JPH11238117A - Method and device for image processing and storage medium thereof - Google Patents

Abstract

PROBLEM TO BE SOLVED: To increase the capacity efficiency when image data and a recognition result are held by correcting an image file according to the result of character recognition while making use of image data included in the image file. SOLUTION: An image file generating means 101 encodes one piece of input image data, obtained by electronizing a paper document 104 by an image input means 105 such as a scanner and a camera 105, with plural resolutions and also generate image attribute information regarding the image data and data on a property set, etc., to generate a document image file A106 in such a format that image data are held hierarchically. A character recognizing means 102 performs conversion from image data of a high-resolution image layer included in the document image file A106 to a character code string. An image file correcting means 103 corrects the image file according to the data outputted by the character recognizing means 102. The result obtained by correcting the image file by the image file correcting means 103 by the image file correcting means 103 is an image file A'107.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing method and apparatus related to a filing method when a document including a character image is digitized and input by a scanner or a camera and stored and transferred as an image file. It is a storage medium.

[0002] The present invention relates to an image processing method and apparatus for recognizing character information from an input image and a storage medium therefor.

[0003]

2. Description of the Related Art A document including character information is digitized by a scanner, a camera, or the like, input, and stored as an image file. There was a technology for creating image data and saving it as one file.

[0004] It is necessary to accurately store or output fine images such as small characters with high-resolution image data.
There is no technique for performing various processes suitable for each resolution by using a plurality of image data of different resolutions, such as enabling a reduced display showing an overview of the entire document with low-resolution image data. Was.

There has been a technique for performing character recognition processing on an image input as a character recognition target and converting the image into text data.

[0006]

However, in the above prior art, the recognition result is not reflected in the image file, so that the file capacity efficiency of the image file and the storage data of the recognition result is improved. I never did.

Further, in the above-mentioned prior art, since the recognition result is merely a document file of the recognition result alone, the user himself / herself manages the management data in relation to the recognition result and the image file containing the image data obtained from the recognition result. And had to search for the association while visually checking the created management data.

[0008]

In order to solve the above-mentioned problems in the prior art, the present invention holds an image file containing a plurality of image data of different resolutions, and stores the image file in the held image file. The present invention provides an image processing method and apparatus for recognizing characters, converting the image data into character code strings, correcting the image file according to the character code strings, and a storage medium therefor.

In order to solve the above-mentioned problems of the prior art, the present invention preferably modifies the image file by a process of discarding the image data converted into the character code string.

[0010] In order to solve the above-mentioned problems of the prior art, according to the present invention, preferably, the correction of the image file is a process of storing the character code string in the image file.

[0011] In order to solve the above-mentioned problems of the prior art, the present invention preferably comprises the steps of:
Stores the layout information extracted from the image file.

In order to solve the above-mentioned problems of the prior art, according to the present invention, preferably, the layout information is position information of a text area.

In order to solve the above-mentioned problems of the prior art, the present invention preferably provides that the layout information is row position information.

In order to solve the above-mentioned problems of the prior art, according to the present invention, preferably, the layout information is position information of a character image.

In order to solve the above-mentioned problems of the prior art, according to the present invention, preferably, the layout information is position information of a text area.

In order to solve the above-mentioned problems of the prior art, according to the present invention, preferably, the layout information is character size information.

In order to solve the above-mentioned problems of the prior art, according to the present invention, preferably, the layout information is character color information.

In order to solve the above-mentioned problems of the prior art, according to the present invention, preferably, the layout information is position information of a graphic area.

In order to solve the above-mentioned problems of the prior art, according to the present invention, preferably, the layout information is position information of an image area.

In order to solve the above-mentioned problem of the prior art, the present invention preferably compares the certainty factor in the character recognition processing with a predetermined reference, and performs the character recognition in accordance with the result of the comparison. Switches whether to discard image data.

In order to solve the above-mentioned problems of the prior art, the present invention preferably compares the size of image data with the size of a character code string, and decides whether to discard the image data according to the result of the comparison. Switch.

In order to solve the above-mentioned problem of the prior art, the present invention preferably comprises the steps of: inputting image data; encoding the input image data hierarchically at different resolutions; Then, the image file to be held is generated from.

