JP2007213416A - Apparatus, method and program for retrieving handwritten character string - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve the accuracy of retrieval in the case of retrieving an inputted handwritten character. <P>SOLUTION: A handwritten character string retrieving apparatus is provided with: an accuracy degree determination part 3-4 for determining an accuracy degree (matching degree) between a handwritten character inputted from a handwritten character input part 2-1 and recognized by a handwritten character recognition part 3-3 and a standard form of the character; a retrieving candidate range determination part 3-5 for determining a retrieving candidate range on the basis of the determined coincidence degree; and a sentence retrieving part 3-6 for retrieving a sentence file on the basis of the retrieving candidate range of the handwritten character recognized by the handwritten character recognition part 3-3 and displaying the retrieved result. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention searches a document file recorded in the past using an input handwritten character as a search key, and causes a computer to execute a handwritten character string search device, a handwritten character string search method, and the character string search method for displaying a search result. Is related to the program.

  Conventionally, a search device that searches a document file based on an input character string and displays a search result has been realized.

  For example, in the text search device disclosed in Patent Document 1 below, first, information relating to character quality and state is extracted from the input text image as search auxiliary information, and the character recognition result and the search auxiliary information are used as search text data. Stored in a text database. Then, the collation according to the extracted search auxiliary information is performed at the time of collation between the keyword character of the text search input from the keyword input means and the text data for search.

  Further, for example, in the image search device of Patent Document 2 below, first, a new character string set obtained by expanding a specified character string with reference to a database storing character string sets that are easily misread by character recognition. Is generated. Then, the generated set of character strings is displayed so as to be selectable as search target character string candidates, and the character string search of the sentence image is performed based on the character string selected by the user and further narrowed down and the specified character string. Yes.

JP 2002-312398 A Japanese Patent Laid-Open No. 2002-92016

  The text search device of Patent Document 1 is configured to input a search character string by keyword input means. In this search device, the input keyword is a character code string, which is input by a keyboard or a conventional method. If the user clearly recognizes the character string to be searched, this method can be used to input the search character string without any problem. However, if the user inputs the search character string with ambiguous memory, , You must repeat the act of entering keywords after looking at the search results.

  A configuration in which a handwritten character string recognition function is incorporated in the keyword input means is also conceivable. However, in this case as well, if it is simply incorporated, only the character code is input for the same operation as the input by the keyboard etc., and if the search character string is input with ambiguous memory, the search result You must repeat the act of entering keywords after seeing.

  In the text search device of Patent Literature 1, the text recognition unit that is the search target is configured to output search auxiliary information. This auxiliary information is auxiliary information related to the text to be searched, This is information such as presence / absence and whether the input character is printed or handwritten. That is, when a handwritten character string is input as a search keyword, the user wants to make effective use of the added information because it is a handwritten character string. I couldn't.

  Further, in the image search device of Patent Document 2, in order to efficiently collate with the recognition result of the text to be searched, the image search apparatus is input with reference to a database storing character string pairs that are easily misread by character recognition. A new character string set obtained by extending the search character string is generated from the search character string. Then, the character string search of the sentence image is performed based on the character string selected by the user from the plurality of character strings generated based on the input search character string (input keyword). At this time, if the user selects an appropriate character string, an effective search can be performed. If not, the search result desired by the user may not be displayed. Further, in the image search device of Patent Document 2, it is effective when an input keyword is input from a keyboard, but when a new character string set is generated from a result of recognizing a handwritten character, a more uncertain candidate is present. It may be displayed, and it is difficult to perform a search as it is.

  The present invention has been made in view of the above-described problems, and an object of the present invention is to improve the search accuracy when searching for input handwritten characters.

The handwritten character string search device according to the present invention is recognized by a handwritten character input means for inputting a handwritten character, a handwritten character recognition means for recognizing a handwritten character input from the handwritten character input means, and the handwritten character recognition means. A degree-of-matching determination unit that determines the degree of matching between the handwritten character and the standard font of the character, and a candidate-of-search range determination unit that determines a search candidate range based on the degree of matching determined by the degree-of-matching determination unit And a search means for searching for a character string file based on the search candidate range for the handwritten character recognized by the handwritten character recognition means and displaying the search result.
In another aspect of the handwritten character string search device of the present invention, a handwritten character input unit for inputting a handwritten character, a handwritten character recognition unit for recognizing a handwritten character input from the handwritten character input unit, and the handwritten character A size determining unit that determines the size of the handwritten character input from the input unit from a relative relationship with the input screen, and a search area is determined based on the size of the handwritten character determined by the size determining unit. Search area determination means, and search means for searching for a character string file based on the search area for the handwritten characters recognized by the handwritten character recognition means and displaying the search results.
In another aspect of the handwritten character string search device of the present invention, the handwritten character input means for inputting a handwritten character, the handwritten character recognition means for recognizing the handwritten character input from the handwritten character input means, Input time detecting means for detecting an input time of a handwritten character input from the handwritten character input means; search candidate range determining means for determining a search candidate range based on the input time of the handwritten character detected by the input time detecting means; And a search means for searching for a character string file based on the search candidate range for the handwritten character recognized by the handwritten character recognition means and displaying the search result.

