continuous literal decide the maximum likelihood text strings with the testing process of interior wrong literal.

In Figure 27 A, by maximum likelihood text strings determination section 247 at a normal solution text strings " South-Korea state "; select by a succession of character area candidate a among Figure 26 C, a group that b, d form; about initial start node a (radical of " Dry " (left side radicals by which characters are arranged in traditional Chinese dictionaries)); will check text point X and be made as initial value X=1; in checking text point q, set to check and begin text point q=X and next text point q=X+1, thereby check with the 1st literal " South-Korea " and the 2nd literal " state " of normal solution text strings successively.

Obtain identification candidate character " Zhuo " Ji “ Testis at node a " identification mark 780 and 760.Because “ South-Korea Zhong the identification candidate character " or " state " do not exist, so " Zhuo " of selecting identification mark maximum is as selecting literal.At this moment because the literal of having selected and check literal “ South-Korea " or " state " all inconsistent, be used as checking text point so give X-1=0, give the inconsistent minimum of expression (for example, 650) and be used as the literal verify.Owing to be initial text point, the text strings verify equates with the literal verify.Because there is not the node that connects before the node a, so the text point of character area candidate a is recorded as Y=1, last selection nodes records is " nothing ".

About next node b (radical of " Dry " (right side radicals by which characters are arranged in traditional Chinese dictionaries)), the text point 0 of checking of the node a that will connect before node b adds 1 and make it become X=1, set and to check beginning text point q=X and next text point q=X+1 checking text point q, and with the 1st and the 2nd literal “ South-Korea of normal solution text strings " reach " state " and check successively.Obtain the identification candidate character and " make " the identification mark 770 and 740 that reaches " the present ".Because “ South-Korea Zhong the identification candidate character " or " state " do not exist, so " order " of selecting to have maximum identification mark is used as selecting literal.The literal of having selected " makes " and checks literal “ South-Korea " or " state " all inconsistent, therefore give the inconsistent minimum of expression (for example, 650) and be used as the literal verify, give X-1=0 as checking text point.The text strings verify is at the mean value of the literal verify of literal several 2 promptly 650.The text point 1 of the node a that will connect before node b adds 1, and the text point of character area candidate b is recorded as Y=2, and the last selection nodes records that will be connected with node b is " a ".

With reference to Figure 27 B, about next node d (" state "), the text point 0 of checking of the node b that will connect before node d adds 1 and make it become X=1, set and to check beginning text point q=X and next text point q=X+1 checking text point q, and reach " state " with the 1st and the 2nd literal " Dry " of normal solution text strings and check successively.Obtain identification candidate character " state " reach " Gu " identification mark 860 and 790.In the identification candidate character, there be " state ", so the literal of selecting to have maximum identification mark " state " is used as selecting literal.Because the literal selected is used as checking text point with to check literal consistent so give X=2, identification mark 860 is added the additive value (for example 50) of expression unanimity and gives 910 and be used as the literal verify.The text strings verify is the mean value 736 at the literal verify of literal several 2.As the text point of character area candidate d, record Y=3, and as the last selection node that is connected with node d, record " b ".

Then, in Figure 26 A, at normal solution text strings “ South-Korea state "; select by a succession of character area candidate c among Figure 26 C, a group that d forms; about initial start node c (" Dry "); will check text point X and be made as initial value X=1, set to check and begin text point q=X and next text point q=X+1 checking text point q, and with the 1st and the 2nd literal “ South-Korea of normal solution text strings " reach " state " and check successively.Then, obtain the identification mark 830 and 820 that reaches " court " at the identification candidate character " Dry " of node c.Because “ South-Korea Zhong the identification candidate character " or " state " do not exist, so select " Dry " of identification mark maximum to be used as selecting literal.At this moment because the literal selected and check literal “ South-Korea " or " state " all inconsistent, be used as checking text point so give X-1=0, give the inconsistent minimum of expression (for example, 650) and be used as the literal verify.Owing to be initial text point, the text strings verify equates with the literal verify.Because the node that connected before node a does not exist, so the text point of character area candidate a is recorded as Y=1, is " nothing " with last selection nodes records.

With reference to Figure 27 C, about next node d (" state "), the text point 0 of checking of the node c that will connect before node d adds 1 and make it become X=1, set and to check beginning text point q=X and next text point q=X+1 checking text point q, and with the 1st and the 2nd literal “ South-Korea of normal solution text strings " reach " state " and check successively.Obtain identification candidate character " state " reach " Gu " identification mark 860 and 790.Because in the identification candidate character, have " state ", so the literal of selecting to have maximum identification mark " state " is used as selecting literal.Because the literal selected is used as checking text point with to check literal consistent so give X=2, identification mark 860 is added the additive value (for example, 50) of expression unanimity and gives 910 and be used as the literal verify.The text strings verify is the mean value 780 at the literal verify of literal several 2.As the text point of character area candidate d, record Y=2, and as the last selection node that is connected with node d, record " c ".

About node d, text strings verify 736 when finding out a group's who is made up of a succession of character area candidate a, b, d path and the text strings verify 780 when finding out a group's who is made up of a succession of character area candidate c, d path compare the side's that storage text strings verify is high data (check text point, check literal, select literal, literal verify, text strings verify, text point, last selection node).

At last, seek the text strings of corresponding with the end node text strings verify maximum of storing.Under the situation of Figure 26 C and Figure 27 A～27C, the end node only is d, thus begin to seek successively last selection node from node d, and the path that decision will be selected is as the maximum likelihood text strings.As Figure 27 D, the text strings decision that will begin to connect determined path c-d from the start node of literal intercepting is preserved for maximum likelihood text strings " Dry state ".

Figure 28 A～28D illustrates by maximum likelihood text strings determination section 247 and obtains the literal verify of identification candidate character of Figure 26 C and text strings verify, decision another (“ Da South-Korea the Republic of China at the normal solution text strings ") the example of order of maximum likelihood text strings of literal intercepting candidate a～d of Figure 26 B.This example is as follows: literal is being intercepted regional candidate a～d and normal solution text strings “ Da South-Korea the Republic of China " when checking, comprise the testing process of 3 continuous literal, decision maximum likelihood text strings with interior wrong literal.

In Figure 28 A, select by a succession of character area candidate a among Figure 26 C, a group that b, d form at another normal solution text strings " big South-Korea the Republic of China " by maximum likelihood text strings determination section 247, the 1st～the 4th literal " greatly ", " South-Korea ", " people " of node a (radical of " Dry " (left side radicals by which characters are arranged in traditional Chinese dictionaries)), b (radical of " Dry " (left side radicals by which characters are arranged in traditional Chinese dictionaries)) and normal solution text strings are reached " state " and check successively.

Identical with the situation of Figure 26 B, obtain identification mark successively at the identification candidate character of node a, b." greatly ", “ South-Korea in the identification candidate character ", " people " reach " state " and do not exist, so selection " Zhuo " of discerning the mark maximum reaches " order " and be used as selecting literal respectively.Give X-1=0 respectively and be used as checking text point, give the inconsistent minimum of expression (for example, 650) respectively as the literal verify.The text strings verify equates with the literal verify.

With reference to Figure 28 B, identical with the situation of Figure 27 B, about node d (" state "), the text point X=0 that checks of the node b that will connect before node d adds 1 and make it become X=1, set and to check beginning text point q=X and next text point q=X+1, X+2 and X+3 checking text point q, and with the 1st, the 2nd, the 3rd and the 4th literal " greatly ", “ South-Korea of normal solution text strings ", " people " reach " state " and check successively.Obtain identification candidate character " state " reach " Gu " identification mark 860 and 790.In the identification candidate character, there be " state ", so the literal of selecting to have maximum identification mark " state " is used as selecting literal.Because the literal selected and to check literal consistent, therefore giving X=4 is used as checking text point, and identification mark 860 is added the additive value (for example 50) of expression unanimity and gives 910 and be used as the literal verify.The text strings verify is the mean value 736 at the literal verify of literal several 3.As the text point of character area candidate d, record Y=3, and as the last selection node that is connected with node d, record " b ".

Then, in Figure 28 A, at normal solution text strings “ Da South-Korea the Republic of China ", select by a succession of character area candidate c among Figure 26 C, a group that d forms, with the 1st～the 4th literal " greatly ", “ South-Korea of node c (" Dry ") and normal solution text strings ", " people " reach " state " and check successively.Then, obtain identification mark at the identification candidate character of node c.Because " greatly ", “ South-Korea in the identification candidate character ", " people " reach " state " and do not exist, and is used as selecting literal so " Dry " of mark maximum discerned in selection.Give X-1=0 as checking text point, give the inconsistent minimum of expression (for example, 650) and be used as the literal verify.The text strings verify equates with the literal verify.

