JP2001195636A - Image data recognizing device of automatic transaction device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the precision of character recognition by making a preferential comparison with past transaction information on a customer for the character recognition. SOLUTION: This device is equipped with an image input part 10 which reads an image entered into respective fields of a document, a transaction destination information storage part 20 which stores information on transaction destinations of past transactions, and a recognition part 19 which analyzes the read image to rank candidates for transaction destination information and corrects and presents the ranking according to information stored in the transaction destination information storage part.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an automatic transaction apparatus installed in a financial institution or the like and performing a transfer process or the like by a customer operation, and more particularly to an automatic transaction apparatus that performs a data recognition process by inputting an image.

[0002]

2. Description of the Related Art Japanese Unexamined Patent Publication No. 6-139265 discloses a system in which a handwriting input device is provided, image data input through the handwriting input device is read, and transaction processing is performed with a host device based on the read content. An automatic transaction device is shown.

In recent years, a transaction apparatus has been developed which recognizes the type of a form, performs character recognition in accordance with the recognized form of the form, and obtains data necessary for processing a transaction. In this apparatus, for example,
As shown in Japanese Patent No. 185, character recognition processing is performed by collating with a word dictionary.

[0004]

The automatic transaction apparatus installed in a financial institution or the like is often used repeatedly by a specific customer. The character recognition device attached to the conventional automatic transaction device has a word dictionary, and performs character recognition with reference to the dictionary.

Therefore, even when the customer performs the same processing many times using the automatic machine, the characters are recognized using the same word dictionary, and the accuracy of character recognition is not improved. Therefore, it is necessary for the customer to correct the character recognition result each time the user uses the character recognition. SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned problems, and improves the accuracy of character recognition by preferentially collating with past transaction information of a customer at the time of character recognition.

[0006]

The present invention employs the following means in order to solve the above-mentioned problems.

[0007] Images written in each field of the form are read, and the read images are analyzed, and the candidates of the supplier information are ranked and presented. At this time, the recognition unit corrects and presents the ranking based on the information stored in the supplier information storage unit, and the user determines the supplier information based on the ranking.

The ranking can be corrected and presented based on the information stored in the supplier information storage unit and the information stored in the dictionary unit. In addition, the ranking can be corrected and presented based on information on the number of transactions with a business partner.

[0009]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to FIGS. FIG. 1 is a diagram showing an automatic transaction apparatus according to an embodiment of the present invention. In the figure, reference numeral 1 denotes an automatic transaction apparatus, which is provided with an image input section for reading a form such as a transfer slip inserted by the customer into the automatic transaction apparatus and a payment slip of a utility bill, as will be described later. Reference numeral 2 denotes a higher-level device, which performs data exchange for performing online processing by the automatic transaction device. 10 is an image input unit,
Reference numeral 11 denotes a display device made up of a CRT or a liquid crystal, which displays a transaction operation guide screen, an image data recognition screen, a transaction result, and the like. An input unit 12 includes a keyboard or a touch panel, and performs selection of a transaction type, correction of a recognition result, and the like. Note that the display unit 11 and the input unit 12 can be operation display boards with a touch panel. 13
A transfer card issuing unit issues a card in which information of a transfer destination is recorded based on an operation performed before. Reference numeral 14 denotes a form handling device, which takes in forms such as a transfer form and a utility bill payment form, and performs processing such as classification, storage, discharge, and separation of a receipt portion on the taken-in form. Reference numeral 15 denotes a card reader / writer which reads and writes a cash card, reads a transfer card, and the like.
Reference numeral 16 denotes a receipt issuing unit for issuing a receipt according to the transaction content. Reference numeral 17 denotes a passbook printer, which prints transaction contents. Reference numeral 18 denotes a cash depositing / dispensing device, which performs processing such as cash depositing / dispensing, automatic counting of cash, etc., in transactions involving depositing / dispensing. Reference numeral 19 denotes a control device that controls operations of the devices, exchanges data between devices, performs recognition processing, storage, and the like, and controls data communication with the host device 2. Further, the control device 19 includes a word dictionary and a bank name dictionary, and performs character recognition processing of the form image data read by the image input unit using the word dictionary, the bank name dictionary, and the business partner information accumulated in the storage unit 20. . Reference numeral 20 denotes a storage unit of the host device 2, which stores the transaction information of the customer who has made a transaction with the automatic transaction device 1 while sequentially updating the transaction information, and transfers the transaction information according to a request from the automatic transaction device 1.

