JP7490861B2 - Character candidate suggestion device, handwritten character identification system, handwritten character identification method and program - Google Patents

Description

The present invention relates to a character candidate suggestion device, a handwritten character recognition system, a handwritten character recognition method, and a program.

Patent document 1 describes that "the information processing device includes a character extraction unit 224 that extracts a character string from image data, a list creation unit 244 that creates a conversion candidate list including one or more conversion candidate characters corresponding to an input character, an identification unit 248 that identifies a conversion target character based on a comparison between the one or more conversion candidate characters included in the conversion candidate list and the character string extracted by the character extraction unit, and a display control unit 252 that generates a display screen including the conversion target character identified by the identification unit."

JP 2018-163413 A

The above technology can present character candidates, but if the characters are not recognized correctly, it is difficult to obtain appropriate results.

The objective of the present invention is to provide a technology that more appropriately suggests similar string candidates.

The present application includes a number of means for solving at least part of the above-mentioned problems, examples of which are as follows: An information processing device according to one aspect of the present invention includes a similar character storage unit that stores similarity relationships in the appearance of characters, an existing character storage unit that stores existing character strings, a request processing unit that receives one or more characters and accepts a request to devise character candidates similar in appearance to the characters, a matching processing unit that identifies character candidates similar in appearance to the received character by matching them with the similar character storage unit and matches combinations of the candidate characters similar in appearance with the existing character strings stored in the existing character storage unit, and a candidate character transmission processing unit that outputs the combinations of candidate characters similar in appearance that are found as a result of matching by the matching processing unit as candidate characters.

For example, the character candidate suggestion device may also include a similar character registration processing unit that, after outputting the candidate characters, when the entrusted processing unit receives a character similar to any of the characters received, stores the similar character in the similar character storage unit if the similarity relationship is not stored in the similar character storage unit.

For example, in the character candidate suggestion device described above, the existing character storage unit may store character strings of existing personal names, the request processing unit may be commissioned to suggest personal names as the character candidates to be suggested, and the candidate character transmission processing unit may output combinations of candidate characters having similar appearances as candidate characters for personal names.

For example, the character string of the existing person's name may be divided into a surname and a given name, and the device may include a name discrimination processing unit that divides the received characters into a surname and a given name and causes the matching processing unit to match them.

For example, the character candidate suggestion device may be characterized in that the first name discrimination processing unit changes the position separating the first name and last name for the received characters to identify combinations of first names and last names, and when a match is obtained as a result of matching by the matching processing unit for both the first name and last name for each combination, the combination is designated as a candidate character.

For example, in the above character candidate suggestion device, the candidate character transmission processing unit may be characterized in that it outputs the candidate characters after arranging them according to their frequency of appearance.

For example, the character candidate suggestion device described above may be characterized in that the characters received by the processing unit are characters that have been recognized by reading handwritten characters.

In addition, a handwritten character identification system according to another aspect of the present invention is a handwritten character identification system including a character candidate suggestion device, a business terminal communicably connected to the character candidate suggestion device via a network, and a reading device that reads handwritten characters in response to instructions from the business terminal, the business terminal including a character recognition unit that acquires the handwritten characters from the reading device and performs character recognition, a commissioning processing unit that transmits the characters recognized by the character recognition unit to the character candidate suggestion device and commissions the suggestion of character candidates, and a visual confirmation processing unit that displays the character candidates output from the character candidate suggestion device, and the character candidate suggestion device is configured to The system is characterized by comprising a similar character storage unit that stores the similarity relationship of character appearances, an existing character storage unit that stores existing character strings, a request processing unit that receives one or more characters and accepts a request to create character candidates similar in appearance to the received characters, a matching processing unit that identifies character candidates similar in appearance to the received characters by matching them with the similar character storage unit and matches combinations of the candidate characters similar in appearance with the existing character strings stored in the existing character storage unit, and a candidate character transmission processing unit that outputs the combinations of candidate characters similar in appearance that are found as a result of matching by the matching processing unit as candidate characters.

In addition, a handwritten character identification method according to another aspect of the present invention is a handwritten character identification method using a handwritten character identification system including a character candidate suggestion device, a business terminal communicably connected to the character candidate suggestion device via a network, and a reading device that reads handwritten characters in response to instructions from the business terminal, in which the business terminal executes a character recognition step of acquiring the handwritten characters from the reading device and performing character recognition, a commissioning processing step of transmitting the characters recognized in the character recognition step to the character candidate suggestion device and commissioning the suggestion of character candidates, and a visual confirmation processing step of displaying the character candidates output from the character candidate suggestion device, and The device includes a similar character storage unit that stores the similarity relationship between the appearances of characters, and an existing character storage unit that stores existing character strings, and executes an order processing step of receiving one or more characters and accepting the creation of character candidates similar in appearance to the received character, a matching processing step of identifying character candidates similar in appearance to the received character by matching them with the similar character storage unit, and matching combinations of the candidate characters similar in appearance with the existing character strings stored in the existing character storage unit, and a candidate character transmission processing step of outputting the combination of candidate characters similar in appearance that is found as a result of the matching in the matching processing step as candidate characters.