With reference to Figure 28 C, identical with the situation of Figure 26 B, about node d (" state "), the text point X=0 that checks of the node b that will connect before node d adds 1 and make it become X=1, set and to check beginning text point q=X and next text point q=X+1, X+2 and X+3 checking text point q, and with the 1st, the 2nd, the 3rd and the 4th literal " greatly ", “ South-Korea of normal solution text strings ", " people " reach " state " and check successively.Obtain identification candidate character " state " reach " Gu " identification mark 860 and 790.Because in the identification candidate character, have " state ", so the literal of selecting to have maximum identification mark " state " is used as selecting literal.Because the literal of having selected is used as checking text point with to check literal consistent so give X=4, identification mark 860 is added the consistent additive value of expression (for example 50) and gives 910 and be used as the literal verify.The text strings verify is the mean value 780 at the literal verify of literal several 2.As the text point of character area candidate d, the last selection node that record Y=2 and conduct are connected with node d, record " c ".

About node d, text strings verify 736 when finding out a group's who is made up of a succession of character area candidate a, b, d path and the text strings verify 780 when finding out a group's who is made up of a succession of character area candidate c, d path compare the side's that storage text strings verify is high data (check text point, check literal, select literal, literal verify, text strings verify, text point, last selection node).

At last, seek the text strings of corresponding with the end node text strings verify maximum of storing.Under the situation of Figure 26 C and Figure 27 A～27C, the end node only is d, so begin to find out successively last selection node from node d, the path that decision will be selected is used as the maximum likelihood text strings.As Figure 28 D, the text strings decision that will begin to connect determined path c-d from the start node of literal intercepting is preserved for maximum likelihood text strings " Dry state ".

Like this, the literal of checking that each literal is intercepted N adjacency in regional candidate and the normal solution text strings is checked, even thus unnecessary literal and lack word continuously and be N-1 individual with interior situation under, also can intercept regional candidate and the normal solution text strings is checked to literal.

The conduct that is determined by maximum likelihood text strings determination section 247 is sent to the text strings similarity at the maximum likelihood text strings of the answer text strings of each normal solution text strings and text strings verify thereof and calculates portion 248.The text strings similarity is calculated portion 248 as previously mentioned, calculates as the answer text strings of maximum likelihood text strings and the similarity between each normal solution text strings.

Figure 29 is used for the method that explanation utilizes answer correspondence normal solution text strings determination section 250 to decide at the corresponding normal solution text strings of answer of maximum likelihood text strings (" Dry state ").

With reference to Figure 29, the corresponding normal solution text strings of answer determination section 250 is according to maximum likelihood text strings " Dry state " and normal solution text strings “ South-Korea state " and “ Da South-Korea the Republic of China " determine benchmark literal number " 2 ", " 4 ", and calculate the text strings similarity.The text strings similarity for example can be by representing with following formula.

Text strings similarity=text strings verify+100 * (consistent literal number/benchmark literal number)

Here, the some big number in benchmark literal number=" the literal number of normal solution text strings " and " the literal number of maximum likelihood text strings ".

The corresponding normal solution text strings of answer determination section 250 will have the text strings “ South-Korea state of the highest text strings similarity (830) among the normal solution text strings group " determine to be the corresponding normal solution text strings of answer.When existence has the corresponding normal solution text strings of a plurality of answers of identical maximum text strings similarity, also can determine the corresponding normal solution text strings of a plurality of answers.

The example of the correctness result of determination of the answer text strings when Figure 30 illustrates the non-normal solution that the study assistance system 200 by Figure 25 carries out.At this moment, in the answer frame, show handwriting.Other display formats are identical with the situation of Fig. 4 A or 4B.Structure additional to the study assistance system 200 of Figure 25 and Fig. 5,12 study assistance system 200 is identical appends textural element, in the answer frame, show handwriting rather than type printing body thus, in addition, can be by carrying out same demonstration with the same processing of Fig. 7,9,11,14,16,18,20,22 and 24 situation.

Figure 31 A and 31B illustrate the process flow diagram of being judged the processing of normal solution/non-normal solution by the study assistance system 200 of Fig. 2 being used to of carrying out.

With reference to Figure 31 A, in step 502, the answer text strings is read portion 242 according to the indication (order) that normal solution/non-normal solution is judged of judging instruction unit 212 from correctness, from answer text strings group storage part 252, take out the answer text strings data in the answer frame, offer the text strings similarity and calculate portion 248.In step 504, the text strings similarity is calculated the normal solution text strings identiflication number of the more than one normal solution text strings that portion 248 will can be used as answer or the value of serial number i is set at initial value i=1, in addition maximum similarity S is set at initial value S=0.In step 508, the text strings similarity is calculated portion 248 and is judged that numbering i are whether greater than the number N (i＞N), promptly whether do not have the normal solution text strings of identiflication number i (i) of normal solution text strings.When being judged to be numbering i greater than the number of normal solution text strings, when promptly not having i normal solution text strings, flow sequence enters the step 532 of Figure 31 B.

When in step 508, be judged to be the numbering i be not more than the number N of normal solution text strings, when promptly having i normal solution text strings, the text strings similarity is calculated portion 248 in step 510, calculates the text strings similarity Si between i normal solution text strings CSi and the answer text strings.

In step 512, the text strings similarity is calculated portion 248 and is judged that whether the text strings similarity Si of the normal solution text strings of numbering i is greater than current maximum similarity S.When being judged to be text strings similarity Si greater than maximum similarity S, the text strings similarity is calculated portion 248 and in step 514 text strings similarity Si is set at maximum similarity S.In step 516, the text strings similarity is calculated portion 248 and is selected i normal solution text strings CSi as the corresponding normal solution text strings of answer, is stored in the operating area of self temporarily.At this moment, when having more than one described answer normal solution text strings, the text strings similarity is calculated portion 248 with its deletion.In step 518, the 248 numbering i with the normal solution text strings of text strings similarity operator portion are set at Next Serial Number i=i+1.Then, flow sequence is returned step 508.

When in step 512, being judged to be text strings similarity Si when being not more than current maximum similarity S, calculating portion 248 in step 524 Chinese character string similarity and judge whether the text strings similarity Si of the normal solution text strings of numbering i equals maximum similarity S.When being judged to be text strings similarity Si when equaling maximum similarity S, in step 526, the text strings similarity is calculated portion 248 and is selected i normal solution text strings CSi to be used as the corresponding normal solution text strings of answer, and is appended in self the operating area and stores temporarily.Then, flow sequence enters step 518.

Structure instead, when being judged to be text strings similarity Si when being not more than maximum similarity S in step 512, flow sequence is execution in step 524 and 526 not also, and enters step 518.At this moment, select to have junior one normal solution text strings of identical maximum similarity, as answer correspondence normal solution text strings.

When being judged to be text strings similarity Si when being not equal to maximum similarity S in step 524, flow sequence enters step 518.

With reference to Figure 31 B, when in the step 508 of Figure 31 A, being judged to be numbering i greater than the number N of normal solution text strings, when promptly not having i normal solution text strings, the text strings similarity is calculated portion 248 in step 532, corresponding normal solution text strings CSi of answer and maximum similarity S are offered the corresponding normal solution text strings of answer determination section 250, and the corresponding normal solution text strings of answer determination section 250 offers correctness detection unit 254 with the value of the corresponding normal solution text strings of answer CSi.

In step 534,250 judgements of the corresponding normal solution text strings of answer determination section whether store normal solution text strings CSi in the temporary storage area of the corresponding normal solution text strings of answer or whether maximum similarity S is S＞0.When being judged to be when in the corresponding normal solution text strings of answer, storing normal solution text strings CSi or to be judged to be maximum similarity be S＞0, in step 536, the normal solution text strings of the corresponding normal solution text strings of answer determination section 250 beyond the corresponding normal solution text strings of temporary storage area storage answer that the non-corresponding text strings of answer is used provides this answer non-corresponding normal solution text strings to normal solution text strings display part 216.In step 540,216 differences of normal solution text strings display part show corresponding normal solution text strings of answer and the non-corresponding normal solution text strings of answer (with reference to Fig. 4 A).Structure instead, flow sequence be execution in step 536 and enter step 540 (with reference to Fig. 4 B) not also.At this moment, in step 540, normal solution text strings display part 216 only shows the corresponding normal solution text strings of answer.

When in step 534, being judged to be when in the corresponding normal solution text strings of answer, not storing normal solution text strings CSi, the corresponding normal solution text strings of answer determination section 250 is in step 538, the whole normal solution text strings CS1～CSN of temporary storage area storage to the non-corresponding normal solution text strings of answer is used provides this answer non-corresponding normal solution text strings to normal solution text strings display part 216.

In addition, when in step 512, being judged to be text strings similarity Si when being not more than maximum similarity S, flow sequence is execution in step 524 and 526 not, and enter step 410, when junior one normal solution text strings of only selecting to have identical maximum similarity is used as answer correspondence normal solution text strings, the corresponding normal solution text strings of answer determination section 250 is only stored normal solution text strings group's the initial normal solution text strings with maximum similarity S to the temporary storage area that the corresponding normal solution text strings of answer is used in step 536, and only provides this answer corresponding normal solution text strings to normal solution text strings display part 216.Then, flow sequence enters step 540.

In step 542, correctness detection unit 254 judges whether the corresponding normal solution text strings of answer text strings and answer is consistent.When being judged to be the corresponding normal solution text strings of answer text strings and answer when consistent, in step 544, correctness result of determination display part 218 carries out normal solution (for example " zero's ") demonstration as the normal solution result of determination in the answer frame of display device 112.Then, flow sequence withdraws from the flow process of Figure 31 A and 31B.