In order to solve the above-mentioned problems of the prior art, according to the present invention, preferably, the discarding of the image data is performed as a part of the image data on which the character recognition processing has been performed.

In order to solve the above-mentioned problems of the prior art, according to the present invention, preferably, the discarded image data is a text area.

In order to solve the above-mentioned problems of the prior art, according to the present invention, preferably, the discarded image data is a character image area.

In order to solve the above-mentioned problems of the prior art, according to the present invention, preferably, the discarded image data is an area other than an image area.

In order to solve the above-mentioned problems of the prior art, according to the present invention, preferably, the discarded image data is an area other than a graphic area.

In order to solve the above-mentioned problems of the prior art, according to the present invention, preferably, the image data to be discarded is a character image area in which the result of low confidence in the character recognition processing is obtained.

In order to solve the above-mentioned problem of the prior art, the present invention preferably reproduces the image data based on the corrected image file.

In order to solve the above-mentioned problems of the prior art, according to the present invention, preferably, the image data is image data input by a scanner.

In order to solve the above-mentioned problems of the prior art, according to the present invention, preferably, the image data is image data input by a camera.

In order to solve the above-mentioned problem of the prior art, the present invention preferably outputs the character recognition result.

[0033]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 6 is a block diagram of the apparatus according to the present invention.

Reference numeral 61 denotes a CPU (central processing unit).
The control of the process according to the present invention is performed according to the control program stored in the MEM 63. The processing shown in the flowchart described later is also executed under the control of the CPU 61. 63
Is a MEM (composed of RAM and ROM) and CP
A control program for processing executed by U61, various parameters used for the processing, an input image, a dictionary for character recognition, etc.
Various data is stored in the MEM 63. & ＄
A display device such as a CRT or an LCD displays an input image, a text of a processing result, an operation instruction screen, a recognition result read from a file according to the document identification information specified by the input unit 69, and the like. The character recognition result stored in the MEM 63 is displayed on the display 64, and editing such as selecting a correct character from a plurality of candidate characters using the input means 69 is possible. A printer 65 such as an LBP or BJ printer prints images, texts, and the like. Reference numeral 66 denotes a scanner, which optically reads an image of a document and inputs it to the apparatus as an electrical signal. A communication I / F 67 controls data transmission and reception via a public line, a LAN, or the like. The input of an image according to the present invention and the output of the result of processing the image can be transmitted and received to and from another terminal via this communication I / F. Reference numeral 68 denotes a storage medium that is detachable from the apparatus and readable and writable by a computer, such as a CD-ROM, a CD-R, and an FD. The image according to the present invention may be read from the storage medium 68 and the processing result may be written to the storage medium 68. Further, the control program stored in the MEM 63 may be installed from another terminal via the communication I / F, or may be installed from the storage medium 68. Reference numeral 69 denotes input means such as a keyboard and a pointing device, and an instruction from an operator is made through this input means. Reference numeral 60 denotes a bus for transmitting and receiving data between the units.

FIG. 1 is a functional block diagram of the apparatus according to the present invention. The image file generating means 101, the character recognizing means 102, and the image file correcting means 103 are controlled by the CPU 61 based on a control program stored in the MEM 63. Is achieved. Reference numeral 104 denotes a document manuscript, which is input as an electric signal by a scanner or a camera (105) and is subjected to various types of processing. The document image file A 106 is an image file generated by the image file generation unit 101 from image data input from the image input unit 105 such as a scanner or a camera. The document image file A ′ 107 is a document image file generated by the image file correction unit 103. This is an image file obtained by modifying A106.

FIG. 2 shows an example of a hierarchically encoded image file, FlashPixTM (FlashPix is from Eastman, USA).
FIG. 2 is a diagram showing a file format of Kodak (registered trademark). Each property and data in the file can be accessed by storage and stream corresponding to the relationship between directory and file in the hierarchical file system. In FIG. 2, a shaded block (20
1, 208 to 212) are storages, and the blocks without shadows (202 to 207, 213, 214) are streams. Image data including pixel information and image attribute information such as the size, depth, and resolution of the image are stored in the stream portion. The image data is hierarchized at different resolutions, and an image of each resolution is stored as a Subimage (213) for each resolution (Resolution) indicated by Resolution 0, 1,..., N (208 to 212). For each resolution image, image attribute information necessary to read the image is stored as a subimage header (214) for each resolution. The property set (202-205)
Summary Info. Property that defines the attribute information about one object of the file, classifying it according to the purpose and contents of the object, and stores the title and author name of the file.
set202, Image Co storing image data storage method
ntents Property set 203, Image Info. Property set that stores various information that can be used when using images
An extension list property set 205 is an area for adding and storing information not included in the basic specifications.