In addition, a program according to another aspect of the present invention is a program for making a computer perform a process to assist in the identification of handwritten characters, the computer comprising a similar character storage unit that stores similarity relationships in the appearance of characters, an existing character storage unit that stores existing character strings, and a processing unit, and is characterized in that the processing unit performs an order processing step of receiving one or more characters and accepting the creation of character candidates having a similar appearance to the received character, a matching processing step of identifying character candidates having a similar appearance to the received character by matching them with the similar character storage unit, and matching combinations of the candidate characters having a similar appearance with the existing character strings stored in the existing character storage unit, and a candidate character transmission processing step of outputting the combinations of candidate characters having a similar appearance that are found as a result of the matching in the matching processing step as candidate characters.

The present invention provides a technology that can appropriately evaluate the content of a text.

Issues, configurations, and advantages other than those described above will become clear from the description of the embodiments below.

1 is a block diagram illustrating an overview of a handwriting identification system according to an embodiment. FIG. 1 is a diagram illustrating an example of the configuration of a handwriting recognition system. FIG. 13 illustrates an example of a data structure of a receipt data storage unit; FIG. 13 is a diagram illustrating an example of a data structure of a surname storage unit. FIG. 4 illustrates an example of a data structure of a name storage unit; 13 is a diagram illustrating an example of a data structure of a similar character storage unit; 13 is a diagram illustrating an example of a data structure of a candidate character storage unit; FIG. 4 illustrates an example of a data structure of a character storage unit; FIG. 4 is a diagram illustrating an example of a data structure of an image storage unit. FIG. 2 is a diagram illustrating an example of a hardware configuration of a character candidate suggestion device. FIG. 1 is a diagram illustrating an example of an overall processing flow of handwriting recognition. FIG. 13 is a diagram illustrating an example of a flow of a first name category matching process. FIG. 13 is a diagram illustrating an example of the flow of a candidate character receiving process. FIG. 11 is a diagram illustrating an example of a flow of a similar character registration process. FIG. 13 is a diagram showing an example of a candidate presentation screen. FIG. 13 is a diagram showing an example of surname candidate presentation.

A handwriting recognition system 1 to which an embodiment according to one aspect of the present invention is applied will be described below with reference to the drawings. In the following embodiment, when necessary for convenience, the description will be divided into multiple sections or embodiments. However, unless otherwise specified, they are not unrelated to each other, and one is a partial or complete modification, detail, supplementary explanation, etc. of the other.

In addition, in the following embodiments, when referring to the number of elements (including the number, numerical value, amount, range, etc.), unless otherwise specified or clearly limited in principle to a specific number, the number is not limited to that specific number and may be more than or less than the specific number.

Furthermore, it goes without saying that in the following embodiments, the components (including element steps, etc.) are not necessarily essential unless specifically stated otherwise or considered to be clearly essential in principle.

Similarly, in the following embodiments, when referring to the shapes, positional relationships, etc. of components, etc., it is intended to include shapes that are substantially similar or similar to those, unless otherwise specified or considered to be clearly different in principle. The same applies to the above numerical values and ranges.

In addition, in all drawings used to explain the embodiments, the same components are generally given the same reference numerals, and repeated explanations will be omitted.

FIG. 1 is a block diagram illustrating an example of an overview of a handwriting identification system according to this embodiment. A user uses a business terminal 200 to connect to a character candidate suggestion device 100 via a browser or the like to use the handwriting identification system 1, but this is not limited thereto, and each function of the character candidate suggestion device 100 may be installed on the business terminal 200 as application software.

When connecting from the business terminal 200 to the character candidate suggestion device 100, the connection is made via a network 50 which is a communication network such as a LAN (Local Area Network), a WAN (Wide Area Network), the Internet, a mobile phone network, or a combination of these. The network may be a VPN (Virtual Private Network) on a wireless communication network such as a mobile phone communication network.

Example applications of the handwriting identification system 1 include business systems that handle specific documents (such as insurance applications and bank account opening applications, where names are entered by hand according to a specific format).

In this case, it is expected that the user will handwrite their name in a designated field on a designated document, but in order to process a large amount of documents, it is more efficient to process large amounts of information using a computer, so generally handwritten notations such as names are often digitized using OCR (Optical Character Recognition).

Although technology for digitizing character data from paper documents has improved, there is still a certain percentage of misreadings and non-readings, and no technology exists that can perfectly recognize all handwritten characters. As a result, characters must ultimately be checked visually by a human to confirm match, but this visual checking is one of the factors that hinders reductions in the cost of digitizing document character data. Specifically, when a name is read, a system has been introduced that compares it with a database of names such as first and last names to improve character recognition rates, but some characters are difficult to read or ambiguous, and visual checks to see if there have been any misreadings are still costly.

