JP2000137765A - Processor and method for image processing and storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To increase the reproducibility of the whole document image and to obtain an output image having high quality characters by outputting character codes for image parts which are highly possible to be the character document image and outputting images corresponding to the original document image for the other parts. SOLUTION: When a recognition copy mode is selected at an operation part 116, an image read by an image read part 1002 is stored in an image buffer memory 118 and recognized by an image recognition part 4000. Here, a partial image to be recognized is cut by a character segmentation part 108 and a vector calculation part 109 extracts the features quantity. Then it is decided that the partial image can be outputted as characters when the difference between the extracted features quantity and the that of a standard character that a recognition control part 111 has is minimum and less than a threshold, and then its character code are outputted, and an image data conversion part 110 generates a character pattern. When the difference is larger than the threshold, the partial image is outputted in the form of image data as it is.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus, method, and storage medium, and more particularly to an image processing apparatus, method, and storage medium for recognizing a character image in a read image.

[0002]

2. Description of the Related Art With the recent increase in the speed of CPUs and semiconductors, character recognition technology is capable of processing in substantially real time at a copying speed of a low-speed copying machine. or,
In recent years, most of the originals to be copied have been created and printed by a word processor or the like, and thus their recognition accuracy has been improved, and has reached a practical level.

The applicant of the present invention has disclosed, for example, Japanese Patent Application Laid-Open No.
If the technology already proposed in JP 07876 or the like is used, if the character image information of the document is converted into recognized font image information and the image is reproduced, image deterioration due to repeated copying in principle is eliminated.

[0004]

However, in the case where the original is not only characters but also graphics, etc.,
A part of a figure may be converted into a similar character or symbol, resulting in an extremely unnatural reproduced image. Further, when recognizing the character font in order to reproduce the recognized character in the character font of the document, highly accurate character recognition cannot be performed because features such as graphic portions are mixed.

[0005]

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and outputs a character code for an image portion which is highly likely to be a character in a document image. It is an object of the present invention to provide an image processing apparatus and method which can output a corresponding image to enhance the reproducibility of the whole original image and obtain a high-quality output image for characters.

In order to solve this problem, for example, an image processing apparatus according to the present invention has the following configuration. That is, an image processing apparatus for recognizing a character in a character candidate image portion in a document image, wherein a search is performed to search for a character having a small difference between a feature amount of the character candidate image portion and a standard feature amount stored in a recognition dictionary. Means, a comparing means for comparing the difference finally obtained by the searching means with a predetermined threshold value, and a comparison result of the comparing means. If the difference is smaller than the threshold value, The candidate image part is a character,
A comparison result between the character code output unit that outputs the corresponding character code based on the recognition dictionary and the comparison unit. If the difference is equal to or greater than the threshold, the target character candidate image portion is regarded as a non-character, and the character candidate image portion As image data.

[0007]

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

FIG. 1 is a block diagram of a copying machine according to an embodiment, and each component will be described below together with its operation.

An original 1001 passes through a CCD 102 from a lens 101 of an image reading unit 1002, enters an A / D converter 103 for performing A / D conversion, and a signal after the A / D conversion is input to an image processing unit 1003. Image processing unit 10
In step 03, the shading correction circuit 104 corrects the read image data appropriately, and the mode switching circuit 1
05.

In this case, the mode switching circuit 105 switches the output destination of the operation unit 116 depending on whether the operator is in the recognition copy mode or the normal copy mode. In the normal copy mode, the luminance data is converted into recording density data by the light density conversion circuit 106 and output to the image editing unit 107. The image editing unit 107 performs various editing processes based on the data processed according to the image processing settings stored in the RAM 115 in the CPU circuit 113. For example, trimming, color conversion, and the like. The image editing unit 107
Has a built-in memory for image development for one page, and performs various editing processes using this memory.
As described later, a font pattern or the like can be developed and output to the image recording unit 1004.

An image recording unit 1004 includes a control circuit such as a motor for transporting transfer paper and the like, a laser recording circuit unit for writing a video signal input from the image processing unit 1003 to a photosensitive drum, and a development control circuit for performing development. And an image is recorded on a recording medium such as recording paper.

On the other hand, when the recognition copy mode is set,
The output of the mode switching circuit 105 is the image recognition unit 400
It is switched to zero. The image recognition unit 4000 temporarily stores the output image data in the image buffer memory 1.
18 and recognizes characters in the document (generates character codes). The result is passed to the image data conversion unit 110, where an output font image for outputting the recognized characters is generated. The image data conversion unit 110 includes:
It has a font memory for each typeface (in the embodiment, outline font data), and generates a font pattern of a specified typeface.

The operation unit 112 is connected to the image processing unit 100.
It has various key groups and various LED groups for instructing an image copy operation such as image editing contents, the number of copies, and a magnification ratio for 3 and a display unit for displaying the contents at the time of operation.

