TW407248B - Character recognition device - Google Patents

Abstract

To eliminate the necessity of inputting character attribute information in the case that a character can be correctly recognized only by character information, and to enable correct recognition of the character only by the character information even when the character attribute information is not inputted. A recognition means 22 executes character recognition based on written information inputted from an input means 21, stores the recognition result in a recognition result storing means 23 and displays the result on a recognition result display means 24. A recognition result converging means 28 converges a recognition result to be displayed on the means 24 out of plural recognition results stored in the means 23 based on character attribute information selectively inputted from displayed character attribute information by a selective input means 27 so as to recognize characters.

Description

V. Description of the invention (1) The technical field to which the invention belongs The present invention relates to the recognition of input information about the note information and the use of text; 1 Knowing Techniques for Recognizing Texts Knowing Techniques 1 FIG. 28 shows the situation of the report block document processing piece of a conventional text recognition device disclosed in, for example, Japanese Unexamined Patent Application Publication No. Hei 5-2030! Sorry, Xunzi reads the name device 2 $ Ml ^ ^ The reading device that takes the image of the front and back of the business card. 2 is to identify the front and back characters of the business card to determine whether the back text recognized by the identification device 2 is It is an English judging device, 4 is when the judging device 3 judges that it is an English mark; the ^ surface ^ characters and text are divided into address, i.e., phone number and other address data. 5 is a selection device that compares the registered text of the table text and the back text divided by the dividing device 4 and selects the marking of the back text that is consistent with the surface ^ characters, 6 is based on the selection of Pei Buzhi 5 A device for generating a side-note pseudonym corresponding to the mark on the back text. 7 is the memorization of the memory 0 associated with the side-note pseudonym generated by the generating device 6 and the surface text corresponding to the side-note pseudonym. ‘&Quot; 衮 The action will be explained next. First ’use the reading device 1 to read the front and back images of business cards ΙΒηΠΗΗ page 4

I. V. Description of the Invention (2) The recognition device 2 recognizes the read image as a text on the surface and the back judging device 3 judges whether the back text is an English mark. If it is displayed in English, the dividing device 4 will register the address information on the front and back: Erbai μ name / telephone number. Select the address registration of the front and back text divided by the option "J. Select the back text that is the same as the front text. The identification of the recognition result on the front page and the back text are inconsistent ^ Situation ' The selection device 5 performs the memorization with the recognition of the surface text, and selects the most consistent one for subsequent comparisons. Then, the generating device 6 generates a marginal pseudonym corresponding to the »», and the generated marginal pseudonym and the The mark of the surface text corresponding to this side-main kana is associated and stored in the memory device. Known technology 2 FIG. 29 is a conventional character recognition device disclosed in, for example, Japanese Patent Application Laid-Open No. 3_1 958 9 (optical character reading device). The block diagram of the structure. The text recognition setting also uses the text attribute information of the text to identify the recognition result of the text used part and the recognition result of the side-note kana part using the early word information to improve the accuracy. After the bill is recorded, the scanning part of the quantized pattern (from 2 sub-character patterns) output after the photoelectric conversion of the bill and the next β main kana text is recorded. 12 is a character recognition unit that recognizes the Chinese characters on the ledger by using the χ word list & based on the transmission from the scanning unit U = quantized pattern U.] 3 is a unit of sigma and katakana characters memorizing word units in advance Recognition 纟 (dictionary) Recognition result of Chinese characters and ^ characters output from the character recognition unit 12 ^

V. Description of the invention (3) The word recognition unit of each word composed of each character. The kanji and katakana characters output by the word recognition unit 13 correspond to each other. The result of the recognition of each word unit in the comparison result. Production department. In addition, ^ the control of outputting each recognition result as the final answer 丨 In addition, the control unit 14 controls the operations of the 011 and the word recognition unit 13. The operation of the sub-identification unit 12 will be described next. Fig. 30 is an explanatory diagram showing an example of inputting a bill.和 圮 录 料 = ί Chinese character area 15a of Kanji 15 Kana if3 Katakana text used for kana (hereinafter referred to as Katakana area 16a. First, scan the picture using the scanning unit 11 ', and the account will be recorded on the account book The katakana 16 of the name region 163 of the Kanji region 15a is converted into a quantity Γ self-character recognition unit 12. In the character recognition unit 12, each quantized pattern is used to identify the Kanji 15 and Katakana 16 in accordance with the tracing unit. = 31 is an explanatory diagram showing the recognition result of the character recognition unit 12 and shows the recognition result of katakana 16 '18 is the recognition of kanji 15-the result. Here, the recognition result 17 of katakana 16 is the first The i word and the second child are identified as "Dad" and "Fang", and there are two candidate words "<" and "Xi." Similarly, the recognition result of the interjection is 18, the second Each word is identified as "well"-a candidate ', but there are two candidate words "ban" and "saka" in the first word.