The handwritten character string search method of the present invention is a handwritten character string search method in a handwritten character string search device comprising a handwritten character input means for inputting a handwritten character, the handwritten character input from the handwritten character input means A handwritten character recognition step for recognizing a handwriting character recognized in the handwritten character recognition step, a matching degree determination step for determining a matching degree between a standard character form of the character, and a matching determined by the matching degree determination step A search candidate range determining step for determining a search candidate range based on the degree, and searching for a character string file based on the search candidate range for the handwritten character recognized by the handwritten character recognition means, and displaying the search result A search step.
Another aspect of the handwritten character string search method of the present invention is a handwritten character string search method in a handwritten character string search device including a handwritten character input means for inputting a handwritten character, the input from the handwritten character input means. A handwritten character recognizing step for recognizing a handwritten character, a size determining step for determining the size of the handwritten character input from the handwritten character input means from a relative relationship with respect to the input screen, and the size determining step. A search area determining step for determining a search area based on the size of the determined handwritten character, and searching for a character string file based on the search area for the handwritten character recognized in the handwritten character recognition step, and the search result And a search step for displaying.
Another aspect of the handwritten character string search method of the present invention is a handwritten character string search method in a handwritten character string search device including a handwritten character input means for inputting a handwritten character, the handwritten character input means A handwritten character recognizing step for recognizing a handwritten character input from the handwriting, an input time detecting step for detecting an input time of the handwritten character input from the handwritten character input means, and an input of the handwritten character detected by the input time detecting step A search candidate range determining step for determining a search candidate range based on time, and a character string file is searched based on the search candidate range for the handwritten character recognized in the handwritten character recognition step, and the search result is displayed. A search step.

  The program of this invention is for making a computer perform the said handwritten character string search method.

  According to the present invention, when searching for input handwritten characters, the search accuracy can be improved.

  In the present invention, when an accurate handwritten character is input, a sentence including a character string that matches 100% is presented as a search result. When a character string of a broken character is input, a wider range of sentences is displayed as a search result. Present as. The user can input a handwritten character string with an accurate number of strokes if he / she wants to search for a perfect sentence, and can enter a broken character by hand if he / she wants to search for a sentence with an ambiguous memory.

  In the present invention, when the user inputs a large amount of handwritten characters, the text is searched from the Internet, and when the user inputs a small amount of handwritten characters, the search is performed from the local disk. As described above, since the character string serving as the search key and the size of the search range can be specified at the same time, the search operation can be simplified.

  Further, in the present invention, when the user is at a loss in inputting the character of the search key, since the sentence including the similar character of the character is set as the search target, the search process can be performed more accurately.

(First embodiment)
FIG. 1 is a schematic configuration diagram of a handwritten character string search system according to an embodiment of the present invention.
In FIG. 1, 1 is a handwritten character string search device. The handwritten character string search device 1 includes a liquid crystal display unit 2, a display-integrated coordinate input unit, a CPU, ROM, RAM, HD, a slot into which a recording medium is inserted, an interface such as a USB for connecting to other information devices, and wireless communication And an antenna 4 and the like. The wireless communication unit communicates with, for example, a personal computer having another communication function.

  The handwritten character string search device 1 recognizes the locus of the handwritten character input by the coordinate input unit, and displays the text as a result of searching from the character code of the recognition result and its additional information. In FIG. 1, reference numeral 2 denotes a general liquid crystal display unit, which displays a handwriting trajectory input to the liquid crystal display unit, a text of a search result, and the like.

  In FIG. 1, reference numeral 3 denotes a general local area network cable (LAN cable), whereby the handwritten character string search device 1 can communicate with the personal computer 5. In FIG. 1, reference numeral 4 denotes an antenna for wireless communication, which shares a screen displayed by the handwritten character string search device 1 and exchanges input text information and the like. In FIG. 1, a general personal computer 5 is connected to the handwritten character string search device 1 via the local area network 4.

FIG. 2 is a block diagram showing a system configuration of the handwritten character string search device 1 according to the embodiment of the present invention.
In FIG. 2, reference numeral 2-1 denotes a general coordinate input unit, which is composed of a transparent resistive film digitizer or the like, and is connected to the CPU 6 via the system bus 11. In FIG. 2, reference numeral 2 denotes a general liquid crystal display unit as described above, which includes a liquid crystal display element, a liquid crystal control circuit, and a display memory, and is connected to the CPU 6 via the system bus 11.

  In FIG. 2, reference numeral 6 denotes a general CPU, which is connected to the RAM 7, ROM 8, coordinate input unit 2-1, liquid crystal display unit 2, and the like via the system bus 11, and is processed by a program stored in the ROM 8. Perform the action. In the present embodiment, display of a handwritten locus or display of a character font is executed according to an instruction from the CPU 6.

  In FIG. 2, reference numeral 7 denotes a general RAM, which is used as a work area for a captured image processing program or the like in storage of image data or processing in the CPU 6. In FIG. 2, 8 is a general ROM, and a program for executing a handwriting recognition processing procedure, a text search processing procedure, an application processing procedure on another handwritten character string search device 1 and the like according to the present invention. It is remembered. The ROM 7 also stores a CCD control program, a liquid crystal display control program, and other processing procedure programs.