When in step 542, being judged to be the corresponding normal solution text strings of answer text strings and answer when inconsistent, in step 546, correctness result of determination display part 218 carries out non-normal solution in the answer frame of display device 112 demonstration (for example " * ") is used as the correctness result of determination.Then, flow sequence withdraws from the flow process of Figure 31 A and 31B.

In step 546, corresponding normal solution text strings of 256 pairs of determined answers of the wrong text point extraction unit of Fig. 5 and answer text strings compare, and extract the wrong text point in the corresponding normal solution text strings of determined answer.Wrong text point in 216 pairs of normal solution text strings display parts and the corresponding normal solution text strings of answer (different words, unnecessary word, lack word) corresponding character or text point is emphasized to show or underscore shows.In addition, the check on one's answers literal (different words, unnecessary word, lack word) of the wrong text point in the text strings of the answer text strings changing unit 259 of Figure 12 emphasizes to show that (Figure 14,16,18) or underscore show (Figure 16).

Figure 32 A～32C is illustrated in the step 546 of Figure 31 B the process flow diagram of processing that is extracted the wrong text point of the corresponding normal solution text strings of answer CSi by the study assistance system 200 (mistake text point extraction unit 256) of Fig. 5 being used for of carrying out.

Mistake text point extraction unit 256 is in step 602, read the answer text strings that receives and store into the temporary storage area from the corresponding normal solution text strings of answer determination section 250, in step 604, read the corresponding normal solution text strings of answer that receives and store into the temporary storage area from the corresponding normal solution text strings of answer determination section 250.Mistake text point extraction unit 256 is in step 606, the Position Number m of the literal in the corresponding normal solution text strings of answer is set at initial value m=1, the Position Number n of the literal in the answer text strings is set at initial value n=1, checking of normal solution text strings is begun text point numbering x be set at initial value x=1.

In step 608, whether whether the Position Number m that mistake text point extraction unit 256 is judged the literal in the corresponding normal solution text strings of answers greater than length (literal number) M of the corresponding normal solution text strings of answer, promptly extract at the wrong literal that is through with of the whole literal in the corresponding normal solution text strings of answer.When being judged to be Position Number m greater than the length of the corresponding text strings of answer, flow sequence enters the step 640 of Figure 32 C.

When in step 608, being judged to be Position Number m when being not more than the length M of the corresponding normal solution text strings of answer, in step 612 wrong text point extraction unit 256 judge the normal solution text strings check beginning text point numbering x whether greater than length (literal number) N of answer text strings, i.e. all comparisons of literal in the text strings that check on one's answers that whether have been through with.When be judged to be the normal solution text strings check beginning text point numbering x greater than length (literal number) N of answer text strings the time, flow sequence enters step 632.

When in step 612, being judged to be the checking beginning text point numbering x and be not more than length (literal number) N of answer text strings of normal solution text strings, in step 616, the Position Number n that mistake text point extraction unit 256 is judged the literal in the answer text strings whether greater than length (literal number) N of answer text strings, promptly whether at the whole literal in the answer text strings be through with literal relatively or extract mistake.The Position Number n of the literal in being judged to be the answer text strings is during greater than length (literal number) N of answer text strings, and flow sequence enters step 634.When the Position Number n of the literal when be judged to be the answer text strings in step 616 in was not more than length (literal number) N of answer text strings, flow sequence entered the step 620 of Figure 32 B.

During with reference to Figure 32 B, in step 620, mistake text point extraction unit 256 judges whether the m literal of the corresponding normal solution text strings of answer is consistent with the n literal of answer text strings.When the m literal that is judged to be the corresponding normal solution text strings of answer is consistent with the n literal of answer text strings, in step 624, mistake text point extraction unit 256 with the normal solution text strings check beginning text point numbering add 1 and be set be x=n+1, the Position Number A[m of the literal of answer text strings that will be consistent with the m literal of the corresponding normal solution text strings of answer]=n is stored in the temporary storage area.In step 626, mistake text point extraction unit 256 adds 1 and be set and be m=m+1 with the Position Number of the literal of the corresponding normal solution text strings of answer, check beginning text point numbering with what the Position Number n of the literal of answer text strings was set at the normal solution text strings, promptly be set at n=x.Then, flow sequence is returned step 608.

When in step 612, be judged to be the answer text strings check beginning text point numbering x greater than length (literal number) N of answer text strings the time, the literal in the answer text strings that the m literal of corresponding normal solution text strings with answer is corresponding not, so wrong text point extraction unit 256 in step 632, in order to be illustrated in the situation that has unnecessary word in the answer text strings, set A[m as text point numbering at the answer text strings of the m literal of the corresponding normal solution text strings of answer]=-1 and store, the Position Number of the literal of the corresponding normal solution text strings of answer is added 1 and be set and be m=m+1.Then, flow sequence is returned step 608.

The Position Number n of the literal when be judged to be the answer text strings in step 616 in is during greater than length (literal number) N of answer text strings, do not have will corresponding normal solution text strings with answer the answer text strings checked of m literal in literal, so in step 634, mistake text point extraction unit 256, in order to be illustrated in the situation that has different words or scarce word in the answer text strings, set A[m as Position Number at the literal of the answer text strings of the m literal of the corresponding normal solution text strings of answer]=0 and store, the Position Number of the literal of the corresponding normal solution text strings of answer is added 1 and be set and be m=m+1, check beginning text point numbering with what the Position Number n of the literal of answer text strings was set at the answer text strings, promptly be set at n=x.Then, flow sequence is returned step 608.

When the n literal of m literal that is judged to be the corresponding normal solution text strings of answer in step 620 and answer text strings is inconsistent, in step 636, mistake text point extraction unit 256 Position Numbers with the literal in the answer text strings add 1 and be set and be n=n+1.Then, flow sequence is returned step 616.

With reference to Figure 32 C, the Position Number m of the literal when be judged to be the corresponding normal solution text strings of answer in step 608 in is during greater than the length M of the corresponding normal solution text strings of answer, in step 640, mistake text point extraction unit 256 is set at the text point numbering m of answer correspondence normal solution text strings the length M of the corresponding normal solution text strings of answer and makes m=M, and the gap d that the text point of the corresponding normal solution text strings of answer is numbered the text point numbering of m and the answer text strings that will check with it is set at initial value d=0.

In step 642, mistake text point extraction unit 256 judges that whether the text point numbering m of the corresponding normal solution text strings of answers is 0 is m=0.When being judged to be when not being m=0, in step 646, mistake text point extraction unit 256 is judged the text point numbering A[m at the answer text strings of the m literal of the corresponding normal solution text strings of the answer of having stored] whether be-1.

As the text point numbering A[m that in step 646, is judged to be at the answer text strings of the m literal of the corresponding normal solution text strings of answer] when being-1, in step 648, mistake text point extraction unit 256 is appended the text point m of the corresponding normal solution text strings of answer and is stored in lacking the word location tabulation.Then, flow sequence enters step 668.

In step 668, mistake text point extraction unit 256 deducts 1 and make it become m=m-1 from the text point of the corresponding normal solution text strings of answer.Then, flow sequence enters step 642.

As the text point numbering A[m that in step 646, is judged to be at the answer text strings of the m literal of the corresponding normal solution text strings of answer] when not being-1, mistake text point extraction unit 256 is judged the text point numbering A[m at the answer text strings of the m literal of the corresponding normal solution text strings of the answer of being stored] whether be 0.Be judged to be A[m] be under 0 the situation, wrong text point extraction unit 256 judges that whether gap d are 0, promptly whether are d=0 in step 654.When be judged to be checking of corresponding normal solution text strings of answer and normal solution text strings do not have gap d between the text point, when being d=0, in step 656, mistake text point extraction unit 256 is appended the text point m of normal solution text strings and is stored in different word location tabulations.Then, flow sequence enters step 668.

The text point of checking that is judged to be corresponding normal solution text strings of answer and normal solution text strings in step 654 has gap d, is d ≠ 0 o'clock, mistake text point extraction unit 256 is appended text point m and is stored in the tabulation of scarce word location, gap d is deducted 1 and be set and be d=d-1.Then, flow sequence enters step 668.

When in step 650, being judged to be A[m] when being not 0, wrong text point extraction unit 256 judges that gap d are whether greater than the text point numbering A[m at the answer text strings of the m literal of the corresponding normal solution text strings of answer in step 660] and the text point numbering m of the corresponding normal solution text strings of answer between difference absolute value (| A[m]-m|).When being judged to be gap d greater than text point numbering A[m at the answer text strings of the m literal of the corresponding normal solution text strings of answer] and the text point of the corresponding normal solution text strings of answer when numbering the absolute value of difference of m, at the size of the gap d that checks text point numbering m of the text point of the answer text strings of the m literal of the corresponding normal solution text strings of answer and the corresponding normal solution text strings of answer the gap size of checking text point less than the position after leaning on than text point m, so in step 662, mistake text point extraction unit 256 is appended text point m and is stored in unnecessary word location tabulation.

In step 664, mistake text point extraction unit 256 is set at d=|A[m with gap d]-m|.Enter step 668 then.