In other words, if the scanned characters contain a misreading, for example if a character recognition engine such as an OCR misreads the correct surname "Kawamoto" as "Ogi," the incorrect "Ogi" will be compared with each database, and the correct "Kawamoto" may be ranked lower or missed as a candidate.

Also, when using a format in which the name is entered as the full name without clearly separating the spaces for the surname and given name, a misreading may not even be noticed unless the correct surname is identified. For example, if the "田" in the full name "Oyamada Shin" is misread as "由", the surname may be recognized as "Koyama" and a name beginning with "由" such as "Yoshinobu" may be selected as a candidate.

In light of these issues, this application assumes that characters read by a character recognition engine such as OCR may be incorrect, and in addition to comparing with a database of surnames and given names, compares with similar characters that have been frequently misread by a character recognition engine such as OCR in the past. This appropriately widens the range of candidates for read characters to include characters that are prone to misreading, and also makes characters that may be inferior or have been omitted from the list of candidates, and selects candidate characters after performing a process to distinguish between surnames and given names. Furthermore, the aim is to improve the accuracy of character recognition by allowing new similar character information to be constantly accumulated through the visual confirmation of characters.

The business terminal 200 according to this embodiment is a terminal used for the business of digitizing such specified documents. The reading device 300 is a scanner, an imaging device, or the like, and reads the specified document as an image or the like in response to an instruction from the business terminal 200, and transfers the read handwritten image to the business terminal 200.

In this business, when the business terminal 200 receives a handwritten character image, it digitizes it using OCR and converts it into character string information, transmits the converted character string to the character candidate suggestion device 100, and receives candidate characters proposed by the character candidate suggestion device 100. The business terminal 200 then selectively displays the candidate characters, and uses the candidate characters selected by the user as digitized information.

The reading device 300 is not limited to a scanner, but may be a device capable of reading images as data, such as a mobile phone terminal with a camera or a tablet terminal with a camera, and may be configured to be communicably connected to the business terminal 200 via the network 50.

In addition, the reading device 300 is not a required device, and the character candidate suggestion device 100 or the business terminal 200 may receive image data that has already been read or data that has undergone character recognition from other devices via the network 50 and suggest character candidates.

Figure 2 is a diagram showing an example of the configuration of a handwritten character identification system. In the character candidate suggestion device 100, a memory unit 110, a processing unit 120, and a communication unit 130 are connected to each other so that they can communicate with each other via a bus or the like.

The storage unit 110 includes a received data storage unit 111, an existing character storage unit 112, a similar character storage unit 115, and a candidate character storage unit 116. The existing character storage unit 112 stores character strings such as existing personal names and proper nouns. For example, the existing character storage unit 112 includes a surname storage unit 113 and a given name storage unit 114. For example, it is said that there are more than 300,000 different Japanese surnames, and surnames whose existence has been confirmed are stored in the surname storage unit 113.

Figure 3 is a diagram showing an example of the data structure of the received data storage unit. The received data storage unit 111 includes an acceptance date 111A, a requesting company ID 111B, a document number 111C, a character identification number 111D, and a recognized character code 111E. The acceptance date 111A is information that identifies the date on which a request for character candidate proposals was accepted from the business terminal 200. The requesting company ID 111B is information that identifies the company that requested the character candidate proposals. The document number 111C is information that identifies the document (document image data) that was the subject of the request for character candidate proposals. The character identification number 111D is information that identifies each character included in the document to be recognized. The recognized character code 111E is the character code of the character obtained as a result of recognizing each character from its appearance.

Figure 4 is a diagram showing an example of the data structure of the surname storage unit. The surname storage unit 113 includes a surname identifier 113A, a surname 113B, a character count 113C, and an occurrence frequency 113D. The surname identifier 113A is information that distinguishes a surname from other surnames. The surname 113B is spelling information for the surname. The character count 113C is the number of characters in the spelling of the surname. The occurrence frequency 113D is the frequency of occurrence within the entire surname.

Figure 5 is a diagram showing an example of the data structure of the name storage unit. The name storage unit 114 includes a name identifier 114A, a name 114B, a character count 114C, and an occurrence frequency 114D. The name identifier 114A is information that distinguishes a name from other names. The name 114B is information on the notation of the name. The character count 114C is the number of characters in the notation of the name. The occurrence frequency 114D is the frequency of occurrence within the entire name.

Figure 6 is a diagram showing an example of the data structure of the similar character storage unit. The similar character storage unit 115 includes character code 115A, similar character code 115B, occurrence frequency 115C, and storage date 115D. Character code 115A is the code of a reference character among characters that are similar to each other. Similar character code 115B is the code of a character that is similar in appearance to the character identified by character code 115A. Occurrence frequency 115C is the frequency with which the character appears as a similar character in use. Storage date 115D is information specifying the date on which the character was stored as a similar character.

Figure 7 is a diagram showing an example of the data structure of the candidate character storage unit. The candidate character storage unit 116 includes a character identification number 116A and a candidate character code 116B. The character identification number 116A is information that identifies an individual character included in the document to be recognized. The candidate character code 116B is a character candidate that is different from the character recognized by OCR, and is a code for a similar and existing character.