  In FIG. 2, 9 is a small hard disk (HD) used in a general portable device, and is connected to the CPU 6 via the system bus 11. In the present embodiment, the input text file or the like is stored in the HD 9.

  In FIG. 2, reference numeral 10 denotes a general local area network (LAN) interface, which communicates with the personal computer 5 as an external device. In FIG. 2, 11 is a general system bus for exchanging data between the CPU 6 and the ROM 6, RAM 8, and other devices.

  In FIG. 2, reference numeral 12 denotes a general wireless communication card, which has a built-in wireless LAN controller and antenna. By passing a specified command to the controller in the communication card 12, it is possible to connect to the wireless local area network and execute transmission / reception of digital images and transmission / reception of control commands. The transmission / reception protocol itself is the same as the conventional method.

FIG. 3 is a block diagram illustrating a functional configuration of the handwritten character string search device 1 according to the first embodiment.
In FIG. 3, reference numeral 2-1 denotes the coordinate input unit described above, which is composed of a transparent resistive film digitizer or the like, and can read XY single table data of a handwritten locus written with a pen.

  In FIG. 3, reference numeral 3-1 denotes a character cutting unit, which separates the position coordinate data representing the locus of the handwritten character string input by the coordinate input unit 2-1 into data for each character according to a specified rule. I do.

  In FIG. 3, 3-2 is a handwritten character recognition dictionary, in which normalized data of general handwritten trajectory data of characters to be recognized is stored.

  In FIG. 3, reference numeral 3-3 denotes a handwritten character recognition unit. The handwritten character recognition unit 3-3 extracts a feature point of a locus of handwritten characters written by a general online handwritten character recognition algorithm, and the data stored in the handwritten character recognition dictionary 3-2 Matching processing is performed, and the most similar dictionary data is set as the first candidate.

  In FIG. 3, 3-4 is a handwritten character accuracy determination unit, which determines the accuracy (coincidence) from the difference between the input handwritten character and the standard pattern stored in the handwritten character recognition dictionary 3-2. Then, information on the accuracy is output.

  In FIG. 3, reference numeral 3-5 denotes a search candidate range determination unit, which determines the display range of the search result based on the accuracy information sent from the handwritten character accuracy determination unit 3-4.

  In FIG. 3, reference numeral 3-6 denotes a text search unit, which stores based on the display range determined by the search candidate range determination unit 3-5 using the character string output by the handwritten character recognition unit 3-3 as a search key. Search for sentences.

  In FIG. 3, reference numeral 2 denotes the above-described general liquid crystal display unit, which displays a text image or the like indicating the result of search by the text search unit 3-6.

  FIG. 4 is a diagram showing a handwritten character input screen for search keys, and FIG. 5 is a diagram showing a screen after the handwritten character recognition process shown in FIG. 6 and 7 are diagrams showing the relationship between one input character and the standard pattern stored in the handwritten character recognition dictionary 3-2. FIG. 8 is a flowchart of the handwriting search key input process of the handwritten character string search device 1 according to the first embodiment.

Hereinafter, the processing of the handwritten character string search device 1 will be described with reference to the flowchart of FIG.
When the user inputs handwritten characters in the search keyword input area 4-1 shown in FIG. 4 and the search key 402 is pressed (clicked) or a certain time has elapsed, recognition of the handwritten characters is started. At this time, a handwritten character string is input as XY position coordinate data by the coordinate input unit 2-1 in FIG. Then, data for the character string is stored in the input buffer. After this general handwritten character recognition start process is performed, the process of the flowchart of FIG. 8 is started.

  First, in step S8-1, a handwriting search key input process is started, and a work area and the like are secured.

  Subsequently, in step S8-2, a process of cutting out character trajectory data for each character from the handwritten input character string is performed. This character cutting process is performed by the character cutting unit 3-1, and the process may be a conventional general algorithm. For example, a virtual one character frame is created based on the height of the entire character string, and the characters are separated by a process of sequentially separating the locus included therein as one character data. If it is the handwritten data shown in FIG. 4, it will be divided into three characters.

  Subsequently, in step S8-3, matching processing of the one-character handwriting data cut out in step S8-2 and the standard handwriting pattern in the handwritten character recognition dictionary 3-2 is sequentially performed. This matching process is performed by the handwritten character recognition unit 3-3, and the process itself may be a general algorithm. For example, the feature points are extracted from the normalized input handwriting data, and the standard in the handwritten character recognition dictionary 3-2 that is most similar is calculated from the distance calculation with the feature point data previously created from the standard handwriting data. A handwriting pattern is output as a recognition result.

Here, FIG. 6 and FIG. 7 will be described.
In FIG. 6, 6-1 is an input character input by handwriting, and shows an example in which “齊” is written with eight brushes. In FIG. 6, 6-2 is a standard handwriting pattern in the handwritten character recognition dictionary 3-2, and handwriting written with 12 brushes is made into a dictionary.

FIG. 7 shows another input example.
In FIG. 7, reference numeral 7-1 denotes an input character input by handwriting, and shows an example in which “齊” is written with 12 brushes.