V. Description of the invention (4) In this way, each recognition of the character unit recognized by the character recognition unit 12 is given to the word recognition unit 13. The word recognition unit 13 firstly recognizes the literacy "5 of kanji kana 16" for each character based on the received results (generally, there are multiple candidate words for each character). Fruit Γ: Produce V "and :: eve" 2 words Candidates are generated from the Han recognition result 18, "Itai" and "Sakai". Secondly, the word recognition unit 13 checks whether the words stored and decorated as such exist in the recognition table, and selects the existing word candidates. The output of the recognition results in Volumes 2 and 13. Figure 2 of Tianchazao's Word Recognition Department Figure 2 is an explanatory diagram showing the contents of the word recognition unit 1 using the word recognition unit 13 as an example. The word list = recognition table and Chinese characters indicate lions .Word Recognition Unit 13 =: = 成 片 假 不在 此 词 units of the recognition table. A candidate for each outstanding word is the word comparison in kana kan 19, which compares the Kanji single kana word with the kana word. In this book For example, output katakana bar 19;: / "Sakai" and "Itaisu" in the mark column 20 in the pre-mark bar 20 and the Kanji department 14. Output as early word candidates to the control d control unit 1 to 4 For this word recognition section, 3 word candidates will be located in Figure 3 and Figure 2. Shows that the single gas n is the word candidate from one column of the recognition result. As the final answer, enter the same name in the recognition table of 单位 unit 襕 19 in the "Daddy;" Han of the same column "° _ In this example, in and The fake word is only "Sakai", and it is used as the final% = 20 order, which is output in response to the problem to be solved by the invention and the answer. The sum of the moments shown in the conventional technique 1 and the conventional technique 2 is identified by text attribute information. Page 7 L. Description of the invention (5) The character recognition device for the text is configured as described above, because the reading device i or the scanning unit 11 is used to read the text printed on the business card or the text written on the account book and its text attribute information. (Side note on kana), the problem is that in the case where only text information can be correctly identified, the text attribute information needs to be rotated. On the contrary, in the case where no text attribute information is input, it cannot be improved at all. ^ Also, because the text recognition devices of the cognition technology 1 and cognition technology 2 perform recognition processing on the text and sexual information, in the case that the text attribute information cannot be correctly identified, it may be able to correctly recognize There is also a problem that the recognition result cannot be obtained when the text information is used. Also, the text recognition device of the conventional technology 2 uses handwriting to take note of the text attribute information. At that time, because of the need to take both Chinese characters and its side-note kana, the amount of notes increased. The problem that the load of the note-writer increases. A and the 'text recognition device of the conventional technology 2 uses word information, and the words constituting the word and its side-note kana need to be noted. Without the note, the word information cannot be used. The problem of words cannot be recognized. Also, the text recognition devices of "Knowledge Technology 1 and 2" are all situations where the text recognition if on a line that should be recognized immediately after using a pen on a graphic board or other note text is due to the fact that Most of the text recognition devices do not have the problem of increasing the load of inputting text attribute information. The present invention was conceived in order to solve the above-mentioned problems, and an object of the present invention is to obtain a text recognition device, which can be easily used for text attribute information on a line without a physical keyboard. Also, the object of the present invention is to obtain a character recognition device.

Page 8 Enter the text if it is incorrect. —A kind of character recognition device, which uses the amount of calculation and can improve the text of the second article. By only attribute recognition information in the text recognition, the character recognition can be improved A'The purpose of the present invention is to obtain the attribute information to reduce the text The recognition process is based on the accuracy of the recognition. Another object of the present invention is to obtain a device for identifying the characters Debai Qiu eight times in the text of the word Sword and Reed. In the status, use the stand-by: thinking information to output the word information of the word candidate to the attribute information, which can be deleted. Reduce the number of word candidates. In addition, the purpose of the present invention is to obtain an early word and recognize a single temple, and to identify the device of the word * M, and when the condition of the word is λ as early as Ί, the result of the early word recognition is incorrect. Introduce text attribute information to improve word recognition results. Means for solving the problem After the recognition device for a character recognition device of the present invention recognizes a character based on note information input from an input device, the recognition result storage device memorizes the recognition result and displays it on the character attribute display device according to the use The text attribute information of the selected round is selected from the text attribute information of the text attribute display device selected by the turn-round device, and the recognition result is indented by the recognition result. The recognition result of the memory device is displayed on the recognition of the recognition result display device. The result is indented and the device is identified. The text recognition device of the present invention is provided with a text attribute information indentation device, and the text attribute information displayed on the text attribute display device is indented according to the text attribute information of the recognition result stored in the recognition result storage device. 0 The text recognition of the present invention Device, set the extraction memory to the recognition result record

407248 V. Description of the invention (7) Recalling the text attribute information of the recognition result of the device and then indenting the text attribute information displayed on the text attribute display device of the text attribute display device to select the best from a variety of text attribute information Text attribute information. According to the character recognition device of the present invention, based on the text attribute information selected and input from the text attribute information displayed on the text attribute display device using the selection input device, the recognition object indentation device is used to indent the text of the recognition object, and then the recognition device is used according to the After indenting the text information of the recognition object and the information recognition text obtained from the input device for inputting the note information of the text, the recognition result is displayed on the recognition result display device. The character recognition device of the present invention uses the vowel information of characters on the character attribute information. The word recognition device of the present invention memorizes the word association result of the word association result from the word association result memory device of the word association device self-recognition device, and displays the word association result of the word association result on the word association result display device. The input device, according to the text attribute information of the words constituting the word selected and input by the self-selection input device, the word association result displayed on the word association result display device is indented and recognized from the word association result memorized in the word association result storage device Text. The character recognition device of the present invention is provided with a word attribute information indentation device, and the text attribute information of the word association result memorized in the word association result memory device will be displayed in the text attribute information indentation of the text attribute display device. The character recognition device of the present invention is configured to extract the text attribute information of the words memorized in the word association result memory device of the word association result memory device and display the text attribute information.