The processing unit 120 includes an acceptance processing unit 121, a name discrimination processing unit 122, a matching processing unit 123, a similar character registration processing unit 124, and a candidate character transmission processing unit 125.

The acceptance processing unit 121 accepts a request for proposal of character candidates from the business terminal 200. When the name discrimination processing unit 122 receives a proposal of character candidates of a plurality of character strings, it identifies a combination of a surname and a given name separated by changing the division position of the name one character at a time and judges whether it is valid as a name. The matching processing unit 123 checks against the existing character storage unit 112 to judge whether it is an existing character candidate. The similar character registration processing unit 124 registers the characters reported from the business terminal 200 as unregistered similar relationships in the similar character storage unit 115. The candidate character transmission processing unit 125 outputs the combination of candidates of characters with similar appearances that are hit as a result of the matching by the matching processing unit 123 as candidate characters. Specifically, the candidate character transmission processing unit 125 realizes the idea by outputting (transmitting) the candidates stored in the candidate character storage unit 116 to the business terminal 200. The communication unit 130 communicates with other devices via the network 50.

The business terminal 200 has a memory unit 210, a processing unit 220, and a communication unit 230, which are connected to each other so that they can communicate with each other via a bus or the like.

The memory unit 210 includes a character memory unit 211, a candidate character memory unit 212, and an image memory unit 213.

Figure 8 is a diagram showing an example of the data structure of the character storage unit. The character storage unit 211 includes an image data number 211A, a reading date 211B, a character identification number 211C, and an image 211D.

Image data number 211A is information that identifies the image of the character string read by reading device 300. Reading date 211B is information that identifies the date on which reading device 300 read the document. Character identification number 211C is information that identifies individual characters contained in the document to be recognized. Image 211D is information that identifies image data of a single character identified by character identification number 211C (part of the image read by reading device 300).

Figure 9 is a diagram showing an example of the data structure of the image storage unit. The image storage unit 213 includes an image data number 213A, a document number 213B, a read date 213C, and an image 213D.

Image data number 213A is information that identifies the image of the character string read by reading device 300. Document number 213B is information that identifies the document (document image data) for which a request for character candidate proposals was made. Reading date 213C is information that identifies the date on which reading device 300 read the document. Image 213D is information that identifies the image read by reading device 300.

The processing unit 220 includes a character recognition unit 221, a commission processing unit 222, a visual confirmation processing unit 223, and a similar character report processing unit 224. The character recognition unit 221 is a so-called OCR engine, and analyzes handwritten character images acquired from the reading device 300 to recognize characters.

The commissioning processing unit 222 commissions the character candidate suggestion device 100 to propose candidate characters with high accuracy in order to eliminate rework and procedural errors caused by misreading by the character recognition unit 221.

The visual confirmation processing unit 223 selectively displays each of the character candidates output from the character candidate suggestion device 100 so that they can be visually confirmed, and accepts the user's selection of one of the candidates.

The similar character report processing unit 224 accepts input of characters that the user recognizes as similar characters and reports to the character candidate suggestion device 100 that the characters should be registered as similar characters.

The communication unit 230 communicates with other devices via the network 50.

In addition, the business terminal 200 includes an input unit (not shown) that accepts input from a user, and an output unit that outputs various output information such as a screen or a report. For example, the input unit accepts various types of input, such as various contact inputs such as typing, touching, or flick input, or voice input or gaze input.

Figure 10 is a diagram showing an example of the hardware configuration of character candidate suggestion device 100. Character candidate suggestion device 100 has a hardware configuration realized by the housing of a so-called server device, workstation, personal computer, smartphone, or tablet terminal. Character candidate suggestion device 100 has a calculation device 101, a main memory device 102, an auxiliary memory device 103, a communication device 104, and a bus 107 connecting each device. The same is true for business terminal 200. In addition, business terminal 200 has input/output devices such as a touch panel, keyboard, microphone, and display.

The calculation device 101 is, for example, a calculation device such as a CPU (Central Processing Unit).

The main memory device 102 is a memory device such as a RAM (Random Access Memory).

The auxiliary storage device 103 is a non-volatile storage device capable of storing digital information, such as a hard disk drive, a solid state drive (SSD), or a flash memory.

The communication device 104 is a wired communication device such as a network card, or a wireless communication device.

The above-mentioned entrustment processing unit 121, name discrimination processing unit 122, matching processing unit 123, similar character registration processing unit 124, and candidate character transmission processing unit 125 of the character candidate suggestion device 100 are realized by a program that causes the calculation device 101 to perform processing. This program is stored in the main memory device 102, the auxiliary memory device 103, or a ROM device not shown, and is loaded onto the main memory device 102 for execution and executed by the calculation device 101.

The memory unit 110 of the character candidate suggestion device 100 is realized by the main memory device 102 and the auxiliary memory device 103. The communication unit 130 is realized by the communication device 104. The above is an example of the hardware configuration of the character candidate suggestion device 100.