  Here, in this embodiment, as shown in FIG. 7, when an input character is written in the same shape as the standard handwriting pattern in the handwritten character recognition dictionary 3-2, the accuracy is 80% or more. Further, as shown in FIG. 6, when an input character is written with a stroke number different from the standard handwriting pattern in the handwritten character recognition dictionary 3-2, the accuracy is 80% or less.

  In step S8-4, the recognition result dictionary pattern is checked. If the dictionary pattern with the correct number of strokes is the recognition result dictionary pattern, the process proceeds to step S8-5. If the collapsed character pattern dictionary pattern is the recognition result dictionary pattern, the process proceeds to step S8-6. This process stores a plurality of patterns for a certain character in the standard pattern in the handwritten character recognition dictionary 3-2, and one pattern is a handwriting pattern written with the correct number of strokes, The pattern is a case where the pattern is a handwritten pattern (like a curse) written in a broken manner. In this case, if written with the correct number of strokes, the recognition result will match the handwriting pattern written with the correct number of strokes, so if the result is output and written with a broken handwriting pattern, it will match the broken character pattern in the dictionary. That is why. Note that this processing is not necessary in a system in which only one correct handwriting pattern is registered for one character type as the standard pattern in the handwritten character recognition dictionary 3-2. The process of step S8-4 is performed by the handwritten character recognition unit 3-3.

  In step S8-5, a process of calculating the accuracy from the difference in similarity between the character dictionary pattern and the input handwriting pattern is performed. The process of step S8-5 is performed by the handwritten character accuracy determination unit 3-4. In this embodiment, when the number of strokes is the same as that of the dictionary pattern and the shape is almost the same, the accuracy is 100%, 90% if the number of strokes is ± 1 stroke, and 70% if the number of strokes is ± 2 strokes. This numerical value is an example in the embodiment, and the process of subtracting −1% or the like according to the distance difference between the feature point and one point is performed in more detail.

  On the other hand, in step S8-6, the accuracy is calculated from the difference between the coefficient in the character recognition dictionary pattern storage × the dictionary pattern. The process of step S8-6 is performed by the handwritten character accuracy determination unit 3-4. Here, the coefficient is stored in advance according to the breaking method of the characters stored in the handwritten character recognition dictionary 3-2. For example, if a coefficient 0.6 is stored in a certain broken dictionary pattern, the coefficient 0.6 is read here. Then, the difference between the coefficient 0.6 × the collapsible character dictionary pattern = accuracy. Therefore, even if a trajectory pattern that is 100% identical to the collapsible dictionary pattern is input, the accuracy is 60%.

  Subsequently, in step S8-7, it is checked whether the recognition process for all handwritten characters has been completed. If the recognition processing for all handwritten characters has been completed, the handwriting search key input processing is terminated. If not, the processing from step S8-2 is performed on the unprocessed characters. Do. By performing processing in this way, the recognition result and accuracy of the character string of the input handwriting search key are determined and output.

  FIG. 9 is a flowchart of processing for determining the search candidate range from the character string of the recognition result and the accuracy information of the character in the handwritten character string search device 1 according to the first embodiment. The processing of FIG. 9 is started after the processing in the flowchart of FIG. 8 is completed and the input handwriting search key is converted into a character code string and accuracy information as a recognition result. Here, the processing of the flowchart of FIG. 9 is performed by the search candidate range determination unit 3-5.

  First, in step S9-1, search candidate range determination processing is started and a work area and the like are secured.

  Subsequently, in step S9-2, a character code string as a recognition result generated by the processing in the flowchart of FIG. 8 and accuracy information reading processing are performed. Here, the character code string and accuracy information of the read recognition result are stored in a buffer.

  Subsequently, in step S9-3, read processing of accuracy information of one character of the search key character string is performed.

  Subsequently, in step S9-4, the read accuracy information is judged. If the accuracy is 80% or more, the process proceeds to step S9-6, and if the accuracy is 80% or less, the process proceeds to step S9-5.

  In step S9-5, the read accuracy information is determined. If the accuracy is 50% or more, the process proceeds to step S9-7. If the accuracy is 50% or less, the process proceeds to step S9-8.

  In this embodiment, the processing in steps S9-4 and S9-5 is classified into the following three types of processing in steps S9-6 to S9-8 according to the accuracy. You may make it control more finely by classifying into 10 types in% increments.

  In step S9-6, the recognition candidate is set in the search key buffer. At this time, if the accuracy of the character is 80% or more, the recognition result is used as it is as a search key, so it is stored in the search key buffer as it is.

  In step S9-7, the first candidate candidate of the recognition result and the second candidate candidate of the recognition result are set in the search key buffer. For example, the recognition result of the first handwritten input character is “齋”, the accuracy is 60%, the second candidate is “齊”, and the recognition result of the second handwritten input character is “Fuji”. The case where the accuracy is 90% is considered. In this case, “Saito” is set as the first search key and “Saito” is set as the second search key.