Page 10 _ 407248 V. Description of the invention (8) The word attribute information indentation device shown in the text attribute information of the text attribute display device is used to select the best text attribute information from a variety of text attribute information. Embodiments of the Invention Embodiments of the present invention will be described below. Embodiment 1 FIG. 1 is a block diagram showing the structure of a character recognition device according to Embodiment 1 of the present invention. In FIG. 1, 21 is an input device for inputting note information, such as a graphic tablet or a scanner, 22 is a recognition device that outputs recognition results after recognizing text based on the note information output by the input device 21, and 23 is a memory for the recognition device 2 2 The output of the recognition result storage device of the recognition result, 2 4 is a recognition result display device that displays the recognition result in the recognition result storage device 2 3, and 2 5 is the storage of text attribute information such as pronunciation, stroke number, radical information, etc. The text attribute dictionary, 26 is a text attribute display device that displays the text attribute information. 27 is a selection input device that selects text attribute information to be displayed on the text attribute display device 26. 28 is selected based on the selection input device 27. The text attribute information is indented from the recognition result stored in the recognition result storage device 23 and is displayed in the recognition result indentation device of the recognition result displayed on the recognition result display device 24. The 29 series controls these input devices 21, recognition device 2, recognition result memory device 2, 3. recognition result display device 2, 4, character code 25, text attribute display device 26, selection input device ", and recognition ... fruit indentation device 2 8 The control device will be described next. Here, FIG. 2 shows the control device 29 of the present invention].

V. Description of the invention (9) Flow chart of the process of 1 word processing. The controller ST10 instructs the input device 21 to input the input device ^ in the first step = indicates the input pattern of the input device 21 and the input pattern. Then, enter the pattern of the Chinese text "中" in Diji, and in the example shown in the figure, input the recognition result of pattern 30 to identify the note box. The display of the device 24 is followed by step ST11, and the control device 90 controls the batch approval device 22 to output the acquired note data t "into the device 21 to identify the note information. The recognition device 22 inputs the recognition result from the other device 22 and transmits the recognition result to the recognition result storage device 23. After identifying the characters, D will store 23 in the recognition result storage device 23 of the recognition result of = Γ. At step STi2, those recognition results will be stored in the result storage device 23 with a high probability. FIG. 4 shows an example of the memory contents of the recognition result storage device 23 when 4 different addresses are received: 3 己 = recognition. This;,:; is the candidate text of the recognition result and its ranking. In the example shown in the figure, the four recognition results of "towel", "shen", "帀", and "medium" are ranked from the first to the third. 4 ranks are stored in the recognition result storage device 23. Next, in step S T1 3, the recognition result display device 2 4 reads the recognition result 6 and the first character in the recognition result memorized by the memory device 2 3 is displayed on the display frame 24a of the recognition result display device 24. A display example of the recognition result display device 24 is shown in FIG. 5. In addition, 32 in the figure is displayed in the display frame 24a of the recognition result display device 24, which has the first order of the fifth, the description of the invention (10) The text of the recognition result "towel, pre-attribute display device 26 in the step & Be text attribute information institute of the spiritual counseling but Xijin _ Shi, step M14 into order to choose the display of text attributes such as 卩% _ such as / and the sub attribute display device 26 king as shown in Figure 6. In this example, to fight against the thumb _ The text means Chinese 夂 发 立 俨 +, the examples are spelled with the standard notation (in English. Among all, t is not), it means that there are 5 kinds of vowels and 5 kinds of vowels in Chinese characters. In the case of Pinyin,-must contain one of u, e, and 0 ". In the case of the text, 15, in the selection input device 27, it is checked whether the recognition result is correct. If the text attribute information is not rotated, it is deemed to end the recognition processing of the one character. However, in the selection pendulum input device 27, the text attribute information is entered. The process branches to step ST16. In the selection wheel in this case, the recognition result with the first order is the text "scarf", = the recognition result is incorrect, use the selection input device. 27 Enter the text attribute of correct text, that is, the vowel information of Pinyin. FIG. 7 shows the text attribute information of the correct text selected and input by the selection input device 27. 33 in the figure is the text attribute information of the selected correct rounded text, which is indicated by a network line. In addition, in the example shown in the figure, the vowel information "0" with the pinyin "zhong" in the rounded text "中" is selected to be entered in the five kinds of vowel information 26a. Therefore, the processing from step ST1 5 to step ST1 6 'the recognition result indentation device 2 8 first searches the text attribute dictionary 25 and extracts the text attribute information of the recognition result character stored in the recognition result storage device 23' as a new recognition result memory. Here, FIG. 8 shows the contents of the recognition result storage device 23 after extracting the character attribute information of the recognition result. The 34 in the picture is the pinyin information extracted. As shown in FIG. 8, the recognition result indentation device 2 8

Page 13-40ma V. Description of the invention (11) What is memorized by the fruit memory device 23 (Pinyin mark m In addition, 3 'result text extracted from the text attribute information is fixed as one. Bu, most of the text of the Chinese text Pronunciation social fruit :::::: The device 28 is still in step ST16 'based on the extraction identification, "Kou Woyouzi's essay attribute 杳 · * η The content of the recognition result memory 23 after ϋ κ is selected by the device 27 The text of the entered text attribute information is £ 23 Α 4 is stored as a new recognition result in the recognition result memory. Γ 二 Γ The result table cannot be indented: Do n’t indent the result. 28 According to the text :, don't, ... The content map of the recognition result storage device 23 after the entry ϋ The recognition result. That is, the attribute information "." Selected by the input device 27 selected by the selection input device 27 is used to delete the result of the recognition result storage device indentation. The pinyin does not include the characters from the "○" order to the 3rd order shown in Figure 8, and only the pinyin i is displayed on the recognition result 35 to indicate the word "中" (zh〇ng) with the vowel information "〇". Pinyin Next, go to step ST17 to identify The result display device 24 regards the text indented in the result of the recognition result storage device 23 which is indented into the device 2 8 into a new 1 result, and displays it in the display of the recognition result. 24a. The display on the recognition result display device 24 = the new recognition result of 24a is shown in Figure 10. 36 in the figure is the recognition result, which is displayed on the recognition result 36 in the example shown in the figure. The correct text is "medium". In this way, the correct recognition result can be obtained. In addition, in this embodiment 1, the recognition result 32 displayed on the display frame 24a of the recognition result display device is set to be memorized only by the recognition result 3 4 device 23 The first text in the first order, but it can also make the guest 7 1 fixed waiting page 14 Fifth, the description of the invention (12) Select the text. 里 Ϊ In this embodiment 1 ′ 4, the result display device 24 displays recognition After the result, the text attribute information is displayed on the text attribute display device 26, but since the content displayed by the text attribute display device 26 is not related to the recognition result, as long as the text attribute information is selected and inputted by the selection input device 27, The text attribute information may be displayed. ≫ In the first embodiment, the inner attribute displayed by the text attribute display device 26 is again the vowel information of the Chinese text, but the pronunciation of the text, the number of strokes, and the radical information are used. Other text attribute information may also be used. Also in this embodiment 1, the recognition result indentation device 2 8 deletes the text result attribute information in the recognition result memory device 23 according to the text attribute information selected and input by using the selection input device 27. The recognition result is left intact in the recognition result memory device 23 without being deleted, and it may be used as the recognition result without selection. As described above, if this embodiment is used, it can be displayed and can be easily input. The text attribute information 'also selects the case where the text attribute information is input. Because the text attribute information is used to indent the recognition result, it is easy to enter the correct text in the case of cluttered notes and insufficient text recognition accuracy. 'If using the first embodiment, in the case where the character recognition result is wrong', it is unnecessary to input the note data of the same character. , Reduce the load of the Notes pen reporter 'particularly in the candidate character (typically about 1 0) not containing the correct recognition result of the case of words, you can no longer enter text notes, notes with a load reducing effect can be light pen of the reporter.