The format of the image file to which the present invention can be applied is not limited to FlashPix. For example, an image file including a plurality of hierarchically encoded image data, such as CIFF (Canon Image File Format). Needless to say, any data that can store text data such as a character string in addition to image data may be used.

Next, the flow of processing according to the present invention will be described.

The image file generating means 101 encodes one input image data obtained by digitizing a paper document by the image input means 105 such as a scanner or a camera at a plurality of resolutions, and stores image attribute information and properties relating to the image data. A document image file A10 having a format having image data as shown in FIG.
6 is generated. FIG. 3 schematically shows the structure of the document image file A106.

Here, the high-resolution image data 305 and the high-resolution image data 30 belonging to the high-resolution layer 303
Image header 306 storing image attribute information related to image 5
A low-resolution image data 307 belonging to the low-resolution layer 304, an image header 308 storing image attribute information relating to the low-resolution image data 307, and a property in which data described as the property sets 202 to 205 in FIG. A description will be given by taking an image file composed of 302 as an example.

The character recognition means 102 converts the image data 305 of the high-resolution image layer 303 included in the document image file A106 into a character code string by using a known character recognition technique.

The character recognizing means 102 outputs the image data 305
By extracting a character image from each character and comparing each character image with the dictionary data for recognition, a candidate character and a certainty factor of the candidate character (similarity between the input character image and the dictionary data) are obtained.
Output. The text region information, line position information, character image position information, character size information, character image color information, or image regions that have not been subjected to character recognition processing are identified in the process of cutting out character images. The character recognition unit 102 also outputs the position information of the graphic area.

The character recognizing means 102 is provided for the high resolution layer 30.
Although it has been described that the image data 305 of No. 3 is recognized, the image data to be recognized by the character recognizing unit 102 may be any image data suitable for character recognition processing. And image data close to the resolution specified in advance, or the type of input means of the image data, for example, whether it is a scanner or a digital camera, what is their resolution, etc. The image data of the appropriate resolution layer may be selected according to the condition (1). The selection criteria for the image data are stored in the MEM 63 in advance. Or,
At the time of processing, image data of the resolution layer specified by the operator may be selected.

The image file correcting means 103 corrects the image file according to the data output from the character recognizing means 102. The result of correcting the image file of FIG. 3 by the image file correcting means 103 is an image file A ′ shown in FIG.

The image file correcting means 103 discards the image data 305 of the high-resolution layer 303 which has been subjected to the recognition processing, and stores the data relating to the image data 305 stored in the image header 306 and the property 302 in accordance with the discarding processing. Also edit. As a result, the image data storage portion 405 in the high resolution layer 403 becomes empty, and the total size of the document image file decreases. Further, a character code string as a candidate character of the recognition result is stored in the property 30 of the image file.
2 is stored. When a character code string is stored in the property 302, information other than the candidate character string output by the character recognizing unit 102, information of a text area, information of a line position, information of each character image position, character size information, and a character image The color information, the position information of the image area or graphic area that has not been subjected to the character recognition processing, and the identification information of the resolution layer that has been subjected to the character recognition may be stored in the property 302 together.

When the image file correction means 103 discards image data, only the text area subjected to the character recognition processing may be discarded according to the text area information output by the character recognition means 102. Is a character recognition unit 10
In accordance with the position information of the image area or graphic area which has not been subjected to the character recognition processing and output by 2, the area other than the image area or graphic area may be discarded. As described above, when partially discarding, the position information of the discarded area or the position information of the area not discarded may be stored in the property 302.

The image file correcting means 103 does not unconditionally discard the image data 305 of the high-resolution layer 303 in accordance with the extraction of the character code string, but instead executes image data and character code in accordance with the following conditions. Either one of the columns may be selectively left, or both may be left without being discarded.