Be not more than the text point numbering A[m of the answer text strings of the m literal of corresponding normal solution text strings when in step 660, being judged to be gap d at answer] and the text point m of the corresponding normal solution text strings of answer between the absolute value of difference the time, flow sequence enters step 664.

When in step 642, being judged to be m=0, in step 672, mistake text point extraction unit 256 is judged the text point A[M of the answer text strings that the literal of last text point M that will corresponding normal solution text strings with answer is checked] whether less than the length N of answer text strings (A[M]＜N).When being judged to be A[M] during less than N, in step 674, mistake text point extraction unit 256 is appended last text point M in unnecessary word location tabulation.Then, flow sequence withdraws from the flow process of Figure 32 A～32C.When being judged to be A[M] during less than N, flow sequence withdraws from this flow process.

Like this, obtain the text point of the normal solution text strings that the literal of each text point of corresponding normal solution text strings with answer checks, according to its result, extract the text point of different words, scarce word and unnecessary word, be used as the wrong literal in the corresponding normal solution text strings of determined answer.

Extract the text point of checking by step 608～636 of Figure 32 A and 32B at the answer text strings of the corresponding normal solution text strings of answer.Extract wrong text point in the corresponding normal solution text strings of answer by step 640～664 of Figure 32 C.

Figure 33 A～33C is illustrated in the process flow diagram of the processing of the wrong text point of being carried out by the study assistance system 200 (mistake text point extraction unit 256) of Figure 12 in the step 546 of Figure 31 B, be used for extracting the answer text strings.

With reference to Figure 33 A, step 602～606 are identical with the step of Figure 32 A.

In step 708, the Position Number n that mistake text point extraction unit 256 is judged the literal in the answer text strings whether greater than length (literal number) N of answer text strings, i.e. the be through with extraction of wrong literal of whole literal in the text strings that whether checked on one's answers.Text point in being judged to be answer text strings numbering n is during greater than the length N of answer text strings, and flow sequence enters the step 740 of Figure 33 C.

When text point when be judged to be the answer text strings in step 708 in numbering n is not more than the length N of answer text strings, in step 712 wrong text point extraction unit 256 judge the corresponding normal solution text strings of answers check beginning text point numbering x whether greater than length (literal number) M of the corresponding normal solution text strings of answer, i.e. the comparison of the whole literal in the corresponding normal solution text strings of checking on one's answers that whether has been through with.When be judged to be the corresponding normal solution text strings of answer check beginning text point numbering x greater than length (literal number) M of the corresponding normal solution text strings of answer the time, flow sequence enters step 732.

When the checking beginning text point numbering x and be not more than length (literal number) M of the corresponding normal solution text strings of answer of the corresponding normal solution text strings of answer in step 712, in step 716, the Position Number m that mistake text point extraction unit 256 is judged the literal in the corresponding normal solution text strings of answers whether greater than length (literal number) M of the corresponding normal solution text strings of answer, promptly whether checked on one's answers whole literal in the corresponding normal solution text strings be through with wrong literal relatively or extract.The Position Number m of the literal in being judged to be the corresponding normal solution text strings of answer is during greater than length (literal number) M of the corresponding normal solution text strings of answer, and flow sequence enters step 734.When the Position Number m of the literal when be judged to be the corresponding normal solution text strings of answer in step 716 in was not more than length (literal number) M of the corresponding normal solution text strings of answer, flow sequence entered the step 720 of Figure 33 B.

With reference to Figure 33 B, in step 720, mistake text point extraction unit 256 judges whether the n literal of answer text strings is consistent with the m literal of the corresponding normal solution text strings of answer.When the n literal that is judged to be the answer text strings is consistent with the m literal of the corresponding normal solution text strings of answer, in step 724, mistake text point extraction unit 256 with the corresponding normal solution text strings of answer check beginning text point numbering add 1 and be set be x=m+1, the Position Number A[n of the literal of the corresponding normal solution text strings of answer that will be consistent with the n literal of answer text strings]=m stores the temporary storage area into.In step 726, mistake text point extraction unit 256 adds 1 and be set and be n=n+1 with the Position Number of the literal of answer text strings, check beginning text point numbering with what the Position Number m of the literal of the corresponding normal solution text strings of answer was set at the corresponding normal solution text strings of answer, promptly be set at m=x.Then, flow sequence is returned step 708.

When in step 712, be judged to be the corresponding normal solution text strings of answer check beginning text point numbering x greater than length (literal number) M of the corresponding normal solution text strings of answer the time, because there is not the literal in the corresponding normal solution text strings of answer corresponding with the n literal of answer text strings, so in step 732, there is the situation that lacks word in mistake text point extraction unit 256 in order to be illustrated in the answer text strings, set A[n as text point at the corresponding normal solution text strings of answer of the literal of the Position Number n of the literal of answer text strings]=-1 and store, the Position Number of the literal of answer text strings is added 1 and be set and be n=n+1.Then, flow sequence is returned step 708.

The Position Number m of the literal when be judged to be the corresponding normal solution text strings of answer in step 716 in is during greater than length (literal number) M of the corresponding normal solution text strings of answer, because there is not the literal of the answer text strings will corresponding normal solution text strings checked with answer, so in step 734, mistake text point extraction unit 256 is in order to be illustrated in the situation that has unnecessary word in the answer text strings, set A[n as text point at the corresponding normal solution text strings of answer of the literal of the Position Number n of the literal of answer text strings]=0 and store, the Position Number of the literal of answer text strings is added 1 and be set and be n=n+1, check beginning text point numbering with what the Position Number m of the literal of the corresponding normal solution text strings of answer was set at the corresponding normal solution text strings of answer, promptly be set at m=x.Then, flow sequence is returned step 708.

When the m literal of the corresponding normal solution text strings with answer of n literal that is judged to be the answer text strings in step 720 is inconsistent, in step 736, mistake text point extraction unit 256 adds 1 and be set and be m=m+1 with the Position Number of the literal in the corresponding normal solution text strings of answer.Then, flow sequence is returned step 716.

With reference to Figure 33 C, the Position Number n of the literal when be judged to be the answer text strings in step 708 in is during greater than the length N of answer text strings, in step 740, mistake text point extraction unit 256 is set at the text point numbering n of answer text strings the length of answer text strings, promptly be set at n=N, the gap d of the text point numbering of the corresponding normal solution text strings of answer that the text point of answer text strings numbering n and Ying Yuqi are checked is set at initial value d=0.

In step 742, mistake text point extraction unit 256 judges whether the text point numbering n of answer text strings is 0, i.e. n=0 whether.When being judged to be when not being n=0, in step 746, mistake text point extraction unit 256 is judged as the text point A[n at the corresponding normal solution text strings of answer of the literal of the Position Number n of the answer text strings of being stored] whether be-1.

When in step 746, being judged to be A[n] when being-1, in step 748, mistake text point extraction unit 256 is appended the text point n of answer text strings and is stored in unnecessary word location tabulation.Then, flow sequence enters step 768.

In step 768, mistake text point extraction unit 256 text points from the answer text strings deduct 1 and make it become n=n-1.Then, flow sequence enters step 742.

When in step 746, being judged to be A[n] when not being-1, mistake text point extraction unit 256 is judged the text point A[n at the corresponding normal solution text strings of answer of the n literal of the answer text strings of having stored] whether be 0.Be judged to be A[n] be under 0 the situation, in step 754, mistake text point extraction unit 256 judges whether gap d are 0, i.e. d=0 whether.When be judged to be checking of the corresponding normal solution text strings of answer text strings with answer do not have gap between the text point, when being d=0, in step 756, mistake text point extraction unit 256 is appended the text point n of answer text strings and is stored in different word location tabulations.Then, flow sequence enters step 768.

When in step 754, be judged to be the corresponding normal solution text strings of answer text strings and answer check text point have gap, when being d ≠ 0, mistake text point extraction unit 256 is appended text point n in unnecessary word location tabulation, d is set at d=d-1 with gap.Then, flow sequence enters step 768.

When in step 750, being judged to be A[n] when being not 0, mistake text point extraction unit 256 judges that in step 760 gap d is whether greater than the consistent text point numbering A[n at the corresponding normal solution text strings of answer of the n literal of answer text strings] and the text point n of answer text strings between difference absolute value (| A[n]-n|).When being judged to be d greater than A[n] and during the absolute value of the difference of n, because with the size of the gap d that checks text point of the corresponding normal solution text strings of the n literal of answer text strings answer consistent or that check the gap size of checking text point less than the position after leaning on than text point n, so in step 762, mistake text point extraction unit 256 is appended text point n in scarce word location tabulation.In step 764, mistake text point extraction unit 256 is set at d=|A[n with gap d]-n|.Then, flow sequence enters step 768.

Be not more than A[n when in step 760, being judged to be gap d] and n between the absolute value of difference the time, flow sequence enters step 764.

When in step 742, being judged to be n=0, mistake text point extraction unit 256 is judged the text point A[N of the corresponding normal solution text strings of answer that causes or check with the last text point N-of answer text strings in step 772] whether less than the length M of the corresponding normal solution text strings of answer (A[N]＜M).When being judged to be A[N] during less than M, in step 774, mistake text point extraction unit 256 is appended last text point N in scarce word location tabulation.Then, flow sequence withdraws from the flow process of Figure 33 A～33C.When being judged to be A[N] when being not less than M, flow sequence withdraws from this flow process.