Page 15

Do not conclude = Example 1 'After recognizing the text', if only the misunderstanding of the text; if the brother only enters the text attribute information, there are recognition gaps f VI ·, ± force J to improve the D result of the text, can reduce the text attribute information Effect of input load. If the instructor uses this embodiment 1, because the vowel information (5 types) of Chinese characters is selected as the input attribute of the text, the portable information without a physical keyboard can be entered :: There can be reduced text Effect of input load. Embodiment 2 The following describes Embodiment 2 of the present invention using FIGS. 11 to 15. FIG. Η is a block diagram showing the structure of the character recognition device of Embodiment 2 by f. The same or equivalent parts of Embodiment i shown in FIG. In figure u, 40 is the recognition result of the text attribute information indented by the text attribute information of the text attribute display device 26 according to the recognition result memorized in s, ,, 'α fruit 5 and the memory device 2 3 after being recognized by the recognition device 22. Indentation device. In addition, the character recognition device of the second embodiment is different from the first embodiment in that the character attribute information indenting device 40 has the character attribute information. The operation will be described next. ★ Here 'FIG. 12 is a flowchart of the processing flow of one word in the control device 29 in the second embodiment. The control device 29 processes step ST20 to step ST23 and the example steps ST10 to step ST13, which processes the input of note information on the input device 21, the character recognition on the recognition device 22, and the recognition result to the recognition result storage device 23. Memory, and display the recognition result to the recognition result display device 24. Next to step ST24, the control device 29 indents the text attribute information.

Page 16

V. Description of the invention (14) Set 40 to transmit and store the memory information in the recognition result. The indentation device 40 is based on the recognition result of ^ 23. Text attribute information analysis in dictionary 25: The text attribute branch of each character of $ is less than k, the recognition result of the _ bank mother in the selection branch is selected, and the selection can be classified into the candidates in each selection branch. ^ Located in the range below], select text attribute information. The average and dispersion of the selected numbers become the smallest. The text attribute information is indented into one of the messages, such as the road-each device 40 is used to classify the text attribute information. In the figure, 3 φ, η, the pronunciation of the Chinese characters is the English vowel information expressed in English characters, 42 is the head of the sub-information before the pinyin, 43 # is the sister Li 4 «Zhong Ling Xi Yin Ci # 〇 Pinyin information on the number of words, 44 is the four-tone shell sfl in Chinese (4 types + soft sound recording, the number of words '46 is the radical of the text, and 45 is the total stroke of the text 40,,. ^ P 首 贝 讯, 47 is the number of strokes of the radical. 48 is not in the Chinese input method Xi Bu Tong 丨 recognize ^ ^ μ. | Of the five strokes] input before the stroke shape of the first priority order. The priority order is set as follows: Regarding the recognition result stored in the recognition result storage device 23, the order extracted from the text attribute information indentation device 40—shown in FIG. 14 is shown in FIG. 14 ' -For example, Figures ^ β ^ ^ in, press ..., from the 1st to the 4th order are stored in the recognition result memory device 23 "towel", "shen", "coin" and "medium" results, indicating the pinyin mark Vowel information, initial text and sub-number information, four-tone information, total stroke count of text I, Specific information first consistent cases, the first of Huashi ^] number information, and the shape of the head before five strokes draw input.

Page 17 407248 V. Description of the invention (15) Here, in the case where the text attribute information is too much "text attribute information, select a variety of I · generation information remaining after indentation in Figure i 3. In this embodiment 2 Set ^ 5, j = 1 When indented, display the initial text information and the number of characters marked in Pinyin: Candidate number 1, scattered 0, but because the precedence of the first text before Pinyin mark is higher than that in Pinyin mark Word data resources :: = Selection: Initial text information before the sound mark. = Device 40 chooses the best from a variety of text attribute information. Information: Go to step "25 'The text attribute display device 26 will be in the text attribute The text attribute information and its name are displayed in the text :: ϊ ΐ The text attribute :: of the text attribute display device 26 is displayed as shown in FIG. 15. In FIG. 15, 50 is a character belonging to _ Ziguang, and 51 is a name indicating the type of the character attribute information 50. In FIG. 2, the text attribute information displayed on the text attribute display device 26 is displayed without a pinyin before the initial text information "j, s, b, j", and "front text" is displayed on its name. ^^^ Then, in steps ^ 26 to ST28, and in step 1 to step ST17 of the first embodiment, the processing is performed to check whether the input device 27 has a text input, and whether the recognition result is indented into the recognition result of the recognition result in the device 28. Indentation and display of the recognition result to the recognition result display device 24 are changed. In the second embodiment, the text attribute information indentation device 40 uses the priority order of the text attribute information as shown in FIG. 13 but makes the text attribute information used by the noter and its priority order specified in advance. It is also possible to select from the text attribute information reduction device 4G.