FIG. 2 is a diagram showing supplier information stored in the storage unit 20 of the host device 2. The customer information includes the name of the customer bank that has previously made a transaction for each customer name, the number of transactions for each customer bank, the name of the customer branch, and the number of transactions for each customer branch.

FIG. 3 is a diagram showing an example of a transfer slip. In the figure, an entry frame 1 is an entry frame in which the name of the bank to be transferred is entered by handwriting or the like, and an entry frame 2 is an entry frame in which the name of the branch office of the correspondent bank is entered by handwriting or the like.

FIG. 4 is a diagram showing an example of a transaction screen displayed on the operation display board of the automatic transaction apparatus 1. Field 1
Displays guidance for operation to the customer. Field 2 displays a cut-out image of the entry frame 1 of the transfer slip shown in FIG. Field 3 displays the name of the transfer destination bank, which is the recognition result of the cut-out image. Field 4
Displays another candidate in the second or lower rank of the recognition result. Field 5 is an area for performing key input by a touch panel for correcting a recognition result. When another candidate is selected, the corresponding candidate in the field 4 may be selected using the touch panel.

FIG. 5 shows that the handwritten character "Royal" in field 1 of the transfer slip shown in FIG.
2 shows an example of a result of collation with the supplier information and the bank name dictionary stored in the automatic transaction apparatus 1 shown in FIG. In the figure, the candidate order indicates the order of certainty of the candidate at the time of recognition. Candidate points are points representing the degree of certainty according to the candidate ranking,
The setting is optional. The combination candidates are obtained by combining recognized character candidates. The word points are the sum of the candidate points of each character candidate, and the “royalty” in the example of FIG. 4 is 180 points that is the sum of 100 points of the “king” and 80 points of the “ritsu”. In the actual result addition, when the partner bank name shown in FIG. 2 matches the combination candidate, the number of transactions × arbitrary points is added to the word points. However, in order to prevent the addition point from becoming too large depending on the number of times, an addition point with an arbitrary number as an upper limit is added. In this embodiment, the upper limit is 1
It is set to 0 times, and if it is more times, it is not added. In other words, the customer Ouritsu Hanaco has made 25 transactions with the Royal Bank, but the upper limit is 10 times, and 100 points are added by 10 times 10 points. As a result, the total points of the combination candidate “royalty” are 280 points. Further, the existence confirmation indicates that the combination candidate is an existing bank name, and in the present embodiment, the banks of the moon, royal, meji, and Nishiyama are stored in the bank name dictionary storing the existing bank names. It indicates that.

FIG. 6 is a flowchart showing the operation of the image data recognition device of the automatic transaction device. Hereinafter, based on this flowchart, the details of the operation when processing the transfer slip will be described.