Like this, obtain the text point of the corresponding normal solution text strings of checking with the literal of each text point of answer text strings of normal solution, the text point that according to its result, extract different words, lacks word and unnecessary word is used as the wrong literal in the answer text strings.

Extract the text point of checking by step 708～736 of Figure 33 A and 33B at the corresponding normal solution text strings of the answer of answer text strings.By step 740～764 of Figure 33 C, extract the wrong text point in the answer text strings.

Figure 34 A～34D illustrate carry out by the study assistance system 200 of Figure 23, be used to judge that normal solution/non-normal solution and the decision frame that checks on one's answers distributes the processing flow chart of the corresponding normal solution text strings of answer.

With reference to Figure 34 A, step 502 is identical with the step of Figure 31 A.In step 804, the text strings similarity is calculated portion 248 the identiflication number h of answer textbox is set at initial value h=1, will distribute the number A of determined answer textbox to be set at initial value A=0.In step 806, the text strings similarity is calculated portion 248 and is judged whether identiflication number h (h＞B), are whether the answer frame of identiflication number h does not exist greater than the number B of answer frame.When being judged to be identiflication number h greater than the number B of answer frame, when promptly h answer frame does not exist, flow sequence enters the step 832 of Figure 34 B.

When in step 806, be judged to be identiflication number h be not more than the number B of answer frame, when promptly having the answer frame of identiflication number h, in step 808, the text strings similarity is calculated portion 248 the identiflication number i of normal solution text strings is set at initial value i=1, with identiflication number is the maximum text strings similarity S[h of the answer frame of h] be set at initial value S[h]=-1, with the identiflication number AC[h of the normal solution text strings with maximum text strings similarity of the answer frame of identiflication number h] be set at initial value AC[h]=0.

In step 810, the text strings similarity is calculated portion 248 and is judged whether numbering i (i＞N), are whether the normal solution text strings of identiflication number i does not exist greater than the number N of normal solution text strings.When being judged to be numbering i greater than the number of normal solution text strings, when promptly i normal solution text strings does not exist, flow sequence enters step 822.

Be not more than the number of normal solution text strings, when promptly having i normal solution text strings, calculate portion 248 in step 812 Chinese character string similarity and calculate the answer text strings that is written in the h answer frame and the text strings similarity S[h between the i normal solution text strings CSi when in step 810, being judged to be numbering i] [i].

In step 814, the text strings similarity is calculated portion 248 and is judged text strings similarity S[h] whether [i] greater than the maximum text strings similarity S[h at h answer frame].When being judged to be S[h] [i] greater than S[h] time, in step 816, the text strings similarity is calculated portion 248 with S[h] value be updated to S[h]=S[h] [i], identiflication number AC[h with the normal solution text strings with maximum text strings similarity of the answer frame of identiflication number h] value be set at AC[h]=i, with both values (S[h], AC[h]) corresponding with answer frame identiflication number h store in the temporary storage area.In step 818, the text strings similarity is calculated portion 248 and is set at identiflication number i=i+1.

In step 814, when being judged to be S[h] [i] be not more than S[h] time, flow sequence enters step 818.Then, flow sequence is returned step 810.

In step 822, the maximum text strings similarity S[h that the corresponding normal solution text strings of answer determination section 9 is judged at h answer frame] whether equal 1.When being judged to be S[h] when being not equal to 1, flow sequence enters step 830.When being judged to be S[h] when equaling 1, in step 824, the corresponding normal solution text strings of answer determination section 9 is set the corresponding normal solution text strings of answer of h answer frame as i normal solution text strings.

In step 826, the corresponding normal solution text strings of answer determination section 9 appends identiflication number i to the temporary storage area of determining the normal solution text strings.In step 828, the corresponding normal solution text strings of answer determination section 9 appends identiflication number h and stores to the temporary storage area of determining the answer frame, will distribute the number A of determined answer frame to be set at A=A+1.Then, flow sequence enters step 830.

In step 830, the text strings similarity is calculated portion 248 the identiflication number h of answer textbox is set at h=h+1.Then, flow sequence enters step 806.

With reference to Figure 34 B, in step 832, the corresponding normal solution text strings of answer determination section 250 judges whether the number A that distributes determined definite answer textbox equals the number B (A=B) of answer textbox.When being judged to be A when being not equal to B, in step 834, the corresponding normal solution text strings of answer determination section 250 is set at initial value h=1 with the identiflication number h of answer frame, maximum text strings similarity S is set for the maximum similarity S[h in the answer frame of identiflication number h as initial value], will distribute the identiflication number k of determined definite answer frame to be set at k=h.

In step 838, the corresponding normal solution text strings of answer determination section 250 judges that identiflication number h are whether greater than the number B of answer frame (h＞B).When being judged to be h greater than B in step 836, flow sequence is returned step 832.

When being judged to be identiflication number h when being not more than the number B of answer textbox in step 836, the corresponding normal solution text strings of answer determination section 250 is judged in step 840 whether h is included in and is determined in the answer frame.Be included in when determining in the answer frame when being judged to be h, flow sequence enters step 836.

In step 836, the corresponding normal solution text strings of answer determination section 250 is set at h=h+1 with the identiflication number h of answer frame.Then, flow sequence enters step 838.

Be not contained in when determining in the answer frame when in step 840, being judged to be h, the identiflication number p of the answer frame that the corresponding normal solution text strings of answer determination section 250 will compare in step 842 is set at initial value p=h+1, will make AC[h] the normal solution text strings is that the answer frame k of the corresponding normal solution text strings of answer is set at k=h.

With reference to Figure 34 C, in step 844, the corresponding normal solution text strings of answer determination section 250 judges that identiflication number p are whether greater than the number B of answer frame.When being judged to be identiflication number p greater than the number B of answer frame, the corresponding normal solution text strings of answer determination section 250 appends identiflication number AC[h to the temporary storage area of determining the normal solution text strings in step 856] and store.In step 858, the corresponding normal solution text strings of answer determination section 250 appends k and stores to the temporary storage area of determining the answer frame, and the number A that determines the answer frame is set at A=A+1.Then, flow sequence is returned the step 836 of Figure 35 B.

When being judged to be identiflication number p when being not more than the number B of answer frame in step 844, the corresponding normal solution text strings of answer determination section 250 judges in step 846 whether identiflication number p is included in the numbering of determining answer frame (temporary storage area).When being judged to be identiflication number p when being included in the numbering of determining the answer frame, flow sequence enters step 882.

In step 882, p is set at p=p+1 with identiflication number.Then, flow sequence enters step 844.

When being judged to be identiflication number p when not being contained in the numbering of determining the answer frame in step 846, the corresponding normal solution text strings of answer determination section 250 is judged the identiflication number AC[h of the normal solution text strings with maximum text strings similarity of h answer frame in step 848] whether equal the identiflication number AC[p of the normal solution text strings with maximum text strings similarity of p answer frame] (AC[h]=AC[p]).When being judged to be AC[h] be not equal to AC[p] time, flow sequence enters step 882.

When in step 848, being judged to be AC[h] equal AC[p] time, the corresponding normal solution text strings of answer determination section 250 judges that in step 850 whether maximum text strings similarity S is less than the maximum text strings similarity S[p at p answer frame] (S＜S[p]).Be not less than S[p when being judged to be S] time, flow sequence enters step 882.

When in step 580, being judged to be S less than S[p] time, the corresponding normal solution text strings of answer determination section 250 is set at S=S[p with maximum text strings similarity S in step 852], will make identiflication number AC[h] the normal solution text strings be that the answer frame k of the corresponding normal solution text strings of answer is set at k=p.

With reference to Figure 34 D, the corresponding normal solution text strings of answer determination section 250 is set at i=1 with the identiflication number i of normal solution text strings in step 862, the maximum text strings similarity S[h except determining the normal solution text strings with h answer frame] be set at initial value S[h]=-1, with the identiflication number AC[h of the normal solution text strings except determining the normal solution text strings of h answer frame with maximum text strings similarity] be set at initial value AC[h]=0.

In step 864, whether the identiflication number i that the corresponding normal solution text strings of answer determination section 250 is judged the normal solution text strings is greater than the number N of normal solution text strings.As the identiflication number i that is judged to be the normal solution text strings during greater than the number N of normal solution text strings, flow sequence is returned the step 842 of Figure 34 A.

When the identiflication number i that is judged to be the normal solution text strings in step 864 is not more than the number N of normal solution text strings, the corresponding normal solution text strings of answer determination section 250 judges whether the identiflication number i of normal solution text strings is included in (temporary storage area) numbering of determining the normal solution text strings in step 866.When the identiflication number i that is judged to be the normal solution text strings was included in definite normal solution text strings, flow sequence entered step 872.

In step 872, the corresponding normal solution text strings of answer determination section 250 is set at i=i+1 with the identiflication number i of normal solution text strings.Then, flow sequence is returned step 864.