FIG. 5 is a flowchart showing a process for determining the above conditions performed by the image file correcting means 103.

In step S501, the degree of certainty of the character recognition result output by the character recognition means 102 is evaluated. For example, if the degree of certainty is lower than a predetermined threshold value stored in the MEM 63 and the ambiguity of the recognition result is large. If determined, the high-resolution layer 303 is used for an application such as re-recognition.
The character code string is stored in the property 302 without leaving the image data 305 of FIG.

If it is determined in step S501 that the ambiguity of the recognition result is not large, the process proceeds to step S502.
The size of the image data 305 is compared with the size of the character code string output by the character recognizing means 102, and if the size of the image data is determined to be smaller,
5 without being discarded and the character code string
To be stored.

If it is determined in step S502 that the size of the image data is not smaller than the size of the character code string, the image data is discarded.

The size of the image data determined in step S502 is the size of the image data that can be discarded, that is, the size of the image data of the text area subjected to the character recognition processing by the character recognition unit 102, or And the size of image data other than the image area and the graphic area which have not been subjected to the character recognition processing by the character recognition means 102.

The determination of the ambiguity of the character recognition result in step S501 is performed not for the entire document image but for each sentence region, each line, each character, etc., and the determination whether to discard the image data is partially made. It may be performed for each area.

If it is determined in step S501 that the ambiguity of the recognition result is large, an application for editing the character recognition result is automatically started to display candidate characters and image data on the display screen of the display device 64. After the operator performs an operation of selecting a correct character from the candidate characters and an operation of inputting the correct character, the process may return to S502.

Before discarding the image data, the image file correcting means 103 determines whether a special code such as a discard prohibition code has been added to the image data to be discarded. After confirming that the code or the like is not added, a process of discarding the image data may be performed.

The discard prohibition codes are the image headers 306, 3
08, it is possible to specify whether or not to prohibit destruction for each resolution layer, and if it is stored in the property 302, it is possible to specify whether or not to prohibit destruction for image data of all resolution layers.

Based on the image file as shown in FIG. 4, even if the image data has been discarded by the image file correcting means 103, the input image data can be reproduced. That is, the character code string stored in the property 402, the information of the text area, the information of the line position, the information of each character image position, the character size information, the color information of the character image, or the character recognition processing was not performed. Based on the position information of the image area and the graphic area, and the identification information of the resolution layer in which the character recognition was performed, a character generator generates a character font of the size and color extracted from the image data, and the character in the layout in the input image data is generated. Is arranged, and by further reading and arranging the image data in the image area and the graphic area, an image in which characters, images, and graphics are mixed can be reproduced with the same layout as that at the time of input.

When the resolution layer in which the image data remains is specified in the image reproduction, the image data of the resolution layer may be reproduced.

The character recognition processing by the character recognition means 102 and the image file correction processing by the image file correction means 103 are performed by the input means 105 such as a scanner or a camera.
May be performed when image data is input, or MEM
64, and may be started when a character recognition process or an image file correction process is instructed for the document image file specified by the operator.

[0061]

As described above, according to the present invention,
In an image file including a plurality of image data hierarchically encoded with different resolutions, by correcting the image file according to the result of character recognition using the image data included in the image file, It is possible to increase the capacity efficiency when holding the image data and the recognition result.

In an image file including a plurality of image data hierarchically encoded at different resolutions, after character recognition is performed using image data included in the image file, the recognition result is stored in the image file. By storing and discarding the image data for character reproduction, the capacity efficiency of the image file can be improved in a state where the characters included in the image can be reproduced.

Further, in an image file including a plurality of image data hierarchically encoded at different resolutions, a text obtained as a result of character recognition using image data included in the image file is converted into an image file. , It is possible to use the recognition result in association with a plurality of image data hierarchically encoded with different resolutions, thereby improving operability.

[Brief description of the drawings]

FIG. 1 is a functional block diagram according to the present invention.

FIG. 2 is an exemplary diagram of a format of an image file.

FIG. 3 is a schematic diagram of generated image file data.

FIG. 4 is a schematic diagram of modified image file data.

FIG. 5 is a flowchart of a process of determining whether to discard image data.

FIG. 6 is a configuration diagram of an apparatus according to the present invention.

