JPH06208650A - Method and device for recognizing information - Google Patents

Abstract

PURPOSE:To enable optimum layout for each recognizing object and to extremely and easily improve versatility without being limited by the display position and display range of the recognizing object in the case of recognizing in-frame information surrounded by a frame inside the recognizing object. CONSTITUTION:Under the control of a CPU 1, a frame recognition part 9 recognizes the attributes of the in-frame information inside the recognizing object in the order of frame reading at a reader 4. Then, the attributes are registered on a frame attribute registration part 10 in the order of recognition. When recognized information is read from the recognizing object by the reader 4 after the registration onto the frame attribute registration part 10, the frame recognizing part 9 successively recognizes the frames in the read information in the order of reading. The information in the frame read by the reader 4 is recognized according to the attributes of the in-frame information in the order of recognition registered on the frame attribute registration part 10 in the order of recognized frames at this frame recognizing reader part 9. For example, the frame recognizing part 9 is provided with a dictionary for each information attribute. Otherwise, the information attributes contain the respective attributes that the in-frame information is numerals, English sentences or Japanese sentences.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information recognition method and an information recognition device for recognizing information within a frame surrounded by a frame to be recognized.

[0002]

2. Description of the Related Art In a conventional character recognition device, all read originals have the same attribute. For example, if the manuscript was an English manuscript, it was all English manuscripts.
Further, the position on the document in which the information to be recognized for each document is displayed is uniquely specified in advance, and only the information at the predetermined position is recognized.

Further, there is also a recognition apparatus in which recognition targets having a plurality of attributes can be mixed on one original. But,
Even in this case, the display position for each attribute of the mixed recognition target is predetermined. In particular, when selecting and recognizing only character information on a document in which characters to be recognized and image information other than the recognition target are mixed, the display position of the recognition target on all read documents is determined in advance. I had to keep it. This had a very long recognition time, and the recognition result was also very unreliable.

[0004]

Therefore, when the recognition target and the non-recognition target are mixed, the recognition position is determined, and a constant area is required even when the recognition target is small. However, even if there was a large amount, only a predetermined amount could be represented. Furthermore, this position was also strictly defined, and even if the position changed a little, correct recognition could not be achieved.

For this reason, when a desired recognition target is transmitted by using the facsimile device and the receiver side wants to recognize it, only information according to the standardized format can be transmitted, and the information can be transmitted. It was not possible to make the optimum processing format for each occurrence point. For this reason, for example, even when it is desired to place an order for a product to a wholesaler by using the facsimile communication, the format of the order slip must be a format that can be commonly used for all products common to all products. In addition, if there is a display of unnecessary characters or the like on this slip, it may be erroneously recognized, which is very inconvenient.

Further, if the sending position of the slip or the like is slightly shifted, or if the reading position of the document reading device is slightly shifted, erroneous recognition cannot be avoided.

[0007]

The present invention has been made for the purpose of solving the above-mentioned problems, and the present embodiment has the following structure as one means for achieving this object.
That is, the registration means for registering the attributes of the in-frame information in the recognition target in the frame reading and recognition order, the information reading means for reading the recognition information from the recognition target after the registration in the registration means, and the information read by the information reading means Frame recognizing means for sequentially recognizing the frames in the reading order, and information in the frame read by the information reading means according to the attribute of the in-frame information in the recognition order registered in the registration means in the recognition frame order in the frame recognizing means. And an information recognition means for recognizing.

[0008]

In the above structure, the attributes of the in-frame information in the recognition target are registered in the frame reading and recognition order, and the in-frame information is recognized according to the in-frame information attributes registered in the recognition order of the frames in the read document thereafter. Thus, the display position and the display range of the recognition target are not limited at all, and the optimum layout can be set for each recognition target, which is very simple and versatile.

Further, the portion not surrounded by the frame can be removed from the recognition object, and the restriction of the format of the recognition object can be made very small.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment according to the present invention will be described in detail below with reference to the drawings. FIG. 1 is a block diagram of an information recognition apparatus according to an embodiment of the present invention. In the figure, reference numeral 1 indicates the entire apparatus of the present embodiment in accordance with a control procedure stored in a ROM 2, for example, shown in FIGS. A CPU for controlling 2, a ROM for storing the above-mentioned programs and the like, a RAM 3 for storing the progress and processing results of the CPU 1, and a reference numeral 4 for each line of the read document to be recognized in the main scanning direction. A reader 5 for reading is a printer for printing out a recognition result, received image information and the like.