  In step S9-8, a search similar character is searched based on the recognition result, and set to a search key. This process is performed when the accuracy of the recognition result is 50% or less, and a character similar to the recognition result in that case is registered as a search key. Here, the character similar to the recognition result is a character that is easy for the user to make a mistake during the search. At this time, since it may be different depending on the user, a standard mistaken character may be set at first, and a learned character may be set while the search process is repeated.

In step S9-9, it is checked whether all input character strings have been processed.
If all input character strings have been processed, the process proceeds to step S9-10. If all input character strings have not been processed, the process returns to step S9-4 to process the next character. .

  In step S9-10, the search character string created from the recognition candidates processed in steps S9-6, S9-7, and S9-8 is organized to create a search key character string.

  For example, consider a case where locus data of “Hiroshi Saito” is input as a handwriting input character string. Then, the accuracy of the recognition result of the first character is 60%, the recognition candidates are “齋” and “齊”, the accuracy of the recognition result of the second character is 90%, the recognition candidate is “Wisteria”, the third It is assumed that the accuracy of the recognition result of the character is 40%, the recognition candidate is “Kan”, and “Hiro” is searched as a search similar character. In this case, a search key is created with a combination of the above candidates. That is, the search key 1: Hiroshi Saito, Search key 2: Hiroshi Saito, Search key 3: Hiroshi Saito, Search key 4: Hiroshi Saito are created.

  In step S9-11, the work area and the like are released, and this process ends.

  FIG. 10 is a flowchart of sentence search processing in the search candidate range in the handwritten character string search apparatus 1 according to the first embodiment. The process of FIG. 10 is started when a search key is created by the process of the flowchart of FIG. Here, the process of the flowchart of FIG. 10 is performed by the text search unit 3-6.

  First, in step S10-1, a text search process is started, and an initialization process such as securing a work area to be used in the text search process is performed.

  Subsequently, in step S10-2, a process of reading one file of stored sentences is performed. Here, if the stored text is a text file, it is stored in the buffer as it is. If the stored text is a locus file of handwritten characters, the handwritten character recognition unit 3-3 performs character recognition and converts it into a text file, which is then stored in a buffer.

  Subsequently, in step S10-3, if there is the same character string as the search key, processing for storing the sentence file is performed. More specifically, the text file read in step S10-2 is matched with the search key, and if there is a matching character string, the file is stored there.

  Subsequently, in step S10-4, it is checked whether there is an unprocessed accumulated sentence file. If there is an unprocessed stored sentence file, the file pointer is advanced to the next file and the process returns to step S10-2. On the other hand, if all the files have been searched, step S10-5 is executed. Proceed to

  In step S10-5, a process for displaying the retrieved text file as a candidate is performed. More specifically, in step S10-3, the stored file is read, and a process of displaying a page of a portion where a matching character string exists is performed. The display itself is a process for displaying a general text string.

  By performing processing in this way, handwritten characters that display accurate search results when entering accurate characters that take advantage of the characteristics of handwritten input characters, and display a wide range of search results when entering corrupted characters A column search device can be realized. In addition, by adjusting how characters are broken, the range of text displayed as a search result can be controlled. In addition, when the character string to be searched is uncertain, it is possible to comfortably search by handwritten character by inputting an uncertain broken character.

(Second Embodiment)
The handwritten character string search device according to the first embodiment determines the search candidate output range according to the accuracy of the input character. There is a problem of where to search. In a normal search system, a search area is specified as a separate setting item in addition to the search key input area. Specifically, for example, search from the local disk, search from the intranet, search from the Internet, and the like are designated.

  In the handwritten character string search device according to the first embodiment, the search key and the search area are specified at the same time only by the action of determining the search area based on the relative size of the input character string and inputting the handwritten character. Is what you do.

  FIG. 11 is a block diagram illustrating a functional configuration of the handwritten character string search device 1 according to the second embodiment. The same code | symbol is attached | subjected about the structure similar to the handwritten character string search apparatus which concerns on 1st Embodiment, and the detailed description is abbreviate | omitted.

  In FIG. 11, reference numeral 11-1 denotes a character string size determining unit that determines the size of the input handwritten character string from a relative relationship with the input screen and outputs information related to the size. .

  In FIG. 11, reference numeral 11-2 denotes a search area determination unit that determines a search area according to information relating to the size of the character string sent from the character string size determination unit 11-1. Information on the determined search area is output to the text search unit 3-6. Here, the search area is a place where the text to be searched is stored. In this embodiment, the smallest place is in the local disk, the next designated area is in an intranet or a home server, and the maximum range is in the Internet.

12 and 13 are diagrams illustrating examples of input of handwritten search characters.
12 and 13, reference numeral 12-1 denotes a handwriting search character input area, and handwritten characters written in this area are recognized as search keys. 12 and 13, reference numeral 12-2 denotes a search button area. When pressed (clicked) after the input of the handwritten character is confirmed, the search is started using the handwritten character as a search key. Here, the example in FIG. 12 is an example in which a small character string is entered, and the example in FIG. 13 is an example in which a large character string is entered.

FIG. 14 is a flowchart of a handwriting search area determination process in the handwritten character string search device 1 according to the second embodiment.
First, in step S14-1, when a handwritten character is input in the handwriting search character input area 12-1 and the search button 12-2 is pressed, a handwriting search area determination process is started and used by this process. Work area and buffer area are secured.