FIG. 2 is a sectional view showing the structure of the image copying apparatus in the embodiment.

In FIG. 2, reference numeral 1 denotes a document feeder serving as a document feeder, which feeds the placed documents one by one or two at a predetermined position on the platen glass surface 2 continuously. Reference numeral 3 denotes a scanner constituted by a lamp, a scanning mirror 5, and the like. The scanner 3 is placed on the platen glass surface 2 by the document feeding device 1, and when a copy instruction or a character recognition instruction is given from the operation unit 116,
The scanner scans in a predetermined direction (horizontal direction in the drawing), passes the original reflected light through the lens 8 via the scanning mirrors 5 to 7, and forms an image on the image sensor unit 9.

Reference numeral 100 denotes a control unit (mainly a printed circuit board) on which the various circuits shown in FIG. 1 are mounted. An exposure unit 10 generates a light beam based on image data from an image editing unit 107 in the control unit 100. The light beam is swept in the horizontal direction by a polygon mirror, which scans and exposes the photoconductor 11 via a mirror or the like. Developing devices 12 and 13 visualize the electrostatic latent image formed on the photoreceptor 11 with a developer (toner) of a predetermined color. Reference numerals 14 and 15 denote transfer paper stacking portions, on which recording media of a fixed size are stacked and stored, and which are fed to a position where a registration roller is provided by driving a feeding roller and are aligned with the image formed on the photoreceptor 11. The sheet is fed again in a timed manner.

Reference numeral 16 denotes a transfer / separation charger, which transfers the toner image developed on the photoconductor 11 to a transfer sheet, and then separates from the photoconductor 11 and is fixed by a fixing unit 17 via a transport belt. Reference numeral 18 denotes a paper discharge roller which stacks and discharges the transfer-receiving paper on which image formation has been completed on the tray 20. Reference numeral 19 denotes a directional flapper which switches the transport direction of the transfer-receiving sheet on which image formation has been completed between the paper discharge port and the internal transport direction, and prepares for a multiplex / double-sided image forming process.

FIG. 3 shows an operation unit 116 used in this embodiment.
FIG.

In the figure, reference numeral 5001 denotes a power switch for controlling power supply to the apparatus main body. A reset key 5002 operates as a key for returning to the standard mode (normal copy mode) during standby. Reference numeral 5003 denotes a copy start key. A clear key 5004 is used to clear a numerical value such as the number of copies.

Reference numeral 5005 denotes an ID key.
05 enables copying operation for a specific operator,
For other operators, the copying operation can be prohibited unless an ID is input with the ID key. 5
Reference numeral 006 denotes a stop key, which is used to interrupt or stop copying. Reference numeral 5007 denotes a guide key, which is used when the user wants to know each function. Reference numeral 5008 denotes an up cursor key, which is a key for moving a pointer upward on each function setting screen. 50
09 is a down cursor key, 5010 is a right cursor key and a left cursor key, and the function is the same as that of the up cursor key 508 except that the direction is different.

Reference numeral 5013 denotes 50 in each function setting screen.
Press this key to execute what is output to the lower right of the screen 52. A standard size reduction key 5014 is used to reduce the standard size to another standard size. 501
Reference numeral 5 is used to select the same-size copy. A standard enlargement key 5016 is used to enlarge the standard size to another standard size. A cassette selection key 5017 selects a cassette stage to be copied. A copy density adjustment key 5018 reduces the density. An AE key 5019 automatically adjusts the copy density with respect to the density of the original. A copy density adjustment key 5020 increases the density. 5021 is a key for specifying the operation of the sorter. Reference numeral 5022 denotes a preheating key for turning on a preheating mode.
Used for / OFF. 5023 is an interrupt key,
Press this button to interrupt and copy during copying.
Reference numeral 5024 denotes a numeric keypad, which is used to input numerical values.

Reference numeral 5025 denotes a marker processing key for performing various kinds of image editing processing, for example, setting trimming, masking, and partial processing (contour processing, halftone processing, shadowing processing, negative / positive processing). Reference numeral 5026 denotes a pattern processable key which is used to express a color in a pattern or to express a color by a density difference. A color erasing key 5027 is used to erase a specific color. An image quality key 5028 is used to set the image quality. A negative / positive key 5029 is used when performing a negative / positive process. Reference numeral 5030 denotes an image create key, which includes contour processing, shadowing processing, halftone processing,
Used to perform italic, mirror, and repeat processing. Reference numeral 5031 denotes a trimming key used to designate an area and perform trimming. A masking key 5032 is used to designate an area and perform masking. A partial processing key 5033 specifies an area, and then specifies partial processing (contour processing, halftone processing, shadowing processing, negative / positive processing). Reference numeral 5034 denotes a frame erasing key, which is used when erasing a frame in accordance with a mode. The modes are sheet frame erasing (creating a frame for the sheet size), document frame erasing (creating a frame according to the document size. With original size specified), and book frame erasing (frames matching the spread size of the book). Create a blank in the center. Book spread size specified). A binding margin key 5035 is used to create a binding margin at one end of a sheet.