The configuration of the character candidate suggestion device 100 can be further classified into more components depending on the processing content. Also, a single component can be classified to perform more processes.

In addition, each processing unit (the contract processing unit 121, the name discrimination processing unit 122, the matching processing unit 123, the similar character registration processing unit 124, and the candidate character transmission processing unit 125) may be constructed using dedicated hardware (ASIC, GPU, etc.) that realizes each function. In addition, the processing of each processing unit may be executed by a single piece of hardware, or may be executed by multiple pieces of hardware.

Next, the operation of the handwriting recognition system 1 in this embodiment will be described.

Figure 11 is a diagram showing an example of the overall processing flow of handwriting recognition. The handwriting recognition process is started when the user issues a start command from the business terminal 200.

First, the business terminal 200 requests the reading device 300 to read the image data of the document (step S10). Specifically, the character recognition unit 221 instructs the reading device 300 to start scanning or photographing.

Upon receiving a request from the business terminal 200, the reading device 300 reads the image of the document, creates image data, and responds to the business terminal 200 (step S11).

The character recognition unit 221 of the business terminal 200 receives image data from the reading device 300, stores it in the image storage unit 213, and executes character recognition processing with the OCR engine (step S12). The recognized character data is then stored in the character storage unit 211 (step S13). Note that this OCR engine may be implemented in a PC terminal, a server terminal, or the like that is placed on a network to which the business terminal 200 can be connected, and may be configured to execute character recognition processing in response to an instruction from the business terminal 200.

Then, in the business terminal 200, the commissioning processing unit 222 transmits the recognized characters to the character candidate suggestion device 100 and commissions the process of suggesting character candidates (step S14).

Then, the entrusted processing unit 121 of the character candidate suggestion device 100 receives the recognized characters from the business terminal 200 and stores them in the received data storage unit 111 (step S15).

Then, the matching processing unit 123 matches the received character data with the existing character storage unit 112 and the similar character storage unit 115 (step S16).

Then, the matching processing unit 123 selects candidate characters and stores them in the candidate character storage unit 116 (step S17).

Then, the candidate character transmission processing unit 125 transmits the candidate characters to the business terminal 200 (step S18).

The commissioning processing unit 222 of the business terminal 200 receives the candidate characters and stores them in the candidate character storage unit 212 (step S19).

Then, the visual confirmation processing unit 223 outputs a screen that compares the candidate characters with the image in the image storage unit (step S20). More specifically, for example, the screen displays the handwritten character image read by the reading device 300 and a plurality of candidate characters that can be selected and input.

When an instruction to register similar characters is received from a user, the similar character report processing unit 224 accepts input of similar single characters other than the candidate characters if any, and responds to the character candidate suggestion device 100 (step S21).

The similar character registration processing unit 124 of the character candidate suggestion device 100 receives the similar characters and stores them in the similar character storage unit 115 if they have not been stored therein yet (step S22).

The above is an example of the overall processing flow for handwriting recognition. This overall processing for handwriting recognition makes it possible to more appropriately suggest similar character string candidates.

Figure 12 is a diagram showing an example of the flow of the first name segment matching process. The first name segment matching process is an example of a process that performs steps S16 to S18 as a series in the overall process flow of handwriting recognition.

First, the name discrimination processing unit 122 determines the total number of characters n (step S1601). For example, if the name "Oyamada Shin" is a character string for which candidates are to be presented, the name discrimination processing unit 122 determines the total number of characters n to be "4", with each of the kanji characters "小", "山", "田", and "進" being one character.

Then, the name determination processing unit 122 breaks down the name into d characters (d is 1 or more and less than n, initial value 1) from the beginning as the family name and the remaining characters as the given name (step S1602).

The matching processing unit 123 breaks down the characters that make up the surname character by character and matches them with the similar character storage unit 115, and matches combinations of similar characters with the surname storage unit 113 to select real surname candidates (step S1603).

Then, the name discrimination processing unit 122 determines whether there is no matching surname candidate in the similar character storage unit 115 (step S1604). If there is no surname candidate ("Yes" in step S1604), the name discrimination processing unit 122 advances control to step S1611, which will be described later.

If there is a surname candidate (step S1604: No), the surname discrimination processing unit 122 determines whether there is at least one corresponding surname candidate in the surname storage unit 113 (step S1605). If there is no corresponding surname candidate in the surname storage unit 113 (step S1605: No), the surname discrimination processing unit 122 advances control to step S1611, which will be described later.

If there is a matching surname in the surname storage unit 113 for the surname candidates ("Yes" in step S1605), the name discrimination processing unit 122 temporarily stores all selected surname candidates (step S1606).

Then, the matching processing unit 123 breaks down the characters that make up the name character by character and matches them with the similar character storage unit 115, and matches combinations of similar characters with the name storage unit 114 to select actual name candidates (step S1607).