Page 18 40724b

V. Description of the invention (16) Also 'In this embodiment 2, the text attribute display Pei 26 displays the text attribute information and its name for the text attribute information indented by the sexual information indentation device 40, but only the text attribute is displayed. Information is also available. Sister-in-law, if this embodiment 2 is used, because the sub-attribute information can be selected according to the recognition result, it can reduce the calculation amount of text recognition processing, can improve the accuracy of text recognition, and can reduce the input load of text attribute information, It can also reduce the effect of selecting the correct text. ° Also, if this embodiment 2 is used, in the case where the character recognition result is wrong, the same character can be input without having to take a note, which has the effect of reducing the load of notes. If you do n’t care about yourself, if you use this embodiment 2, after recognizing the text, you can only enter the text attribute information if the recognition result is wrong, which can reduce the input load of the text attribute information. Example 3 Example 3 of the present invention will be described next with reference to FIGS. 16 and 17. Fig. 6 is a block diagram showing the structure of a character recognition device according to Embodiment 3 of the present invention. The same or equivalent parts in Embodiment 16 as those in Embodiment 1 shown in Fig. 1 are assigned the same symbols and their descriptions are omitted. In FIG. 16, 60 is a recognition target indentation device that indents recognition target text based on text attribute information of correct text selected and input by the selection input device 27. In addition, the character recognition device of the third embodiment is different from the embodiment] in that the recognition target indentation device 60 has an alternative recognition result storage device 23 and a recognition result indentation device 28. The operation will be described next. Here 'FIG. 17 shows one character of the control device 29 in the third embodiment.

Page 19 sH Management's flowchart. The control device 29 performs step ST30 and steps as in step ST14 and step ST15 of the first embodiment, so that the temple property exhibition device 26 of the character display device 26 is displayed. Display, use selection input 4 ..., input text attribute information check. The text attribute display device 26 displays the characters here and selects the correct text input at 27; the tone: "message 'self-selected input device" == 3 pastes 2, the control device 29 will select the input device =; the text attribute sent Information-= to have a mother. " In the third embodiment, a character including "0" in the pinyin label of the character is selected as the character to be recognized. Step Sm 'is performed in the same way as step ST10 in the first embodiment, and the note information of the two written characters is loaded to the control. Then, go to step ST34 '. In the same way as in step ST11 of the first embodiment, input 21 is set to ^, and the identification information is transmitted to the identification device 22 using the identification target identification device 6G identification. The recognition device 22 recognizes according to the note information. = The recognition target text is from the recognition target indentation device 60 word 1, which is like text information indentation. In the third embodiment, the text of the recognition result is only a text having vowel information "0". Next, the recognition result display device 24 transmits the recognition result from the recognition device 22 in step ST35. In addition, in the third embodiment, the vowel information of the Chinese characters may also be used.

The content to be displayed on the text attribute display device 26 uses other text attribute information such as pronunciation, stroke number, radical information, and the like. ^ As shown above, if this embodiment 3 is used, the pen is written in front of the text of the note = the text attribute information of the text, because the recognition is based on the information ^ The text is indented, there is acceptable recognition, and the accuracy of text recognition is improved fruit. If you use this embodiment 3, in the case where the text attribute information is not input, the general text recognition without text attribute information can be used to reduce the input load of the text attribute information of the note taker. effect. In addition, if this embodiment 3 is used, the selected input character attribute information is Chinese vowel information, which can be easily selected for input in a portable information terminal without a physical keyboard, etc., which can reduce the text attribute information. Effect of input load. Embodiment 4 Next, Embodiment 4 of the present invention will be described with reference to FIGS. 18 to 27. Fig. 18 is a block diagram showing the structure of a character recognition device according to a fourth embodiment of the present invention. The same or equivalent parts in FIG. 18 as those in Embodiment 1 shown in FIG. 1 are assigned the same reference numerals, and descriptions thereof are omitted. In FIG. 18, 70 is a word association device that outputs word association results after associating words based on the recognition result information of the word portion of the recognition result memory device 23, and 71 stores word association results output from the word association device 70. Word association results memory device. 72 is a word association result indentation device that indents the word association result in the word association result memory device 71 according to the text attribute information of the unnote text inputted by the selection turn-in device 27; 73 shows the indentation using the word association result Device 72 retracted

V. Description of the invention (19) Word association result display device 73 for word association 1 result. In addition, the Λ child two recognition device of this embodiment is different from the embodiment in that it has a word association device 70, a word association result I, a word association result indentation device 72, and a word association result display device 73. The operation will be described next. At =, FIG. 19 is a flowchart showing the characters of the control device 29 in the fourth embodiment. The control device 29 first controls the recognition process at the step circle control structure κι * 二 杳 = 头 文 子. In addition, the processing of this step ^ 40 = = = = Example 丨 the processing of step ST1〇 ~ step sn7 is the same, in this case, Figure 20 shows that the recognition of the initials before the word is completed. Ϊ́ 你 丨 Π One example. In the figure, 36 is the initial character before the word, and the Chinese part is displayed in the display frame 24a of the recognition result display device 24. The other parts are shown in the figure and in the first embodiment shown in FIG. 6-㈣ίί In step ST41, the control device 29 transmits the first-ranked text of the recognition result stored in the recognition result record 7 Λ s Λ to the word association device 2 and the association device 70. It selects it from the word dictionary prepared in it. The text is the word that precedes the word. In this case, because = Ϊ = the Chinese character "中" is uploaded, select the word that starts with the word "中" from the word dictionary. The selected == 'control device 29 will pass the word that starts with the word association device Qilianxiangsha: Zhong "as the word association knot | and pass it to the single = memory device 71. The word association result storing means 71 memorizes the early association result transmitted. In this case, as shown in the word association