Claims (36)

[Claims] 1. An image file including a plurality of image data of different resolutions is held, and image data included in the held image file is recognized as a character and converted into a character code string. Modifying the image file according to the following. 2. The image processing method according to claim 1, wherein the correction of the image file is a process of discarding the image data converted into the character code string. 3. The image processing method according to claim 1, wherein the correction of the image file is a process of storing the character code string in the image file. 4. The image processing method according to claim 1, wherein in modifying the image file, layout information extracted from the image file is stored. 5. The image processing method according to claim 4, wherein the layout information is position information of a text area. 6. The image processing method according to claim 4, wherein the layout information is row position information. 7. The image processing method according to claim 4, wherein the layout information is position information of a character image. 8. The image processing method according to claim 4, wherein the layout information is position information of a text area. 9. The image processing method according to claim 4, wherein the layout information is character size information. 10. The image processing method according to claim 4, wherein the layout information is character color information. 11. The image processing method according to claim 4, wherein the layout information is position information of a graphic area. 12. The image processing method according to claim 4, wherein the layout information is position information of an image area. 13. The method according to claim 1, wherein the degree of certainty in the character recognition processing is compared with a predetermined reference, and whether to discard the image data subjected to the character recognition is switched according to a result of the comparison. The image processing method according to claim 1. 14. The image according to claim 1, wherein the size of the image data is compared with the size of the character code string, and whether to discard the image data is switched according to a result of the comparison. Processing method. 15. An image data input device, wherein the input image data is hierarchically encoded with different resolutions, and the held image file is generated from the encoded plurality of image data. Item 2. The image processing method according to Item 1. 16. The image processing method according to claim 2, wherein the discarding of the image data is a part of the image data on which the character recognition processing has been performed. 17. The image processing method according to claim 16, wherein the discarded image data is a text area. 18. The image processing method according to claim 16, wherein the discarded image data is a character image area. 19. The image processing method according to claim 16, wherein the discarded image data is an area other than an image area. 20. The image processing method according to claim 16, wherein the image data to be discarded is an area other than a graphic area. 21. The image processing method according to claim 16, wherein the image data to be discarded is a character image area in which a result having a low degree of certainty in a character recognition process is obtained. 22. The image processing method according to claim 1, wherein the image data is reproduced based on the corrected image file. 23. The image processing method according to claim 1, wherein the image data is image data input by a scanner. 24. The image processing method according to claim 1, wherein the image data is image data input by a camera. 25. The image processing method according to claim 1, wherein the character recognition result is output. 26. Holding means for holding image files including a plurality of image data of different resolutions, and converting the image data contained in the image files held in the holding means into a character code string by character recognition. An image processing apparatus comprising: a character recognizing unit that performs a character recognition process; 27. The image processing apparatus according to claim 26, wherein the image file correction unit discards the image data converted into a character code string by the character recognition unit. 28. The image processing apparatus according to claim 26, wherein said image file correction means stores a character code string obtained by said character recognition means in an image file. 29. A comparing means for comparing the degree of certainty in the character recognition processing by the character recognizing means with a predetermined reference, and the image file correcting means performs the character recognition according to the result of the comparing means. The image processing apparatus according to claim 27, wherein whether to discard the image data is switched. 30. A size comparing means for comparing the size of image data with the size of a character code string, and the image file correcting means determines whether or not to discard the image data according to the result of the size comparing means. The image processing apparatus according to claim 27, wherein the switching is performed. 31. An image processing apparatus comprising: input means for inputting image data; and image file generating means for generating an image file from a plurality of image data obtained by hierarchically encoding the input image data at different resolutions. The image processing apparatus according to claim 26, wherein 32. The image processing apparatus according to claim 26, further comprising image reproducing means for reproducing the image data based on the corrected image file. 33. The image processing apparatus according to claim 31, wherein said input means is a scanner. 34. The image processing apparatus according to claim 31, wherein said input means is a camera. 35. The image processing apparatus according to claim 26, further comprising an output unit that outputs the character recognition result. 36. A control program for storing image files including a plurality of image data having different resolutions, and converting the image data included in the stored image file into a character code string by character recognition. And a computer-readable storage medium storing a control program for correcting the image file according to the character code string.