First, in step 100, the customer inserts a cash card or a transfer card into a card reader.
When inserted into the writer 15, the control device 19 specifies the customer "Ouritsu Hanako" from the information in the card. Step 1
In 02, data of a bank name, a branch name, and the number of transactions is acquired from past supplier information stored in the storage unit 20 of the higher-level device 2. In step 104, a form image of the form inserted by the customer is obtained. In the case of transfer by cash payment, it is not necessary to insert a cash card, but if there is a cash card or a transfer card to identify the customer, a message prompting the customer to insert it is displayed on the screen. After the card is inserted, a message prompting the user to set a form such as a transfer slip in the image input unit 10 is displayed. When the transfer slip is set in the image input unit 10 and an image capture instruction is input from the input unit 12, the image input unit 10 acquires an image of the transfer slip. The image acquisition may be started when a sensor is detected to confirm that the slip has been set. In step 106, the control device 19 performs a slip recognition process from the obtained slip image, and determines that the read slip is, for example, a transfer slip from the slip ID or ruled line pattern or layout pattern on the slip registered in advance. Is specified, and the transfer business process is started. The customer can also input the transfer slip from the input unit 12 to specify the transfer slip. When the slip is specified in step 108, character recognition of the input items is performed according to the form format of the slip. When character recognition is performed for the bank name “Royal” in the entry frame 1 of the transfer slip shown in FIG. 3, candidates for the king, such as king, moon, eyes, mouth, field, and west, as shown in FIG.
As for “Tate”, candidates such as “Sat”, “Toshi”, “Tate”, “Gen”, “Mountain”, “Three” are extracted from the character recognition dictionary. In step 110, combination candidates are created from the two characters extracted as candidates, and word points, which are the sum of the candidate points, are calculated. In step 112, it is determined whether or not the combination candidate matches the past business partner bank name acquired in step 102, and if so, the process proceeds to step 114;
Proceed to. In step 114, the number of points according to the number of transactions is added.

In step 116, the combination candidates are compared with a bank name dictionary storing actual bank names, and only candidates that match bank names defined in the bank name dictionary are selected. In step 118, among the candidates, the top seven candidates are selected in the order of the total word points obtained by adding the results, and the top candidate "royalty" is recognized as the recognition result in the guidance in the field 1 in FIG.
To be displayed. The remaining six candidates are displayed in another candidate list in field 4.

Here, when it takes a long time to collate with the word dictionary and the processing wait becomes long, when the past business partner information and the combination candidate match, the word point of the matching combination candidate becomes the highest point. May be displayed first as a recognition result, in a state of waiting for confirmation input, then collated with a word dictionary, and those having higher word points may be additionally displayed in another candidate list. Further, when the collation with the word dictionary becomes longer and the influence cannot be ignored, the collation itself may be cut off within a certain period of time, and the combination candidate matched at that time may be displayed as another candidate.

In step 120, if the recognition result is incorrect, the customer selects the correct bank name from the alternative candidates displayed in field 4 of FIG. Enter the correct bank name. The entered bank name is displayed in field 3. In step 122, after confirming that the bank name displayed in the field 3 is correct, the customer inputs the confirmation key in the field 5 to determine the bank name. The above-described confirmation processing is similarly performed for the branch name in the field 2 shown in FIG. Further, for the other input items, the confirmation key is input to execute the confirmation processing, and the process proceeds to step 124. If further correction is necessary, the process returns to step 120 without operating the cancel key to perform the fixing process, and the correction input is performed in step 120.

At step 124, a payment selection screen is displayed on the display unit 11. When paying the transfer in cash, the cash depositing / dispensing device 18 performs a depositing process and a change dispensing process. In the case of withdrawing from the own account for payment, the automatic transaction device 1 is operated with a cash card, a passbook, and a password in accordance with the guidance on the display unit 11 to perform a payment process. When the payment processing is completed, step 126
Proceed to. In step 126, the transaction information is transmitted to the host device 2. In step 128, the host device 2
Updates the supplier information stored in the storage unit 20 shown in FIG. 3 and transmits the transaction completion information to the automatic transaction apparatus 1 after performing processing necessary for the transaction. In step 130, if there is a passbook based on the received transaction completion information, the automatic transaction apparatus 1 records the passbook with the passbook printer 17, and
Issue a receipt at. When the customer inputs a transfer card creation request for this business partner by selecting a menu, a transfer card issuance unit 13 issues a transfer card. Finally, the cash card or the transfer card is returned, and the process ends.