Page 22_407248 Description of the Five 'Inventions (20) "Medium plate", "Mediator", "Medium start word as word association knot result memory device 7" "中班"' 国 ", ... etc. in Chinese characters "Medium" results are stored. Next to step ST43, the control device 29 transmits one of the words of the memory device 7! To the word association? Result: Lian, = Lian I !: I want the result display device 73 to send the L words: single and two; FIG. 22 shows the word association device 7. For the recognition result, the word association result is displayed on the word association result display device after the early girl. In the figure, 74 is the word of association result displayed. In this FIG. 22, for example, X :; ending "shows that the device 73 displays" Middle Class "," Middle Board ", and Zhuo Er = zai ^, and the word that the reporter wants to enter is" China ", because" Medium: Beginning " The number of words is large, and the correct word is not displayed in the word association result display device 73 among the three association result words 74. Therefore, the character attribute display device 26 displays the character attributes of the characters under the word in step ST44. Information. It is shown in the text attribute display device. The text ^ attribute information is shown in Figure 23. In this example, the five vowel information of the Chinese text are represented by pinyin. 75 in the figure is the second character of the word. Text, attribute information, display 5 "a, i, u, e, 0" of various pinyin marks of different sizes. ^ Next, go to step ST45, and check whether the selection input device 27 has input text = attribute information. As a result, without inputting character attribute information, the word is regarded as having been input, and the word input processing is terminated. On the other hand, when the character input information is input by the selection input device 27, the process branches to step 31> 46. 5. Description of the invention on page 23 (21) The word that the person who wants to enter in “F” is “China”, because it is not among the 3 words of the association result displayed in the single 3 association result display device 73, The selection input device 27 is used to input the text attribute information of the second character, that is, the vowel information of Pinyin. Fig. 24 shows the character attribute information selected by the selection input device ". The 75a indicated by the network line in the figure is the character attribute information of the selected input. In addition, the example shown in the figure is because the second word Each word is the Chinese character "国", and input 5 kinds of vowel information—one of the vowel information of the word "国" in the pinyin mark rgu〇 ", for example, 11 〇 followed by step ST46, the control device 29 will select Enter the text attribute information of the M2 characters of the loser ^ 5 Fruit!入 装置。 Into the device. The word association result indentation device that received the information; 2 extracts the text attribute information of the word association result stored in the early word association result memory device 71, and selects the second word to have the text attribute selected by the input device 27 selected Word of information. Here, the word association result indentation device 72 extracts the text attribute information extracted from the words in the word association memory device 71: as shown in FIG. As shown in the figure, for a word association result memory device? , The association result of each word of "Zhongban", "Middleboard", "Zhongbao", "China", ..., and the corresponding pinyin information "zh〇ngban", "zhongban", "zhongbao", "zh" 〇ng gu〇 ", ... In addition, by using the word input information of the selected character attribute information, the final selection of the candidate word for indentation of the association indentation device 7 2 is shown in FIG. 2: Here, only the second word with the pinyin mark is selected. Of vowel information "u"

"中国" (zhong guo) "zhong kui" (zhong hua) "中流" (zhong 1iu) Next to step ST47, the control device 29 will self-memorize in the word association ^ device 71 word association In the result, the word association result in which the word association result is indented into L (3) with a high writing position is regarded as a word: the early S division association result display device 73, and the word association result display device == fselect . Fig. 27 shows word candidates as a result of word association. Candidates in the too " m Γ 5 Division. In the figure, 76 is the updated word "China". By selecting the word, = selected: incorrect: In addition, in the fourth embodiment, the processing ends at the order ^. Text attribute information, but if the second character is displayed after processing by selecting 1 private association, it can be displayed at any time. The text attribute information is displayed before the input is selected. In the fourth embodiment, the text attribute information of the word word is displayed, but the second idea is displayed regardless of the result of the real 4 to make the most appropriate selection of the word association example 2. Set the second word of the extracted word association result memory ^ 71. That is, after the text attribute information of the set text is displayed, the word association result of the Chinese memory of the self-remembered word is indented. The text is attributed to the attribute display device 26. The attribute information reduction device can be set by multiple By using the pre-attribute information to make the best selection of the attribute information, select the best Wen Xi Ling. In the fourth embodiment, make a note? · Think about the second word of the result. Sub-attribute information, but it is also possible to rotate the second character after the initial character to the eleventh character and the character attribute information of the π + l character. Also in the fourth embodiment, after displaying the result of word association processing based on the preceding character, the character attribute information of the second character is entered for indentation, but the character attribute of the second character is input after the first character of the note. Information condition 'allows word association using word attribute information of the second character. As shown above, if this embodiment 4 is used, the word attribute of unwritten text is used for the word association result to the word association result after the words of the word are formed by partially writing a word of two or more words. Use the text attribute information to reduce the word candidate. There is a T effect that makes the word candidate more than the text "word association input more efficient." Houkou lingering if: Use Example 4 ‘Because you can take note of words and recognize words, you can improve word recognition results with text attribute information’ Effect of invention Degree of effect of inputting words. As shown above, if the present invention is used, the recognition result is stored in the memory, and the recognition result is set to a memory. The recognition device recognizes the fruit according to the note information, and displays it on the recognition result display device. ^ Recognize the article… The displayed text attribute information is selected for input and displayed on the recognition result display device. Second, the recognition result text is 'indented only because the text recognition result is incorrect.' Simple text is the input of self-selection and input of attribute information__407248 V. Description of the invention (24) '' 'Enter' can input correct text, which can suppress the increase of the load of the note-holder, improve the recognition result of the text, The effect of a character recognition device capable of recognizing characters with high accuracy can be obtained. If the present invention is used, because the text attribute information indentation device is also provided in the structure, the text attribute information displayed on the text attribute display device is indented by the text attribute information stored in the recognition result memory device to recognize the text. Because the correct text is rotated only when the text recognition result is not correct, it has the effects of making it easier to enter text attribute information and reducing the amount of calculation required for text recognition processing. If the present invention is used, the text attribute information indented by the text attribute information displayed on the text attribute display device is extracted from the text by inducing the text attribute information of the recognition result memorized in the recognition result storage device by using the structure. Attribute information Select the best text attribute information to recognize the text. Entering the correct text only when the text recognition result is incorrect has the effect of making it easier to enter text attribute information. If the present invention is used, the structure of the text is displayed on the display device by selecting the input attribute based on the indentation of the recognition object. The text attribute information of the input object is selected to perform the indentation of the text of the recognition object. Enter the text information of the recognition target and the information obtained from the input device. The result of the recognition of the text by the recognition device is displayed on the recognition result display device. The text of the note can be rotated to the input device. 'There is an effect that a character recognition device capable of recognizing characters at high speed and high accuracy can be obtained. If the present invention is used, the vowel information of the text is used as the text in the structure.