  Subsequently, in step S14-2, a process of reading height width information of an input area for handwritten characters is performed. In the present embodiment, the height width information of the input area frame of the search character input area 12-1 shown in FIG. 12 is read. If the input area is a full screen such as a PDA or pen computer, the size information of the full screen of the device is read. The process of step S14-2 is performed by the character string size determination unit 11-1.

  Subsequently, in step S14-3, a process of reading the size information of the input handwritten character string is performed. More specifically, the height and width information of the circumscribed rectangle of the locus of the input handwritten character string is read out. The processing in step S14-3 is performed by the character string size determination unit 11-1. In the present embodiment, the above method is simply used. However, by adding the height and width of the circumscribed rectangle of the handwriting locus of one character cut by the character cutting unit 3-1 in FIG. If the width is calculated, a more correct value is obtained.

Subsequently, in step S14-4, processing for calculating the relative size of the character string is performed. Specifically, the relative size of the character string is calculated by the following formula.
Relative size of character string = (height of character string × width of character string) / (height of input area × width of input area) × 100.
The process of step S14-4 is performed by the character string size determination unit 11-1.

  Subsequently, in step S14-5, it is checked whether or not the size of the character string is 80% or more. If it is 80% or more, the process proceeds to step S14-7, and if it is less than 80%, the process proceeds to step S14-6. The processing in step S14-5 is performed by the character string size determination unit 11-1.

  In step S14-6, it is checked whether the size of the character string is 60% or more. If it is 60% or more, the process proceeds to step S14-8, and if it is less than 60%, the process proceeds to step S14-9. The process of step S14-6 is performed by the character string size determination unit 11-1.

  In step S14-7, processing for setting the stored text search area to the Internet is performed. That is, when a large character string is written, a full text search is performed from within the Internet, and a process for finding a search result is performed. The process of step S14-7 is performed by the search area determination unit 11-2. In the present embodiment, the stored text search area is in the Internet. However, in the case of a company, for example, information search settings in branch offices around the world may be set in this setting.

  In step S14-8, a process for setting the stored text search area in the home server is performed. That is, when a character string having a standard size is written, a full text search is performed from within the home server to search for the corresponding file. The process of step S14-8 is performed by the search area determination unit 11-2. In the present embodiment, the stored text search area is in the home server. However, in the case of a company, for example, the information search setting in the department server may be set in this setting.

In step S14-9, a process for setting the stored text search area in the local disk is performed. That is, when a small character string is written, a full text search is performed from the local disk to search for the corresponding file. The process of step S14-9 is performed by the search area determination unit 11-2.

  In the present embodiment, an example has been described in which the input character string is discriminated into three types of sizes and three types of search ranges are set. However, the input character string size may be set to two types or may be set to two types by the input user. It is also possible to use four types or five types.

  In step S14-10, the handwriting search area determination process is terminated, the work area and the buffer area are released, and the set accumulated sentence search area is output to the accumulated sentence search process program.

  The subsequent search process itself is the same as a conventional process for searching a sentence file including a character string that matches a text string of a general recognition result by a full-text search, and thus detailed description of the process is omitted. .

  By performing the processing in this way, when a user inputs a character string of an arbitrary size at the time of input, a handwritten character search device that searches a search sentence from an accumulated sentence search area in a range corresponding to the size is realized. it can. In addition, since the input of two input items such as input of a search key and designation of a search range can be performed by a series of actions of inputting a handwritten character string, the search action is simplified.

(Third embodiment)
In the handwritten character string search device according to the second embodiment, the example in which the accuracy information at the time of character input and the character size information are used as control information at the time of search has been shown. Exists. When a human inputs a character by handwriting, an accurately memorized character can be evenly written, but when inputting an ambiguous character, the input often takes time.

  Therefore, in the handwritten character string search device according to the third embodiment, the input time of the input character is measured, and search result candidates are determined according to the input time of the input character.

  FIG. 15 is a block diagram illustrating a functional configuration of the handwritten character string search device 1 according to the third embodiment. The same components as those of the handwritten character string search device according to the first and second embodiments are denoted by the same reference numerals, and detailed description thereof is omitted.

  In FIG. 15, reference numeral 15-1 denotes an input character relative entry time determination unit that measures the input time of handwritten characters, and if there is a character with an input time extremely longer than the normal input time in the character string, the character Remember.

  In FIG. 15, reference numeral 15-2 denotes a search candidate range determination unit, in addition to characters of the recognition result of extremely long input time determined by the input character relative entry time determination unit 15-1, A process of adding the second candidate character is performed.

FIG. 16 is a flowchart of search key character string creation processing in the handwritten character string search device 1 according to the third embodiment.
First, in step S16-1, when a handwritten character is input in the handwriting search character input area and the search button is pressed, the input character relative time determination process is started, and the work area and buffer area used by this process are determined. Secured and initialized.