Reference numeral 6 is a keyboard for inputting operation instructions to the apparatus of this embodiment, 7 is a display unit for displaying and outputting the recognition result and processing progress, and 8 is data read by the reader 4 and received by the communication control unit 14. An image memory for temporarily storing data received from the other device 30, and a frame recognition unit 9 for recognizing and detecting the data to be received and recognized and the frame in the document read by the reader 4. Reference numeral 10 denotes a frame attribute registration unit for registering frame attributes in the order in which each frame is generated. Each of designated data in the frame (control or format designation command display frame) first recognized by the frame recognition unit 9 corresponds to the following. The frame attributes in the frame generation order in the format are registered.

Reference numeral 11 is an information recognition unit for recognizing the data in the frame recognized by the frame recognition unit 9 according to the frame attribute registered in the frame attribute registration unit 10, and 12 is the reception recognized by the information recognition unit 11. A file for filtering data, received data, or the like, and 13 is a dictionary memory for each attribute to be recognized. For example, in the present embodiment, a dictionary memory for "alphanumeric characters", "for katakana, hiragana", and "for kanji". have. However, any dictionary memory can be installed as needed.

Reference numeral 14 denotes a communication control unit for receiving communication data sent from another device 30 via the network 20, and transmitting transmission data from the device of this embodiment to the network 20. Manages the interface with the network 20. In addition,
The communication control unit 14 performs compression processing such as MR when transmitting transmission image data or the like read from the reader 4 or read from the file 12 to another device by compression, and CCITT, which is the international standard for transmission, CCITT. The transmission information is output to the line network 20 by the control procedure according to T.30.

When the communication data from the line network 20 is received, a predetermined expansion process is performed and the process is passed to the CPU 1. The CPU 1 prints out the received data from the printer 5. The apparatus of this embodiment is a G-standard machine with a transmission rate automatic back-back function.

The information recognition processing of the apparatus of this embodiment having the above configuration will be described below. In this embodiment, only the portion surrounded by the frame in the received document or the read document is recognized, and the handling of the document is determined according to the recognition result, and the printout process or the document type is determined as necessary. Then, the file 12 is filed, the predetermined processing is executed in accordance with the instruction of the received document, and the like.

The selection of these execution processes is performed according to the control command in a predetermined frame, for example, according to the designation of the first recognized in-frame information. First, the frame attribute registration processing of the apparatus of this embodiment will be described with reference to FIG. In this embodiment, the attribute of the frame to be recognized is not assigned based on the position on the document, but according to the frame order recognized (scanning order is recognized) when the recognized image data on the document is sequentially scanned. Allotted. Therefore, first in step S1, the document type for which the frame attribute is to be registered is set. This is the document type specified by the data in the frame specified as the control frame in the recognition process, and the number of frames to be recognized thereafter and the attribute type for each frame are set to the optimum format for each process. This is because.

Then, in step S2, attributes are specified in the recognition order of each frame. In this attribute specification, the data in each frame is
The designation is "Japanese with mixed kanji", "Hiragana or Katakana sentence", or "Alphanumeric sentence". Then, in step S3, the attributes of each frame are set in the frame attribute registration unit 1 in the order of frame recognition.
Register to 0. Then, in a succeeding step S4, it is checked whether or not the frame attribute registration processing is performed for another document type. If the next frame attribute is to be registered, the process returns to step S1 and the recognition order of each frame in the next document type. The frame attribute in 1 is registered in the frame attribute registration unit 10.

When the frame attributes for the desired document type have been designated, the registration process is terminated and the process proceeds to the recognition process described later. The keyboard 6 is used to input the document type and the frame attribute.
Alternatively, the document for frame attribute registration may be read by the reader 4 and used. When using a manuscript for registration, the designation is made with a character type such as alphanumeric characters that is easy to recognize, and each frame required for the reader 4 is displayed.
Set by reading the original in which the code indicating the attribute of the frame is displayed in the frame, read each frame by the same process as the recognition process described later, and recognize by the specified code in the frame To do. Then, the recognition result may be displayed on the display unit 7 to confirm the recognition result.

An example of the frame attribute registered as described above is shown in FIG.
Shown in. FIG. 3 is an example in which the present embodiment is applied to an ordering slip in facsimile communication. In FIG. 3, 50 is a document, 51 is a document type designation frame for designating a document type, and 52 to 60 are respective recognition frames in the document type specified by the document type designation frame 51. In the present embodiment, in the frame recognition by the frame recognition unit 9, scanning is performed line by line from the upper left corner of the document of FIG. 3 first in the main scanning direction (right direction), and then in the sub scanning direction (downward). Scan the line. Therefore, the upper left frame of the document 50 is sequentially recognized in the lower right frame. Accordingly, the document type designation frame 51 is first recognized, and thereafter, the frames 52 and 53 are recognized up to the frame 60 in the code order.