Then, the name discrimination processing unit 122 determines whether there are no matching name candidates in the similar character storage unit 115 (step S1608). If there are no matching name candidates ("Yes" at step S1608), the name discrimination processing unit 122 advances control to step S1611, which will be described later.

If there is a name candidate (step S1608: No), the name determination processing unit 122 determines whether there is at least one name candidate that matches the name candidate in the name storage unit 114 (step S1609). If there is no name candidate that matches the name storage unit 114 (step S1609: No), the name determination processing unit 122 advances control to step S1611, which will be described later.

If there is a name that matches the name candidates in the name storage unit 114 ("Yes" at step S1609), the name determination processing unit 122 temporarily stores all selected name candidates (step S1610).

Then, the name discrimination processing unit 122 increments (+1) the number of characters in the family name d (step S1611).

Then, the name determination processing unit 122 determines whether the number of characters in the family name, d, matches the total number of characters, n, (step S1612). If they do not match (step S1612: No), the name determination processing unit 122 returns control to step S1602.

If the number of characters in the surname d matches the total number of characters n (if "Yes" in step S1612), the name discrimination processing unit 122 determines whether there are any pairs of temporarily stored surname candidates and given name candidates (where the surname length + given name length matches the total number of characters n), i.e., whether there are any combinations of surnames and given names that result in a hit (step S1613).

If there is no temporarily stored pair of surname and given name candidates (where the surname length + given name length matches the total number of characters n) ("No" in step S1613), the candidate character transmission processing unit 125 notifies the business terminal 200 that it is unable to distinguish (step S1614).

If there is a pair of temporarily stored surname candidates and given name candidates (where the surname length + given name length is equal to the total number of characters n) (if "Yes" in step S1613), the candidate character transmission processing unit 125 transmits the character candidates to the business terminal 200 (step S1801). At that time, the candidate character transmission processing unit 125 sorts the temporarily stored pairs of surname candidates and given name candidates in order of frequency of occurrence and transmits them. In other words, the candidate character transmission processing unit 125 transmits the surname candidates and given name candidates in the order of characters with the highest frequency of occurrence.

The above is an example of the flow of the name segment matching process. According to the name segment matching process, for a name consisting of one or more characters, the segment position of the name can be changed and similar characters can be applied to identify existing candidates. For example, in the case of the name "Oyamada Shin" mentioned above, "田" and "由" are similar characters, and candidate characters can be identified that include the two-character surname "Oyama" and the remaining part of the given name "Yoshinobu", and the three-character surname "Oyamada" and the remaining part of the given name "Nobu".

Figure 13 is a diagram showing an example of the flow of the candidate character reception process. The candidate character reception process is an example of a process that performs steps S19 to S21 as a series in the overall process flow of handwriting recognition.

First, the entrustment processing unit 222 stores the received candidate characters in the candidate character storage unit 212 (step S191).

Then, the visual confirmation processing unit 223 displays the candidate characters and the handwritten character image for confirmation (step S201). An example of the confirmation display screen will be described later.

Then, the visual confirmation processing unit 223 accepts the selection input of the candidate character (step S202). This confirms the digitization of the handwritten character.

The similar character report processing unit 224 also determines whether or not new similar characters have been input (step S211). If no new similar characters have been input (step S211: No), the similar character report processing unit 224 ends the candidate character reception process.

If new similar characters are input (if "Yes" in step S211), the similar character report processing unit 224 sends information on the new similar characters to the character candidate suggestion device 100 (step S212).

The above is an example of the flow of the candidate character receiving process. According to the candidate character receiving process, candidate characters are received from the character candidate suggestion device 100, and it is possible to select and input a candidate while comparing it with the handwritten character. Furthermore, if a similar character is noticed, it is possible to report this to the character candidate suggestion device 100 and request that it be newly registered as a similar character.

Figure 14 is a diagram showing an example of the flow of the similar character registration process. The similar character registration process is started by the character candidate suggestion device 100 when information that becomes a similar character in the candidate character reception process is transmitted from the business terminal 200 to the character candidate suggestion device 100.

First, the similar character registration processing unit 124 determines whether the received similar character information is already stored in the similar character storage unit 115 (step S2201).

If the received similar character information is not already stored in the similar character storage unit 115 (step S2201: No), the similar character registration processing unit 124 identifies one-to-one combinations of similar characters (breaks down characters that have a similar relationship into one-to-one combinations) (step S2202).

Then, the similar character registration processing unit 124 stores both of the similar characters as similar to each other in the similar character storage unit 115 (step S2203).

If the received similar character information has already been stored in the similar character storage unit 115 (step S2201: Yes), the similar character registration processing unit 124 notifies the business terminal 200 that the information has already been registered (step S2204).

Then, the similar character registration processing unit 124 increases the frequency of occurrence of each of the similar characters by a predetermined amount (step S2205).

The above is an example of the flow of the similar character registration process. According to the similar character registration process, if the similar characters requested to be registered from the business terminal 200 are new, they are registered as similar, and if they have already been registered, their frequency of appearance can be increased.