When the identiflication number i that is judged to be the normal solution text strings in step 866 is not included in when determining in the normal solution text strings, the corresponding normal solution text strings of answer determination section 250 is judged the answer text strings that is imported in the h answer frame and the text strings similarity S[h between the i normal solution text strings in step 868] whether [i] greater than S[h].When being judged to be text strings similarity S[h] [i] be not more than S[h] time, flow sequence enters step 872.

When in step 868, being judged to be text strings similarity S[h] [i] greater than S[h] time, the corresponding normal solution text strings of answer determination section 250 in step 870 with text strings similarity S[h] be set at S[h]=S[h] [i], with the identiflication number AC[h of normal solution text strings] be set at AC[h]=i.Then, flow sequence enters step 872.

Referring again to Figure 34 B, in step 832, be judged to be under the situation that A equals B, finish whole answer textboxs are distributed the corresponding normal solution text strings of answer, flow sequence withdraws from the flow process of Figure 34 A～34D.

The number of textbox of checking on one's answers is illustrated with the situation that the number of normal solution text strings equates, even but also can similarly carry out the distribution of normal solution text strings at the number of normal solution text strings during more than the number of answer textbox.

Figure 35 illustrates the part of other process flow diagrams of being judged the processing of normal solution/non-normal solution by the study assistance system 200 of Figure 25 being used to of carrying out.The process flow diagram of this part is replaceable to be the step 502 of Figure 31 A, the step 602 of Figure 32 A and the step 602 of Figure 33 A.

With reference to Figure 35, in step 1502, the answer text strings is read portion 242 according to the indication (order) that normal solution/non-normal solution is judged of judging instruction unit 212 from correctness, from answer text strings storage part 252, take out the handwriting information in the answer frame, and offer literal and intercept regional candidate's generating unit 244.In step 1504, literal intercepts regional candidate's generating unit 244 and analyzes handwriting information, generate literal and intercept regional candidate (candidate region of literal intercepting), each literal is intercepted regional candidate give each identifying information (a, b, c ...), store literal then into and intercept regional candidate's storage part 254, and this literal is intercepted regional candidate offer indivedual character recognition portion 246 with this identifying information or identiflication number.

In step 1506, indivedual character recognition portion 246 are set at initial value n=1 with identifying information or identiflication number n.In step 1508, indivedual character recognition portion 246 judge whether each group of being made up of a succession of character area candidate or each path have the regional candidate of intercepting of identiflication number n (at this moment, initial n=1).When being judged to be (initial n=1) intercepting that has n during regional candidate, in step 1510, indivedual character recognition portion 246 usefulness known form are carried out literal identification to the person's handwriting that n intercepts in the regional candidate, and generate identification candidate character and identification mark.In step 1512, indivedual character recognition portion 246 are set at n=n+1 with identiflication number n.Then, flow sequence is returned step 1508.

When in step 1508, being judged to be (initial n=1) intercepting that do not have n during regional candidate, in step 1522, the intercepting zone candidate that indivedual character recognition portion 246 will generate, offer maximum likelihood text strings determination section 247 at each regional candidate's identification candidate character and identification mark.Maximum likelihood text strings determination section 247 will intercept regional candidate, intercept regional candidate's identification candidate character and identification mark at each, compare and check with the normal solution text strings of storage in normal solution text strings storage part 258, decision maximum likelihood text strings, and the maximum likelihood text strings that is determined is offered the text strings similarity calculate portion 248.Then, flow sequence enters the step 504 of Figure 31 A, the step 604 of Figure 32 A, the step 604 of Figure 33 A or the step 602 of Figure 34 A.

Figure 36 A～36E illustrate carry out by maximum likelihood text strings determination section 247, other detailed process flow diagrams of step 1522 among Figure 35.At this moment, be used to determine that the processing of maximum likelihood text strings comprises scarce word of detection and unnecessary literal.

In step 1602,247 couples of node serial number n of maximum likelihood text strings determination section set initial value n=1.In step 1604, maximum likelihood text strings determination section 247 judges in the intercepting zone candidate's who has generated node whether have the n node.When being judged to be the n node when not existing, flow sequence enters the step 1662 of Figure 36 E.When being judged to be the n node when existing in step 1604, in step 1606, maximum likelihood text strings determination section 247 judges whether these n nodes are initial intercepting zone candidate.

When in step 1606, being judged to be the n node when not being initial intercepting zone candidate, maximum likelihood text strings determination section 247 is in step 1610, the link number u that will connect before the n node is set at initial value u=1, and maximum text strings verify maxS is set at initial value maxS=0.

In step 1612, maximum likelihood text strings determination section 247 judges whether link number u is below the U.Here, U is illustrated in the total quantity of the link (node) that connects before the n node.When being judged to be link number u when not being (bigger than U) below the U, in step 1616, maximum likelihood text strings determination section 247 is set at n=n+1 with node serial number n.Then, flow sequence is returned step 1604.

When in step 1606, being judged to be the n node when being initial intercepting zone, in step 1618, maximum likelihood text strings determination section 247 is set at initial value X=1 with the position X that checks literal in the normal solution text strings, the text point Y of the n node in the maximum likelihood text strings is set at initial value Y=1, the text strings verify maxS of maximum is set at initial value maxS=0.Then in step 1609, maximum likelihood text strings determination section 247 is set at " nothing " with the node serial number t of last selection node.Then, flow sequence enters step 1621.

When be judged to be link number u in step 1612 is U when following, and in step 1613, maximum likelihood text strings determination section 247 is set at last selection node serial number t the node serial number of the u node that connected before the n node.Then, in step 1621, maximum likelihood text strings determination section 247 adds 1 with the text point of checking that text point X is set at the X=t node of checking in the normal solution text strings, and the text point that text point Y is set at the Y=t node is added 1.Then, flow sequence enters the step 1636 of Figure 36 B.

When reference Figure 36 B, in step 1636, maximum likelihood text strings determination section 247 is judged the length L whether text point X surpasses the normal solution text strings of checking in the normal solution text strings.When being judged to be text point X greater than length L, flow sequence enters the step 1644 of Figure 36 D.

When being judged to be X when being not more than L in step 1636, in step 1682, maximum likelihood text strings determination section 247 will be checked text point q and be set at initial value q=X.In step 1684, judge check text point q whether be X and a and (X+a) more than (q 〉=X+a).Here, a is the numerical value of regulation, but is the scarce word that goes out of continuous detecting and the number of unnecessary word.On the other hand, when be judged to be q X and a when above, in step 1686, maximum likelihood text strings determination section 247 is set at u=u+1 with link number u.Then, flow sequence is returned the step 1616 of Figure 36 A.

When in step 1684, be judged to be q be not X and a when above, maximum likelihood text strings determination section 247 judges that in step 1685 whether q is greater than L.When being judged to be q greater than the length L of normal solution text strings, flow sequence enters step 1686.On the other hand, when being judged to be q when being not more than length L, flow sequence enters the step 1687 of Figure 36 C.

With reference to Figure 36 C, in step 1687, maximum likelihood text strings determination section 247 judges whether the q literal of normal solution text strings is present in the literal candidate in the node.When being judged to be the q literal when not being present in the literal candidate in the node, in step 1645, the candidate character of maximum likelihood text strings determination section 247 setting literal inconsistent values of verify MS=or minimum, largest score is for selecting literal MJ, checking text point MI=X-1.On the other hand, when being judged to be the q literal when being present in the candidate in the node, in step 1647, maximum likelihood text strings determination section 247 is set the literal mark+additive value (homogeneity value) of checking text point q in the literal mark MS=node, the q literal of normal solution text strings is set at selection literal MJ, sets and check text point MI=q.Then, flow sequence enters step 1649.

In step 1649, maximum likelihood text strings determination section 247 is set at text strings verify=(MRS of t node * (Y-1)+MS)/Y.In step 1651, judge that text strings verify MRS is whether greater than the text strings verify maxS of maximum.When being judged to be text strings verify MRS when being not more than largest score maxS, flow sequence enters step 1698, then, returns the step 1684 of Figure 36 B.

When being judged to be text strings verify MRS greater than the text strings mark maxS of maximum in step 1651, in step 1652, maximum likelihood text strings determination section 247 is set at maximum text strings mark maxS=MRS.In step 1696, maximum likelihood text strings determination section 247 will be checked text point MI, selection literal MJ, maximum text strings verify maxS, text point Y, store explicitly as the t node and the n node of last selection node.Then, flow sequence enters step 1698.

With reference to Figure 36 D, in the step 1644 after the "Yes" branch of the step 1636 of Figure 36 B, maximum likelihood text strings determination section 247 is set the candidate character of the literal inconsistent values of verify MS=or minimum, largest score for selecting literal MJ.

In step 1649, maximum likelihood text strings determination section 247 setting text strings verifies=(MRS of t node * (Y-1)+MS)/Y.In step 1651, maximum likelihood text strings determination section 247 judges that whether text strings verify MRS is greater than maximum text strings verify maxS.When being judged to be text strings verify MRS when being not more than largest score maxS, flow sequence is returned the step 1612 of Figure 36 A.