In each frame, 52 is a frame in which the name of the orderer is entered (Japanese characters mixed with Chinese characters), and 53 and 56 are frames in which the ordered product name or the ordered product code is entered (Japanese characters mixed with Kanji characters or alphanumeric character frames). ), 54 and 57 are boxes for entering the number of orders (alphanumeric text box), 55 and 58 are subtotal entry boxes (alphanumeric text box), 59 is tax entry box (alphanumeric text box), and 60 is total entry. It is a frame (alphanumeric sentence frame).

Next, when the frame attribute setting registration for each document type is completed as described above, the registered data held in the RAM 3 is read out, and the frame attribute registration section 10 sets the position corresponding to the designated document type. register. Next, the recognition processing of the recognition target for which the registration processing has been performed in this way (the recognition processing of the data within the frame with respect to the document which is the recognition target transmitted by facsimile communication or read by the reader 4) is performed. Will be described below.

When the frame attribute registration processing shown in FIG. 2 is performed and the frame attribute specified by the document type is registered in the frame attribute registration unit 10, the recognition target is sent thereafter or subsequently. The process moves to the document data recognition process. First, in step S10, the recognition target is read and sequentially stored in the image memory 8. The recognition target is the read data from the reader 4 or the image data received from the other device 30 via the line network 20 and the communication control unit 14 by facsimile.

In the image memory 8, the original shown in FIG. 3 is read sequentially from the upper left to the lower right. When at least a predetermined amount of recognition target is stored in the image memory 8, the CPU 1 activates the frame recognizing unit 9 to detect the frame in the image memory 8 read in steps S10 and S11. When the frame is detected, the process proceeds from step S11 to step S12, and the detected position and size of the frame are held in the RAM 3. Then, in step S13, it is checked whether or not the frame recognition of the recognition target is completed,
If all frame recognition is not completed, the process returns to step S10 to continue the next frame recognition processing.

When the recognition target frame detection processing is completed, the process proceeds from step S13 to step S14. In step S14, the document type setting frame detected first (frame 5 in the example of FIG. 3).
The type code of 1) is recognized, and the frame attribute for each frame detection order in the relevant document type is read out from the document type registered in the frame attribute registration unit 10 in step S16.

When the corresponding frame attribute is not stored, the document type and frame attribute registered immediately before are selected. And
In step S17, it is checked whether or not the document requires the subsequent recognition process of the in-frame data. For example, in the case of facsimile communication data, if the data is simply stored at the position of the file 12 designated by the document type or temporarily stored and transferred to another device, the process proceeds from step S17 to step S20.
The communication data is stored in the corresponding area of the file 12, and the recognition process of the document is completed. In this case, if a plurality of originals are subsequently sent, all original data may be associated and stored in the file 12.

On the other hand, when the recognition processing is required in step S17, for example, when the ordering slip shown in FIG. 3 is received, the process proceeds from step S17 to step S25, and the frame attributes in the order of detection are next. And waits for the frame recognition unit 9 to detect a frame in step S26. Then, when a frame is detected, the process proceeds to step S27, the dictionary memory 13 corresponding to the attribute of the frame registered in advance in the recognized order is selected, and the selected dictionary memory 13 is referred to in the same manner as described above. The recognition processing by the information recognition unit 11 is executed. Then, the recognition result is compared with corresponding code information that can be directly processed by the CPU 1 and stored in the RAM 3.
The recognition process is continued until the recognition process of all the in-frame information is completed in step S28, and when all the recognition processes are completed, the process proceeds to step S29.

This recognition processing is performed by the information recognition unit 11. First, the character strokes and character feature points in the frame are detected, and the detected feature points and strokes are compared with the feature points and strokes in the dictionary memory 13 selected according to the attributes to identify the most promising candidates for recognition. To be recognized. For these character recognition circuits and recognition methods, conventional general circuits and methods can be used.

In step S29, the CPU 1 reads the recognition result in the RAM 3 and performs the corresponding arithmetic processing. If no frame can be recognized in the manuscript, it may be recognized as reception data by normal facsimile communication, reading facsimile transmission data, reading processing of copy data, or the like. The frame attributes are specified simply in the frame recognition order without referring to the frame position on the document. Therefore, there is no restriction on the position of each frame at the time of registering the frame attributes. For example, when the attributes of each frame are registered using the format of FIG. 3, or when the first recognized document is the document shown in FIG. Even in such a case, the document format to be recognized is not limited to this. For example, in FIG.
Subsequent to the original document, it is possible to recognize the in-frame information of the original document as shown in FIG. 5 in which the frame recognition order is the same but the frame positions are completely different. In FIG. 5, the frames corresponding to the respective frames in FIG. 3 are given the same numbers, and “a” is added to the end.