In the present embodiment, the case where the present invention is applied to a customer-operated type automatic transaction apparatus installed in a financial institution is described. In response to the input for specifying the transaction information, the past transaction information and the recognition result can be collated.

Further, in this embodiment, an example in which the past supplier information is stored in the host device 2 has been described. However, the past supplier information is recorded on a medium such as an IC card.
It can also be read from the card reader / writer 15 and referenced as collation data.

Further, in this embodiment, an example is described in which the character recognition of the transferee information is preferentially collated with the past supplier information. However, items that increase the number of candidates when performing the recognition processing, for example, A similar effect can be expected by recognizing the form format itself by preferentially collating the format information of the voucher handled in the past transaction as in the case of the past supplier information. In particular, collation of the form format takes longer than word collation, but at this time, collation with the form format dictionary is terminated for a limited time. Can be avoided.

As described above, according to the present embodiment, in the process of recognizing the contents of a voucher, the corresponding item is preferentially collated with the customer information of the past customer. In particular, the effect is remarkable in the case where the same processing is performed, for example, the same bank and the same branch are repeatedly transferred. That is, even if the customer writes in handwriting characters that are difficult to recognize, the recognition rate can be improved, and the trouble and time required for correction can be significantly reduced.

[0024]

As described above, according to the present invention, the character recognition is preferentially collated with the past transaction information of the customer, so that the accuracy of character recognition is improved.

[Brief description of the drawings]

FIG. 1 is a diagram showing an automatic transaction device according to an embodiment of the present invention.

FIG. 2 is a diagram showing supplier information.

FIG. 3 is a diagram showing an example of a transfer slip.

FIG. 4 is a diagram showing an example of a transaction screen displayed on an operation display board of the automatic transaction device.

FIG. 5 is a diagram showing a result of character recognition.

FIG. 6 is a flowchart showing the operation of the image data recognition device of the automatic transaction device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Automatic transaction apparatus 2 High-order apparatus 10 Image input part 11 Display part 12 Input part 13 Transfer card issue part 14 Report handling device 15 Card reader / writer 16 Receipt issue part 17 Passbook printer 18 Cash deposit and withdrawal device 19 Control device 20 Storage part

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G06K 9/68 G06F 15/30 A 9A001 H 320 360 F term (Reference) 3E040 AA06 BA07 CA11 EA01 FH05 FJ05 5B055 BB03 BB19 CB09 CB10 KK01 LL03 LL07 LL15 MM01 PA21 5B064 AA01 AB03 EA12 EA19 EA21 FA04 5B069 AA07 JA08 5E501 AA09 AC23 BA05 CA03 CB13 FA13 FA22 FA23 FA42 9A001 HH22 HH23 JJ64 KK58

Claims (4)

[Claims] 1. An image input unit for reading an image written in each field of a form, a supplier information storage unit for storing information of business partners who have traded in the past, and An image data recognition apparatus for an automatic transaction apparatus, comprising: a recognition unit that ranks candidates and corrects and presents the ranking based on information stored in the supplier information storage unit. 2. An image input unit for reading an image written in each field of a form, a supplier information storage unit for storing information on business partners previously traded, a dictionary unit for storing business name information, Analyzing the read image to rank the candidates for supplier information, and recognizing and presenting the ranking based on the information stored in the supplier information storage unit and the information stored in the dictionary unit. An image data recognizing device for an automatic transaction device, comprising: 3. An image input section for reading an image written in each field of a form, a supplier information storage section for storing information on previously traded suppliers and the number of transactions, and a dictionary for storing name information of suppliers. Part, analyzing the read image and ranking and presenting the supplier information candidates, and presenting the ranking information on the supplier information and the transaction count stored in the supplier information storage unit, and An image data recognizing device for an automatic transaction device, comprising: a recognizing unit that corrects and presents based on information stored in a dictionary unit. 4. An image data recognition apparatus for an automatic transaction apparatus according to claim 3, wherein an upper limit is set for information on the number of transactions related to said ranking.