  Subsequently, in step S16-2, an average input time of input characters is calculated. Since the locus of the handwritten character is input from the coordinate input unit 2-1 in FIG. 15, the first pen-down time when the locus of the handwritten character is input is the input start time of the locus, and constitutes one character. The time when the last locus is pen-up is the input end time. The input time of the character is from the input start time of the first locus constituting one character to the input end time of the last locus. And each input time of each character is added. Thereafter, the longest input time is omitted, and the average input time is obtained by dividing by the number of (number of characters-1). The process of step S16-2 is performed by the input character relative entry time determination unit 15-1.

  Subsequently, in step S16-3, the input time of each input character is read. The process of step S16-3 is performed by the input character relative entry time determination unit 15-1.

  Subsequently, in step S16-4, processing for checking the input time of each character is performed. The process of step S16-4 is performed by the input character relative entry time determination unit 15-1. If the input time of the input character read in step S16-3 is more than twice the average input time obtained in step S16-2, the process proceeds to step S16-5, and if it is less than twice, the process proceeds to step S16-6. . In this embodiment, although it is calculated | required with the said method, since it may be unsuitable in the case of a character with many strokes, you may make it process considering the average input time for every stroke. In this case, the average input time of one stroke is obtained, the input time of the character is compared with the number of strokes of the character × the average input time, and when the input time of the character is more than twice the average time Proceed to step S16-5.

  In step S16-5, the character is stored as a character for a long time. The input characters whose input time is longer than usual are stored here. For example, an input character and a flag of the character are set in the search key buffer. The process of step S16-5 is performed by the input character relative entry time determination unit 15-1.

  In step S16-6, it is checked whether there is an unprocessed character of the handwriting search character. The process of step S16-6 is performed by the input character relative entry time determination unit 15-1. If there is an unprocessed character, the pointer is set to the next character and the process returns to step S16-4. If all the characters have been processed, the process proceeds to step S16-7.

  Subsequently, in step S16-7, processing for creating a search key character string from the recognition candidate and the search similar character is performed. The process of step S16-7 is performed by the search candidate range determination unit 15-2. In this embodiment, the handwritten character input in the normal input time is determined as the search key for the first candidate of the recognition result. In addition, handwritten characters input with a longer input time than usual are determined by using the first candidate of the recognition result, the second and subsequent candidates of the recognition result, and the character that is often mistaken for the search as a search key. Then, a search key character string is created using a combination of search key characters determined for each character.

  Here, for example, let us consider a case where the memory of whether it was “Shimomaruko's store” or “New Maruko's store” was ambiguous when the user entered a character to be searched. At this time, when “Shimomaruko store” is input by hand, “lower” takes 5 seconds to input, and when the input time of other characters “Marumaru store” is less than 1 second, “Maruko store” Is determined as the search key for the first candidate. In addition, “below” is determined by using the first candidate “below”, the second candidate “T”, and “new”, which is often mistaken for the search, as search keys. Then, when a search key character string is created with the confirmed search key, “Shimomaruko store”, “T Maruko store”, and “New Maruko store” are created as search key character strings. Then, the stored sentence file is searched with the created search key character string.

  Subsequently, in step S16-8, the process in the flowchart is terminated, the secured work area, buffer area, and the like are released, and the process proceeds to the next stored sentence search process.

  By performing the processing in this manner, when the user is confused about inputting a character at the time of input with ambiguous memory, it becomes possible to perform a search by setting a plurality of candidates for the character as a search target.

  Although each processing means of the handwritten character string search device according to the first embodiment, the second embodiment, and the third embodiment described above has been described as being independent from each other, in the present invention, all processing means have been described. The form of the handwritten character string search device provided with the processing means is also included. In this case, the search area is changed depending on the size of the character string, and many search results can be obtained when the input character is lost. Also, if the input character is input more precisely, the search result is narrowed down. Can be achieved. Moreover, since this function may not be necessary depending on the user, a configuration in which it is enabled or disabled by setting is also conceivable.

  2, FIG. 3, FIG. 11 and FIG. 15 constituting the handwritten character string search device according to each embodiment described above, and a control method (handwritten character string search method) of the handwritten character string search device Each of the steps in FIGS. 8 to 10, 14 and 16 can be realized by operating a program stored in a RAM or ROM of a computer. This program and a computer-readable storage medium storing the program are included in the present invention.

  Specifically, the program is recorded in a storage medium such as a CD-ROM, or provided to a computer via various transmission media. As a storage medium for recording the program, a flexible disk, a hard disk, a magnetic tape, a magneto-optical disk, a nonvolatile memory card, and the like can be used in addition to the CD-ROM. On the other hand, as the transmission medium of the program, a communication medium (wired line such as an optical fiber, etc.) in a computer network (LAN, WAN such as the Internet, wireless communication network, etc.) system for propagating and supplying program information as a carrier wave A wireless line or the like.

  Further, by executing the program supplied by the computer, not only the handwritten character string search device according to each of the above-described embodiments is realized, but also the OS (operating system) or other operating system in which the program is running on the computer. When the functions of the handwritten character string search device according to each embodiment described above are realized in cooperation with application software, etc., or all or part of the processing of the supplied program is a function expansion board or function expansion unit of a computer Such a program is also included in the present invention when the functions of the handwritten character string search device according to each of the embodiments described above are realized.