When being judged to be text strings verify MRS greater than maximum text strings mark maxS in step 1651, in step 1652, maximum likelihood text strings determination section 247 is set maximum text strings mark maxS=MRS.In step 1708, maximum likelihood text strings determination section 247 will be checked text point X, select literal MJ, maximum text strings verify maxMRS, text point Y and and store explicitly as the t kind of node and the n node of last selection node.In step 1712, maximum likelihood text strings determination section 247 is set at u=u+1.Then, flow sequence is returned the step 1612 of Figure 36 A.

With reference to Figure 36 E, in the step 1662 after the "No" branch of the step 1604 of Figure 36 A, the maximum likelihood text strings determination section 247 serial number r with the end node are set at initial value r=1, and the maximum text strings verify MAX of end node is set at initial value MAX=0.

In step 1664, whether 247 judgements of maximum likelihood text strings determination section exist as the r node that literal intercepts regional end node.When being judged to be it when not existing, in step 1676, maximum likelihood text strings determination section 247 connects the selection literal, decision maximum likelihood text strings from determined end node L ast to the node searching of front.

When being judged to be r node as the end node when existing in step 1664, in step 1666, maximum likelihood text strings determination section 247 will be set at C as the r node serial number of end node.In step 1668, maximum likelihood text strings determination section 247 judges that whether the text strings verify MRS of C node is less than maximum text strings verify MAX.When being judged to be MRS when being not less than MAX, in step 1672, maximum likelihood text strings determination section 247 is set at r=r+1.Then, flow sequence is returned step 1664.On the other hand, when being judged to be MRS less than MAX, in step 1670, maximum likelihood text strings determination section 247 is set MAX=MRS, L ast=C.Then, flow sequence enters step 1672.

The above only typical case enumerated of Shuo Ming embodiment, those skilled in the art as can be known, also can make up the textural element of this each embodiment or various distortion and variation, so long as those skilled in the art just as can be known, do not break away from the invention scope that the principle of the invention and claim scope are put down in writing, just can carry out the various distortion of above-mentioned embodiment.

The disclosed study assistance system of this instructions and accompanying drawing has: storage unit, and it comprises answer text strings storage area and normal solution text strings group storage area; Answer text strings administrative unit, it will store in the answer text strings storage area an answer text strings of an answer frame input; Answer text strings sensing element, it reads the answer text strings from answer text strings storage area; The text strings similarity is calculated the unit, its calculate in the answer text strings of having read and the normal solution text strings group storage area at the text strings similarity between a plurality of normal solution text strings of an above-mentioned answer frame; The corresponding normal solution text strings decision of answer unit, it is according to each text strings similarity between answer text strings and the above-mentioned a plurality of normal solution text strings, from a plurality of answer text strings, select the more than one normal solution text strings corresponding, its decision is the corresponding normal solution text strings of more than one answer with the answer text strings; And the correctness identifying unit, it is checked the answer text strings with corresponding normal solution text strings, judge the correctness of answer text strings.According to such structure, by selecting the normal solution text strings corresponding to determine to be the such action of the corresponding normal solution text strings of answer with the answer text strings, not only can show the correctness result of determination, can also know the normal solution text strings corresponding, thereby can expect following effect: make the learner easily grasp the corresponding normal solution text strings of importing with the learner of answer with the answer of having imported.

In addition, the disclosed study assistance system of this instructions and accompanying drawing except said structure, also has: the answer input block, by the user answer frame is imported the answer text strings; Correctness is judged indicating member, imports correctness by the user and judges indication; Answer text strings display unit, it will be shown in the answer frame by the answer text strings of user's input; Normal solution text strings display unit, it will be shown with the form different with the normal solution text strings beyond the corresponding normal solution text strings of answer by the corresponding normal solution text strings of answer of the corresponding normal solution text strings decision of answer unit decision; And correctness is judged display unit, the correctness result of determination that its demonstration receives from the correctness identifying unit.According to such structure, by the corresponding normal solution text strings of answer is shown such action with the form different with the normal solution text strings beyond the corresponding normal solution text strings of answer, can be shown as the normal solution text strings corresponding with the answer of being imported come into plain view, thereby can expect following effect: make the learner just can grasp the corresponding normal solution text strings of being imported with the learner of answer moment.

In addition, the disclosed study assistance system of this instructions and accompanying drawing is in the combination in any of said structure, when in a plurality of answer text strings, when corresponding with the described answer text strings number with corresponding normal solution text strings of more than one answer that equates mutually of equal similarity was a plurality of, the decision of the corresponding normal solution text strings decision of answer unit in only will this more than one normal solution text strings was answer correspondence normal solution text strings.According to such structure, by a decision in the corresponding normal solution text strings of the more than one answer that equates mutually that only will have equal similarity is the such action of the corresponding normal solution text strings of answer, can in an answer frame, show the corresponding normal solution text strings of answer all the time, thereby can expect following effect: the corresponding normal solution text strings of answer that the learner can easily be known imported with the learner.

In addition, the disclosed study assistance system of this instructions and accompanying drawing, in the combination in any of said structure, also has wrong text point extraction unit, check on one's answers corresponding normal solution text strings and answer text strings of this mistake text point extraction unit compares, extract the different words in the corresponding normal solution text strings of answer at the answer text strings, the position that lacks word or unnecessary word, normal solution text strings display unit is according to the different words at the answer text strings in the corresponding normal solution text strings of answer, the position that lacks word or unnecessary word shows the different words at the answer text strings in the corresponding normal solution text strings of answer, the position that lacks word or unnecessary word.According to such structure, by showing the such action in position of the different words at the answer text strings, scarce word or unnecessary word in the corresponding normal solution text strings of answer, can under the wrong situation of the part of answer text strings, this wrong position be shown, thereby can expect following effect: the wrong part in the corresponding normal solution text strings of the answer that the learner is easily grasped imported with the learner.

In addition, the disclosed study assistance system of this instructions and accompanying drawing, in the combination in any of said structure, mistake text point extraction unit also extracts the position of the different words at the corresponding normal solution text strings of described answer in the answer text strings, scarce word or unnecessary word, answer text strings display unit shows the position of the different words at the corresponding normal solution text strings of answer in the answer text strings, scarce word or unnecessary word according to the position of the different words at the corresponding normal solution text strings of answer in the answer text strings, scarce word or unnecessary word.According to such structure, by showing the such action in position of the different words in the answer text strings, scarce word or unnecessary word at the corresponding normal solution text strings of answer, can under the wrong situation of the part of answer text strings, this wrong position be shown, thereby can expect following effect: make the learner easily grasp wrong part in the answer text strings of being imported with the learner.

In addition, the disclosed study assistance system of this instructions and accompanying drawing, in the combination in any of said structure, an answer frame is one of a plurality of answer frames that can be corresponding with a plurality of normal solution text strings, the text strings similarity is calculated the unit and is calculated a plurality of answer frames answer text strings and the text strings similarity of a plurality of normal solution text strings between separately separately, the corresponding normal solution text strings decision of answer unit is according to a plurality of answer frames answer text strings and the text strings similarity of a plurality of normal solution text strings between separately separately, a plurality of answer frames are selected to distribute normal solution text strings in a plurality of normal solution text strings uniquely, its decision is the corresponding normal solution text strings of answer.According to such structure, by a plurality of answer frames being selected to distribute uniquely the normal solution text strings in a plurality of normal solution text strings, its decision is the such action of the corresponding normal solution text strings of answer, under situation, not only can show the correctness result of determination with the different a plurality of normal solution text strings of order, can also know the normal solution text strings corresponding, thereby can expect following effect: the corresponding normal solution text strings of answer that under situation, the learner is easily grasped to be imported with the learner with the different a plurality of normal solution text strings of order with the answer of having imported.

In addition, the disclosed study assistance system of this instructions and accompanying drawing, in the combination in any of said structure, storage unit comprises display position information stores zone, this display position information stores zone storage is used for showing at a plurality of answer frames the information of the display position of the corresponding normal solution text strings of answer, storage unit also comprises normal solution text strings display position and obtains the unit, this normal solution text strings display position is obtained the unit and is obtained information at the display position of a plurality of answer frames, the corresponding normal solution text strings of answer display unit is at the display position at a plurality of answer frames, shows the normal solution text strings of the correspondence in the normal solution text strings of described distribution.According to such structure, be assigned to the such action of normal solution text strings by demonstration at the correspondence in the distribution normal solution text strings of the display position of a plurality of answer frames, can be under situation with the different a plurality of normal solution text strings of order, show the corresponding normal solution text strings of answer on respectively at the display position of a plurality of answer frames, thereby can expect following effect: the corresponding normal solution text strings of answer that under situation, the learner is easily known to be imported with the learner with the different a plurality of normal solution text strings of order.