V. Description of the invention (25) The effect of the word attribute information can be easily inputted. If the word memory associated with the present memory is used to associate the word association with the word association, and the word association is selected using the word association to input the word association of the word association device, part of the note construction attribute information can be input to the word if it is deleted. According to the memorized word attribute information, it will be displayed, the text will be recognized, and the text attribute information will be changed. If the text attribute display device of the result memory is used to load the device, it can recognize a variety of characters, only carry the attribute information of the character without a physical keyboard, and reduce the input of the attribute information of the text, such as a tape-less information terminal. ^ 48 Ming, due to self-knowledge results * Texts that want results are condensed into words. Words and characters can be identified, and it is also associated with the fact that texts can only be entered in structures that are easy to construct and obviously indented. The cognition device selected by the cognition text sets the single fruit memory attribute word display recognition. There are words that can be ordered to be displayed in the single set. According to the message, the self-recording is displayed in other characters. Because of the number, some effect word attribute devices are not displayed on the device. Part of the word association of the more southern word association association is selected and recalled in a single word association, and the recognition result of the result record can be obtained in the input. The input word association is expected to be entered into the multi-character text, and the device can be displayed at high speed. The result record shows the text of the loaded word and is highly savvy, because the text using the word association result is placed in the structure. Text attribute information Short text attribute information selection Most word recognition result is incorrect Information indentation device 'The text attribute information of the text of the word association result is indented only when it is correctly entered and the word is extracted and memorized in the word association attribute After the information, it will be displayed in the entered word attribute information. The text attribute information is better.

Page 28 407248 V. Description of the invention (26) Attribute information becomes easy, which has the effect of inputting words at higher speed and accuracy. Brief Description of the Drawings Fig. 1 is a block diagram showing the structure of a character recognition device according to the first embodiment of the present invention. Fig. 2 is a flowchart showing the flow of one word processing in the control device of the first embodiment of the present invention. Fig. 3 is an explanatory diagram showing an example of an input pattern written by the input device according to the first embodiment of the present invention. Fig. 4 is an explanatory diagram showing an example of contents of a recognition result stored in a recognition result storage device after recognition by the recognition device according to the first embodiment of the present invention. Fig. 5 is an explanatory diagram showing an example of the first-digit recognition result displayed by the recognition-result display device according to the first embodiment of the present invention. Fig. 6 is an explanatory diagram showing an example of character attribute information displayed by the character attribute display device according to the first embodiment of the present invention. Fig. 7 is an explanatory diagram showing an example of text attribute information of correct text selected and input by the selection input device according to the first embodiment of the present invention. Fig. 8 is an explanatory diagram showing an example of the content of a recognition result storage device after the recognition result indentation device of the first embodiment of the present invention extracts text attribute information of the recognition result. FIG. 9 is an explanatory diagram showing an example of the contents of a recognition result storage device after the recognition result indentation device according to the first embodiment of the present invention shrinks the recognition result based on the text attribute information. FIG. 10 shows a recognition result display device according to the first embodiment of the present invention.

Page 29

Structure 5. Invention description (27) An explanatory diagram showing an example of a new recognition result. Character recognition device Fig. 11 is a block diagram showing a second embodiment of the present invention. Fig. 12 is a flowchart showing the flow of control processing in the second embodiment of the present invention. Figure 1 is a diagram showing an example of character attribute information used by the device 40 in the second embodiment of the present invention. Fig. 14 is an explanatory diagram showing an example of the identification options for the second embodiment of the present invention. The recognition result memorized by the character attribute information is indented by the device. The text display figure 4 shows the actuality of the present invention. The text attribute information and an example of its name are explained. ISMfi technology master = 丄 4 —. &Amp;. , ^ ... A 乃 ~ ', ΤΓ ~ ... ~ 机 明 图. Fig. 16 is a block diagram showing a character in a third embodiment of the present invention. The structure of the identification device is set to 1 character. Fig. 17 is a flowchart showing the flow of the control process in the third embodiment of the present invention. Fig. 18 is a block diagram of a character recognition device & Fig. 19 is a flowchart showing a flow of processing by the control device in Embodiment 4 of the present invention. 1U 至 FIG. 20 is an explanatory diagram showing an example of a state after the recognition process of the words before the constituent words of the fourth embodiment of the present invention is completed. Fig. 21 is an explanatory diagram showing an example of a word association result memorized in the word association result in the fourth embodiment of the present invention. ...