It is a schematic block diagram of the handwritten character string search system which concerns on embodiment of this invention. It is a block diagram which shows the system configuration | structure of the handwritten character string search apparatus which concerns on embodiment of this invention. It is a block diagram which shows the function structure of the handwritten character string search apparatus which concerns on 1st Embodiment. It is a figure which shows the input screen of the handwritten character of a search key. It is a figure which shows the screen after the recognition process of the handwritten character shown in FIG. It is a figure which shows the relationship between one input character and the standard pattern memorize | stored in the handwritten character recognition dictionary. It is a figure which shows the relationship between one input character and the standard pattern memorize | stored in the handwritten character recognition dictionary. It is a flowchart of the handwriting search key input process of the handwritten character string search device which concerns on 1st Embodiment. In the handwritten character string search device according to the first embodiment, it is a flowchart of processing for determining a search candidate range from a character string of a recognition result and character accuracy information. In the handwritten character string search device which concerns on 1st Embodiment, it is a flowchart of the text search process in a search candidate range. It is a block diagram which shows the function structure of the handwritten character string search device 1 which concerns on 2nd Embodiment. It is a figure which shows the example of input of a handwritten search character. It is a figure which shows the example of input of a handwritten search character. It is a flowchart of the handwriting search area determination process in the handwritten character string search device 1 which concerns on 2nd Embodiment. It is a block diagram which shows the function structure of the handwritten character string search device 1 which concerns on 3rd Embodiment. It is a flowchart of the character string creation process of the search key in the handwritten character string search device 1 which concerns on 3rd Embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Handwritten character string search device 2 Liquid crystal display unit 2-1 Coordinate input unit 3 LAN cable 4 Antenna for wireless communication 5 Personal computer 6 CPU
7 RAM
8 ROM
9 HD
10 LAN interface 11 System bus 12 Wireless communication card 3-1 Character cutting unit 3-2 Handwritten character recognition dictionary 3-3 Handwritten character recognition unit 3-4 Handwritten character accuracy determining unit 3-5 Search candidate range determining unit 3 -6 Text search part

Claims (7)

Handwriting input means for inputting handwritten characters;
Handwritten character recognition means for recognizing a handwritten character input from the handwritten character input means;
A degree of coincidence determination means for determining a degree of coincidence between the handwritten character recognized by the handwritten character recognition means and the standard font of the character;
Search candidate range determining means for determining a search candidate range based on the matching degree determined by the matching degree determining means;
A handwritten character string search device, comprising: search means for searching a character string file based on the search candidate range for the handwritten character recognized by the handwritten character recognition means, and displaying the search result. Handwriting input means for inputting handwritten characters;
Handwritten character recognition means for recognizing a handwritten character input from the handwritten character input means;
A size determining means for determining the size of the handwritten character input from the handwritten character input means from a relative relationship to the input screen;
Search area determining means for determining a search area based on the size of the handwritten character determined by the size determining means;
A handwritten character string search device comprising: search means for searching a character string file based on the search area for the handwritten character recognized by the handwritten character recognition means and displaying the search result. Handwriting input means for inputting handwritten characters;
Handwritten character recognition means for recognizing a handwritten character input from the handwritten character input means;
An input time detecting means for detecting an input time of a handwritten character input from the handwritten character input means;
Search candidate range determining means for determining a search candidate range based on the input time of handwritten characters detected by the input time detecting means;
A handwritten character string search device, comprising: search means for searching a character string file based on the search candidate range for the handwritten character recognized by the handwritten character recognition means, and displaying the search result. A handwritten character string search method in a handwritten character string search device comprising a handwritten character input means for inputting a handwritten character,
A handwritten character recognition step for recognizing a handwritten character input from the handwritten character input means;
A degree of matching determination step for determining the degree of matching between the handwritten character recognized in the handwritten character recognition step and the standard font of the character;
A search candidate range determination step for determining a search candidate range based on the matching degree determined by the matching degree determination step;
A handwritten character string search method comprising: a search step of searching for a character string file based on the search candidate range for the handwritten character recognized by the handwritten character recognition means, and displaying the search result. A handwritten character string search method in a handwritten character string search device comprising a handwritten character input means for inputting a handwritten character,
A handwritten character recognition step for recognizing a handwritten character input from the handwritten character input means;
A size determining step of determining the size of the handwritten character input from the handwritten character input means from a relative relationship to the input screen;
A search area determining step for determining a search area based on the size of the handwritten character determined in the size determining step;
A handwritten character string search method comprising: a search step of searching for a character string file based on the search area for the handwritten character recognized in the handwritten character recognition step, and displaying the search result. A handwritten character string search method in a handwritten character string search device comprising a handwritten character input means for inputting a handwritten character,
A handwritten character recognition step for recognizing a handwritten character input from the handwritten character input means;
An input time detecting step for detecting an input time of a handwritten character input from the handwritten character input means;
A search candidate range determination step for determining a search candidate range based on the input time of the handwritten characters detected in the input time detection step;
A handwritten character string search method comprising: a search step of searching a character string file for the handwritten character recognized in the handwritten character recognition step based on the search candidate range and displaying the search result.   The program for making a computer perform the handwritten character string search method of any one of Claims 4-6.

Images