As described above, according to the present embodiment, since the attributes and contents are specified only by the recognition order of the frame, there is no influence on the graphic information outside the frame, and any explanatory text or advertisement graphic is displayed. It can be printed. In addition, there are very few restrictions on the position of the frame on the document, and the layout becomes very easy.

Furthermore, there is no restriction on the size of the frame,
It can be arbitrarily determined based on the amount of information required in each frame. For this reason, when there is little information in the frame, a small frame
When there is a lot of information in the frame, it is sufficient to draw a large frame, which makes it very efficient and good in appearance.
In the above description, there is one type of frame, and the case of drawing with a ruled line was described, but the present invention is not limited to the above example, and when the frame is drawn with a double ruled line, etc. As described above, a plurality of types of different modes may be used, and the frame attribute may be specified by the frame mode. By controlling in this way, even if the number of frames does not match the number at the time of registration, it is possible to accurately recognize it, and it is possible to increase or decrease a part of the number of frames as necessary.

In the above description, an example of recognizing all the frame positions in advance and then recognizing the in-frame information has been described. However, the present invention is not limited to the above example, and the frame detecting and recognizing process is performed. It may be controlled so that the in-frame information recognition processing is performed in real time. In this case, the first recognition process of the in-frame information is performed instead of steps S12 to S15 of FIG. 4, and the subsequent recognition process recognizes the frame and executes the recognition process from the time when the in-frame information can be specified. do it. As a result, when the recognition process is not required other than the document type recognition, the printout, the filtering, etc. can be performed immediately, and the unnecessary process is not performed.

The recognition of the specific information in the read original document or the received original document has a very high recognition probability, and is not limited to the simple information recognition or the example used only for facsimile communication. . For example, according to the designation information in the document type specifying frame 51 of the document sent by facsimile communication, the sent document is data that is simply printed out, or is facsimile data communication data. It is possible to specify whether the mail data is for the other party specified by, etc., or the registration data for the specific file.

Therefore, the present invention can be widely applied to a remote order entry system, a remote fairing system, a facsimile mail system, etc. which utilize data communication such as facsimile communication.

[0034]

As described above, according to the present invention, the attributes of the in-frame information in the recognition target are registered in the frame reading and recognition order,
By recognizing the in-frame information according to the in-frame information attributes registered in the frame recognition order in the subsequent read document, the display position and display range of the recognition target are not limited at all, and the optimum layout is set for each recognition target. It can be made very simple and versatile.

Further, the portion not surrounded by the frame can be removed from the recognition object, and the restriction of the format of the recognition object can be made very small.

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment according to the present invention.

FIG. 2 is a flow chart showing a registration process of this embodiment.

FIG. 3 is a diagram showing an example of a frame attribute registration format according to this embodiment.

FIG. 4 is a flow chart showing an in-frame information recognition process of this embodiment.

FIG. 5 is a diagram showing an example of a document to be recognized in the present embodiment.

[Explanation of symbols]

 1 CPU 2 ROM 3 RAM 4 Reader 5 Printer 6 Keyboard 7 Display Unit 8 Image Memory 9 Frame Recognition Unit 10 Frame Attribute Registration Unit 11 Information Recognition Unit 12 File 13 Dictionary Memory 14 Communication Control Unit 20 Network Network 30 Other Device

Claims (4)

[Claims] 1. An information recognition method for recognizing information within a frame surrounded by a frame to be recognized, wherein the attributes of the information within the frame to be recognized are registered in the frame reading recognition order, and the subsequent information is read from a document to be read. An information recognition method characterized by recognizing in-frame information according to in-frame information attributes registered in the frame recognition order. 2. An information recognition device for recognizing in-frame information surrounded by a frame in a recognition target, comprising: registration means for registering attributes of the in-frame information in the recognition target in frame reading recognition order; and the registration means. Information reading means for reading the recognition information from the recognition target after registration in the frame, frame recognition means for sequentially recognizing the frames in the read information by the information reading means in the reading order, and the registration means for recognition frame order by the frame recognition means An information recognition device, comprising: information recognition means for recognizing the information in the frame read by the information reading means according to the attribute of the information in the frame in the recognition order registered in. 3. The information recognition apparatus according to claim 2, wherein a dictionary for each information attribute is provided, and the information recognition means performs information recognition by referring to the dictionary according to the attribute of the in-frame information. Information recognition device. 4. The information recognition device according to claim 2, wherein the information attribute includes each attribute in which the information in the frame is a numeral, an English sentence, and a Japanese sentence.