Page 30 -407248- V. Description of the invention (28) FIG. 22 is an explanatory diagram showing an example of the state of the word association device according to the fourth embodiment of the present invention for displaying the result on the word association result display device after recognizing the words associated with the recognition result. Fig. 23 is an explanatory diagram showing an example of a state in which the character attribute display device of the fourth character of the present invention displays the character attribute information of the second character. Fig. 24 is an explanatory diagram showing an example of a state in which the selection input device of the fourth embodiment of the present invention has selected the character attribute information of the second character. Fig. 25 is an explanatory diagram showing an example of a result of extracting text attribute information from a word in the word association result memory device by the word association result indentation device according to the fourth embodiment of the present invention. Fig. 26 is an explanatory diagram showing an example of a state of a word in a word association result indenting device using a character attribute information of a selected input to indent a word association result indentation device according to a fourth embodiment of the present invention. Fig. 27 is an explanatory diagram showing an example of word candidates displayed by the word association result display device according to the fourth embodiment of the present invention. FIG. 28 is a block diagram showing the structure of a character recognition device of the conventional technique 1. FIG. FIG. 29 is a block diagram showing the structure of a character recognition device of the conventional technology 2. FIG. FIG. 30 is an explanatory diagram showing an example of an input form in the conventional technique 2. FIG. FIG. 31 is an explanatory diagram showing an example of a character recognition result in the conventional technique 2. FIG. FIG. 32 shows an example of the contents of a word list in the conventional technique 2. FIG.

Page 31 V. Description of Invention (29) 407248 Explanation of symbols 2 1 input device, 22 recognition device, 23 recognition result memory device, 24 recognition result display device, 2 5 character attribute dictionary, 2 6 character attribute display device, 27 selection input device, 28 recognition result indentation device, 40 characters Attribute information indentation device, 60 recognition object indentation device, 70 word association device, 71 word association result memory device, 72 word association result indentation device, 73 word association result display device.

Page 32

Claims (1)

-407248- 6. Scope of Patent Application 1. A text recognition device, comprising: input device 'input text note information; a recognition device, which recognizes text based on the note information output by the input device. Then outputs its recognition result; recognition result memory device To memorize the recognition result output from the recognition device; a text attribute display device that displays the text attribute information; a selection input device to select the text attribute information displayed on the text attribute display device; and a recognition result indentation device based on use The text attribute information selected by the selection input device is self-stored in the recognition result of the recognition result storage device, and the recognition result displayed on the recognition result display device is indented. 2. For example, the character recognition device of the scope of patent application, wherein the text attribute information of the recognition result stored in the recognition result memory device is extracted, and the indented text attribute information of the text attribute information displayed on the text attribute display device is set. Indentation device. 3. For the character recognition device in the second item of the patent application, the text attribute information indentation device extracts the text attribute information stored in the recognition result of the recognition result storage device, and selects the best text attribute information from a variety of text attribute information. . 4. A text recognition device, comprising: an input device for inputting text note information; a text attribute dictionary that stores text text attribute information; a text attribute display device that displays the text attribute information; Page 33_d〇7248- VI. Patent input range selection input device, used to select the text attribute information displayed on the text attribute display device; identify the object indentation device, according to the text entered by the selection input device The attribute information indents the text of the identification object; the identification device outputs the identification result after identifying the text based on the note information output from the input device and the identification object text information indented by the identification object indentation device; and displaying the identification result; A device that displays a recognition result output from the recognition device. 5. If the character recognition device of item 1, 2, 3 or 4 of the patent application scope, the vowel information of the text is used on the text attribute information. 6. For example, the word recognition device in the scope of patent application No. 1, 2, 3 or 4 further includes: a word association device, which outputs word association results after associating words according to the recognition result of a part of the words of the recognition device; word association result memory A device that memorizes the word association result output from the word association device; a word association result display device that displays the word association result stored in the word association result memory device; and a word association result indentation device that is based on the input selected by using the selection input device The text attribute information of the words constituting the word is indented from the word association result stored in the word association result storage device and displayed on the word association result display device. 7. If the character recognition device of item 5 of the patent application scope, which includes more Page 34_407248_ VI. The scope of patent application includes: word association device, which associates words based on the recognition result of the word part of the recognition device, and outputs word association results; word association result memory device, which memorizes word association output from the word association device Results; the word association result display device displays the word association result memorized in the word association result memory device; and the word association result indentation device, based on the text attribute information of the words constituting the word selected and input using the selection input device, is self-memorized in The word association result displayed in the word association result display device is indented among the word association results in the word association result storage device. 8. If the character recognition device of item 6 of the patent application includes a word attribute information indentation device, extract the text attribute information of the words stored in the word association result memory of the word association result memory device, and display the text attribute information. Indentation for displaying text attribute information of the device. 9. For example, the character recognition device of the 7th scope of the patent application, wherein a word attribute information indentation device is provided, and the text attribute information of the words stored in the word association result memory of the word association result device is extracted and displayed on the text attribute Indentation for displaying text attribute information of the device. 10. As for the character recognition device of the eighth patent application, the word attribute information indentation device extracts the text attribute information of the text memorized in the word association result memory of the word association result memory device, and selects the best one from a variety of text attribute information. Text attribute information. 11. If the character recognition device of item 9 of the patent application scope, wherein the word Page 35 6. Patent application Fan Yue G7248 The attribute information indentation device extracts the text attribute information of the words stored in the word association result memory device and selects the best text attribute information from a variety of text attribute information. Page 36