Figure 15 is a diagram showing an example of a candidate presentation screen. Candidate presentation screen 400 is a screen that is output in step S20 of the handwriting recognition overall processing flow and in step S201 of the candidate character reception processing. Original data (handwritten character image read by reading device 300) to be visually confirmed is displayed in original data display area 401, and candidate characters to be visually confirmed are displayed in surname candidate display area 402 and given name candidate display area 403. Candidate characters proposed by character candidate suggestion device 100 as candidates with the highest occurrence rate are displayed in surname candidate display area 402 and given name candidate display area 403. When input is received, surname candidate decision button 404 confirms the surname displayed in surname candidate display area 402. When input is received, name candidate decision button 405 confirms the name displayed in name candidate display area 403.

The candidate characters from the next candidate onwards are arranged in order of frequency of appearance and displayed in a selectable list in the surname next candidate display area 406 and the first name next candidate display area 407. The surname next candidate display area 406 and the first name next candidate display area 407 accept the selection of listed characters. When input is received, the surname next candidate confirmation button 408 confirms the surname selected in the surname next candidate display area 406. When input is received, the first name next candidate confirmation button 409 confirms the name selected in the first name next candidate display area 407.

Similar character registration area 410 and similar character registration area 411 each accept character input. When registration button 412 accepts input, it accepts the characters entered in similar character registration area 410 and similar character registration area 411 as characters that are similar to each other, and starts the registration process in character candidate suggestion device 100. That is, if the visual inspection result indicates that there is a character that the visual inspection worker (user) feels needs to be registered as a new similar character, the character can be registered by entering the character in similar character registration area 410 and similar character registration area 411 and selecting registration button 412. In addition, if input is accepted to back button 413 or next screen button 414 as necessary, a transition to the previous or next business screen occurs.

Figure 16 is a diagram showing an example of surname candidate presentation. In the handwriting identification system 1 to which this embodiment is applied, even if the reading device 300 and the character recognition unit 221 make a misreading, there is a high possibility that the correct character will be included in the presented candidates, and this is specifically shown.

If a character recognition engine such as OCR misreads the correct surname "Kawamoto" (a) as "Ogi" (b), when comparing the misread "Ogi" (b) with existing surnames, the correct character candidate "Kawamoto" (a) may be ranked lower or may be missed.

The misreading "Oki" (b) is broken down into "O" (c) and "Ki" (d), which are then compared with the similar character storage unit 115 to identify similar characters in order of frequency of occurrence (e, f). Next, a family name formed by combining characters with high frequency of occurrence is compared with the family name storage unit 113 to select and sort candidate characters (g). In this way, even if there is a misreading, by performing a comparison with the similar character storage unit 115, it is possible to avoid the subordination of the correct answer "Kawamoto" (a) or the omission of candidates, and it is expected that the character recognition rate will be improved.

The above is an example of a handwriting recognition system according to an embodiment. According to the example of the handwriting recognition system according to the embodiment, similar character string candidates can be more appropriately suggested for an image in which handwriting has been read, so that a user can smoothly perform the task of reading handwriting by simply selecting from the presented candidates.

In addition, in the above embodiment, a name is used as an example of a handwritten character, but this is not limited to this, and by replacing the real character storage unit 112 to be compared, it can be applied to any other character data, such as an address.

In addition, the technical elements of the above-mentioned embodiments may be applied independently, or may be divided into multiple parts, such as program parts and hardware parts, and then applied.

The present invention has been described above, focusing on the embodiments.

1: Handwritten character identification system, 50: Network, 100: Character candidate suggestion device, 110: Memory unit, 111: Received data memory unit, 112: Existing character memory unit, 113: Family name memory unit, 114: First name memory unit, 115: Similar character memory unit, 116: Candidate character memory unit, 120: Processing unit, 121: Commissioned processing unit, 122: Family name discrimination processing unit, 123: Matching processing unit, 124: Similar character registration processing unit, 125: Candidate character transmission processing unit, 130: Communication unit, 200: Business terminal, 210: Memory unit, 211: Character memory unit, 212: Candidate character memory unit, 213: Image memory unit, 220: Processing unit, 221: Character recognition unit, 222: Commissioned processing unit, 223: Visual confirmation processing unit, 224: Similar character report processing unit, 230: Communication unit, 300: Reading device.