In addition, the disclosed study assistance system of this instructions and accompanying drawing, in the combination in any of said structure, storage unit also comprises literal and intercepts regional candidate memory regions and literal identification dictionaries store zone, answer text strings administrative unit also will store in the answer text strings storage area the handwriting information of an answer frame input, answer text strings sensing element is also read handwriting information from answer text strings storage area, this study assistance system also has: literal intercepts regional candidate's generation unit, and it is a plurality of regional candidate of literal unit with the handwriting information intercepting of reading; Indivedual word recognition units, it is according to literal identification dictinary information in the literal identification dictionaries store zone, and the handwriting information that literal is intercepted in a plurality of regional candidate that regional candidate's generation unit generated carries out literal identification respectively, generates a plurality of candidate characters; And maximum likelihood text strings decision unit, it will be checked at each regional candidate's that may make up among a plurality of regional candidates a plurality of candidate characters and the literal in the normal solution text strings in the normal solution text strings group storage area, decision maximum likelihood text strings, and offer the text strings similarity as the answer text strings and calculate the unit.According to such structure, even under the situation of handwriting input answer text strings, also not only show the correctness result of determination, can also know the normal solution text strings corresponding with the answer of being imported, except answering such results of learning with hand-written, the corresponding normal solution text strings of hand-written answer that can also expect that the learner is easily grasped and imported with the learner.

Claims (10)

1. learn assistance system for one kind, it is characterized in that this study assistance system has:

Storage unit, it comprises answer text strings storage area and normal solution text strings group storage area;

Answer text strings administrative unit, it will store in the above-mentioned answer text strings storage area an answer text strings of an answer frame input;

Answer text strings sensing element, it reads the answer text strings from above-mentioned answer text strings storage area;

The text strings similarity is calculated the unit, its calculate in above-mentioned answer text strings of reading and the above-mentioned normal solution text strings group storage area at the text strings similarity between a plurality of normal solution text strings of an above-mentioned answer frame;

The corresponding normal solution text strings decision of answer unit, it is according to each text strings similarity between above-mentioned answer text strings and the above-mentioned a plurality of normal solution text strings, from above-mentioned a plurality of answer text strings, select the more than one normal solution text strings corresponding, its decision is the corresponding normal solution text strings of more than one answer with above-mentioned answer text strings; And

The correctness identifying unit, it is checked above-mentioned answer text strings with above-mentioned corresponding normal solution text strings, judge the correctness of above-mentioned answer text strings.

2. study assistance system according to claim 1 is characterized in that, this study assistance system also has:

The answer input block, it is used for by the user an above-mentioned answer frame being imported above-mentioned answer text strings;

Correctness is judged indicating member, and it is used for importing correctness by the user and judges indication;

Answer text strings display unit, it will be shown in the above-mentioned answer frame by the above-mentioned answer text strings of above-mentioned user's input;

Normal solution text strings display unit, it will be shown with the form different with the normal solution text strings beyond the corresponding normal solution text strings of above-mentioned answer by the corresponding normal solution text strings of above-mentioned answer of the corresponding normal solution text strings decision of above-mentioned answer unit decision; And

Correctness is judged display unit, and it shows the correctness result of determination that receives from above-mentioned correctness identifying unit.

3. study assistance system according to claim 1 and 2 is characterized in that,

When in above-mentioned a plurality of answer text strings, when the number of the above-mentioned more than one normal solution corresponding normal solution text strings that mutually equate with equal similarity corresponding with above-mentioned answer text strings was a plurality of, the corresponding normal solution text strings decision of answer unit only was the corresponding normal solution text strings of above-mentioned answer with a decision in the above-mentioned more than one normal solution text strings.

4. study assistance system according to claim 1 and 2, it is characterized in that, this study assistance system also has wrong text point extraction unit, this mistake text point extraction unit compares corresponding normal solution text strings of above-mentioned answer and above-mentioned answer text strings, extract the position of the different words at above-mentioned answer text strings, scarce word or unnecessary word in the corresponding normal solution text strings of above-mentioned answer

Above-mentioned normal solution text strings display unit shows the position of the above-mentioned different words at above-mentioned answer text strings, scarce word or unnecessary word in the corresponding normal solution text strings of above-mentioned answer according to the position of the above-mentioned different words at above-mentioned answer text strings, scarce word or unnecessary word in the corresponding normal solution text strings of above-mentioned answer.

5. study assistance system according to claim 4 is characterized in that,

Above-mentioned wrong text point extraction unit also extracts the position of the different words at the corresponding normal solution text strings of above-mentioned answer in the above-mentioned answer text strings, scarce word or unnecessary word,

Above-mentioned answer text strings display unit shows the position of the above-mentioned different words at the corresponding normal solution text strings of above-mentioned answer in the above-mentioned answer text strings, scarce word or unnecessary word according to the position of the above-mentioned different words at above-mentioned corresponding normal solution answer text strings in the above-mentioned answer text strings, scarce word or unnecessary word.

6. study assistance system according to claim 1 and 2 is characterized in that,

An above-mentioned answer frame is one of a plurality of answer frames that can be corresponding with above-mentioned a plurality of normal solution text strings,

Above-mentioned text strings similarity is calculated the unit and is calculated above-mentioned a plurality of answer frame answer text strings and the text strings similarity of above-mentioned a plurality of normal solution text strings between separately separately,

The corresponding normal solution text strings decision of above-mentioned answer unit is according to above-mentioned a plurality of answer frames answer text strings and the above-mentioned text strings similarity of above-mentioned a plurality of normal solution text strings between separately separately, above-mentioned a plurality of answer frames are selected to distribute above-mentioned normal solution text strings in above-mentioned a plurality of normal solution text strings uniquely, its decision is the corresponding normal solution text strings of above-mentioned answer.

7. study assistance system according to claim 6 is characterized in that,

Said memory cells comprises display position information stores zone, and the storage of this display position information stores zone is used to show the information at the display position of the corresponding normal solution text strings of answer of above-mentioned a plurality of answer frames,

Above-mentioned study assistance system comprises that also normal solution text strings display position obtains the unit, and this normal solution text strings display position is obtained the unit and obtained information at the above-mentioned display position of above-mentioned a plurality of answer frames,

The corresponding normal solution text strings of above-mentioned answer display unit is at the above-mentioned display position at above-mentioned a plurality of answer frames, shows the normal solution text strings of the correspondence in the normal solution text strings of above-mentioned distribution.

8. study assistance system according to claim 1 and 2 is characterized in that,

Said memory cells also comprises literal and intercepts regional candidate memory regions and literal identification dictionaries store zone,

Above-mentioned answer text strings administrative unit also will store in the above-mentioned answer text strings storage area the handwriting information of above-mentioned answer frame input,

Above-mentioned answer text strings sensing element is also read handwriting information from above-mentioned answer text strings storage area,

Above-mentioned study assistance system also comprises:

Literal intercepts regional candidate's generation unit, and it is a plurality of regional candidate of literal unit with above-mentioned handwriting information intercepting of reading;

Indivedual word recognition units, it is according to the literal identification dictinary information in the above-mentioned literal identification dictionaries store zone, the handwriting information that above-mentioned literal is intercepted in a plurality of regional candidate that regional candidate's generation unit generated carries out literal identification respectively, generates a plurality of candidate characters; And

Maximum likelihood text strings decision unit, it will be checked at each regional candidate's that may make up among above-mentioned a plurality of regional candidates above-mentioned a plurality of candidate characters and the literal in the normal solution text strings in the above-mentioned normal solution text strings group storage area, decision maximum likelihood text strings, and it is offered above-mentioned text strings similarity as above-mentioned answer text strings calculate the unit.

9. program, it is used to have and comprises answer text strings storage area and the storage unit of normal solution text strings group storage area and the signal conditioning package of processor, this program is characterised in that this program can make above-mentioned signal conditioning package carry out following steps:

To store the step of above-mentioned answer text strings storage area to an answer text strings of an answer frame input into;

From above-mentioned answer text strings storage area, read the answer text strings reading step of answer text strings;

Calculate the step in above-mentioned answer text strings of reading and the above-mentioned normal solution text strings group storage area at the text strings similarity between a plurality of normal solution text strings of an above-mentioned answer frame;

Carry out the following step of handling: according to each text strings similarity between above-mentioned answer text strings and the above-mentioned a plurality of normal solution text strings, from above-mentioned a plurality of answer text strings, select the more than one normal solution text strings corresponding, and its decision is the corresponding normal solution text strings of more than one answer with above-mentioned answer text strings; And

Above-mentioned answer text strings and above-mentioned corresponding normal solution text strings are checked the step of the correctness of judging above-mentioned answer text strings.

10. learn assisted method for one kind, be used to have and comprise answer text strings storage area and the storage unit of normal solution text strings group storage area and the signal conditioning package of processor, this study assisted method is characterised in that this study assisted method may further comprise the steps:

To store the step of above-mentioned answer text strings storage area to an answer text strings of an answer frame input into;

From above-mentioned answer text strings storage area, read the answer text strings reading step of answer text strings;

Calculate the step in above-mentioned answer text strings of reading and the above-mentioned normal solution text strings group storage area at the text strings similarity between a plurality of normal solution text strings of an above-mentioned answer frame;

Carry out the following step of handling: according to each text strings similarity between above-mentioned answer text strings and the above-mentioned a plurality of normal solution text strings, from above-mentioned a plurality of answer text strings, select the more than one normal solution text strings corresponding, and its decision is the corresponding normal solution text strings of more than one answer with above-mentioned answer text strings; And

Above-mentioned answer text strings and above-mentioned corresponding normal solution text strings are checked the step of the correctness of judging above-mentioned answer text strings.

Images