Claims (9)

a similar character storage unit that stores similarity relationships between character appearances;
an existing character storage unit that stores character strings of existing person names ;
a processing unit for receiving one or more characters and generating character candidates for a person's name having a similar appearance to the received characters;
a matching processor that identifies character candidates having a similar appearance to the received character by matching the candidates with the similar character storage unit, and matches a combination of the candidates with a similar appearance with the existing character string stored in the existing character storage unit;
a candidate character transmission processing unit that outputs the combination of candidates of characters having similar appearances that are found as a result of the matching by the matching processing unit as candidate characters for a person's name ;
Equipped with
the candidate character transmission processing unit arranges and outputs candidate characters for the name, which are combinations of the candidates for the characters having similar appearances, in order of frequency of occurrence;
A character candidate suggestion device comprising: The character candidate suggestion device according to claim 1 ,
a similar character registration processing unit that, when receiving a character similar to any of the characters received by the processing unit after outputting the candidate characters, stores the character in the similar character storage unit if the similarity relationship is not stored in the similar character storage unit;
A character candidate suggestion device comprising: The character candidate suggestion device according to claim 1 ,
The character string of the name of an existing person is divided into a surname and a given name,
a name discrimination processing unit that divides the received characters into a family name and a given name and causes the matching processing unit to match the characters;
A character candidate suggestion device comprising: The character candidate suggestion device according to claim 3 ,
The name discrimination processing unit specifies a combination of the received characters divided into a surname and a given name by changing a position separating the surname and the given name, and when a hit is obtained as a result of the matching by the matching processing unit for both the surname and the given name of each combination, the combination is set as the candidate character.
The character candidate suggestion device according to the present invention is characterized in that The character candidate suggestion device according to claim 4 ,
the candidate character transmission processing unit outputs the candidate characters in order according to their appearance frequency.
The character candidate suggestion device according to the present invention is characterized in that The character candidate suggestion device according to claim 1 ,
The characters received by the processing unit are characters obtained by reading and recognizing handwritten characters.
The character candidate suggestion device according to the present invention is characterized in that A handwritten character identification system including: a character candidate suggestion device; a business terminal communicatively connected to the character candidate suggestion device via a network; and a reading device that reads handwritten characters in response to an instruction from the business terminal,
The business terminal has a character recognition unit that acquires the handwritten characters from the reading device and performs character recognition;
a commissioning processing unit that transmits the characters recognized by the character recognition unit to the character candidate suggestion device to commission the suggestion of character candidates;
a visual confirmation processing unit that displays the character candidates output from the character candidate suggestion device,
The character candidate suggestion device,
a similar character storage unit that stores similarity relationships between character appearances;
an existing character storage unit that stores character strings of existing person names ;
a processing unit for receiving one or more characters and generating character candidates for a person's name having a similar appearance to the received characters;
a matching processor that identifies character candidates having a similar appearance to the received character by matching the candidates with the similar character storage unit, and matches a combination of the candidates with a similar appearance with the existing character string stored in the existing character storage unit;
a candidate character transmission processing unit that outputs the combination of candidates of characters having similar appearances that are found as a result of the matching by the matching processing unit as candidate characters for a person's name ;
Equipped with
the candidate character transmission processing unit arranges and outputs candidate characters for the name, which are combinations of the candidates for the characters having similar appearances, in order of frequency of occurrence;
A handwriting recognition system comprising: A handwriting recognition method using a handwriting recognition system including a character candidate suggestion device, a business terminal communicatively connected to the character candidate suggestion device via a network, and a reading device that reads handwriting in response to an instruction from the business terminal,
a character recognition step in which the business terminal acquires the handwritten characters from the reading device and performs character recognition;
a commissioning process step of transmitting the characters recognized in the character recognition step to the character candidate suggestion device to commission the suggestion of character candidates;
A visual confirmation processing step of displaying the character candidates outputted from the character candidate suggestion device is executed;
The character candidate suggestion device,
a similar character storage unit that stores similarity relationships between character appearances;
an existing character storage unit that stores character strings of existing person names ;
a processing step of receiving one or more characters and generating character candidates for a person's name having a similar appearance to the received characters;
a matching process step of identifying character candidates having a similar appearance to the received character by matching them with the similar character storage unit, and matching a combination of the character candidates having a similar appearance with the existing character string stored in the existing character storage unit;
a candidate character transmission processing step of outputting the candidate combination of characters having similar appearances, which is found as a result of the matching in the matching processing step, as candidate characters for a person's name ;
Run
In the candidate character transmission processing step, candidate characters for a name, which are combinations of the candidates for characters having similar appearances, are arranged in order of frequency of occurrence and output.
A handwriting recognition method comprising: A program for causing a computer to perform a process for assisting in the identification of handwritten characters, comprising:
The computer includes:
a similar character storage unit that stores similarity relationships between character appearances;
The apparatus includes an existing character storage unit that stores character strings of existing personal names , and a processing unit,
The processing unit includes:
a processing step of receiving one or more characters and generating character candidates for a person's name having a similar appearance to the received characters;
a matching process step of identifying character candidates having a similar appearance to the received character by matching them with the similar character storage unit, and matching a combination of the character candidates having a similar appearance with the existing character string stored in the existing character storage unit;
a candidate character transmission processing step of outputting the candidate combination of characters having similar appearances, which is found as a result of the matching in the matching processing step, as candidate characters for a person's name ;
Implement the following :
In the candidate character transmission processing step, candidate characters for a name, which are combinations of the candidates for the characters having similar appearances, are arranged in order of frequency of occurrence and output.
A program characterized by:

Images