Reference numeral 5036 denotes a movement key for moving an image in a desired area, which is used when moving is desired.
The movement includes parallel movement (up, down, left and right), center movement, corner movement, and designated movement (point designation). 5037
Is a zoom key, which can set a copy magnification from 25% to 400% in increments of 1%. Further, the main scanning and the sub-scanning can be set independently. The magnification of the image is controlled by controlling the moving speed of the scanner 3 in the sub-scanning direction of the document, and the image data read out from the scanner 3 is thinned out or interpolated in the main scanning direction of the document. Do with.

Reference numeral 5038 denotes an automatic scaling key, which automatically enlarges or reduces the size according to the size of the copy sheet. In addition, the main scanning and the sub-scanning can be automatically scaled independently. Reference numeral 5039 denotes an enlargement continuous shooting key, which is used when one original is enlarged to a plurality of copies and copied. Reference numeral 5040 denotes a reduced layout key, which is used when copying a plurality of originals by enlarging / reducing them. Reference numeral 5043 denotes a continuous shooting key which is used to divide a copy area on the glass surface of the original platen into two right and left sides and to perform continuous copying in which two copies are automatically made (page continuous shooting, double-side continuous shooting). Reference numeral 5044 denotes a double-sided key, which is used when two-sided output is desired (single-sided double-sided, page continuous double-sided, double-sided double-sided). Reference numeral 5045 denotes a multiplex key, which is used when multiplexing is desired (multiplexing, page continuous copying multiplexing). 50
Reference numeral 46 denotes a memory key, which is used when a mode using a memory is desired (memory synthesis, area synthesis, watermark synthesis). A projector key 5047 is used when using the projector. Reference numeral 5048 denotes a printer key, which is used when setting for a printer.
Reference numeral 5050 denotes an original mixed key which is used when originals are mixed when copying using the feeder. Reference numeral 5051 denotes a mode memory key, which is used when registering a copy mode set for copy, when calling the registered copy mode. Reference numeral 5052 denotes a liquid crystal display with a touch panel, which displays the state of the apparatus, the number of copies, the copy magnification, and the copy paper size.

Even when the mode setting of the recognition copying mode or the normal copying mode described above is designated, the mode can be shifted to by touching one of the menus displayed on the liquid crystal display 5052 with a finger. It is possible.

Next, a character recognition unit which is a feature of the embodiment will be described in detail with reference to FIG.

The image stored in the image buffer memory 118 is binarized by the character extracting unit 108 into two orthogonal images.
The average character mesh size including one character is detected from the projection in the direction (histogram of black dots), and the character image is divided accordingly. The character images in each of the divided meshes are normalized, and the vector analysis unit 109 obtains a feature amount as an unknown character vector. Recognition control unit 111
When the feature amount (the feature amount of an unknown character) obtained in this way is input, the feature of the similar standard pattern before and after the obtained feature amount follows the tree structure of the dictionary of the control unit. By following the closer path,
A downward search is performed, and the character having the shortest distance feature obtained in the final stage is output as a recognized character. The mesh size is notified to the CPU 113, and the CPU 113 notifies the image data conversion unit 110 of the size.

FIG. 4 shows the above recognition method using a tree-structured dictionary in the case of an unknown character (here, the character “king”).

Here, i of one stage in the tree structure dictionary
The standard pattern vector (feature amount) of the group
vi, the input vector of the unknown character (the characteristic amount of the character image “king”) is * u, and the distance di between the vector * vi and the vector * u is: di = Σ | * Vik- * uk | Here, Σ indicates the sum of K = 0 to (r−1). Also, * vi = {vi0, vi1, vi2,
... vi (r-1)} and * uk = {u0, u1, u2, ... u (r-
1)}.

The group with the smallest distance di calculated in this way is selected, and thereafter, similarly, descends to the second stage and the third stage to form a path reaching the character "king" in the final stage. Will be.

By the way, when the image included in each mesh is recognized by using the dictionary, when the image in the mesh is not a character, for example, a part of a line segment forming a table graphic is In this case, a specific character or symbol stored in the dictionary is determined to be the shortest distance regardless of the absolute value of the distance. Therefore, for example, a part of the ruled line
When going straight to the number "one", that part is converted to the number "ten". Therefore, it is determined that the distance to the character obtained in the final stage is Ve.

In the case of Ve <K (a constant set in advance)..., A character is not in the mesh otherwise, it is determined that the inside of the mesh is not a character.

That is, a character is determined only when the distance Ve is equal to or less than a predetermined value K, that is, when the likelihood with the character listed as the first candidate is higher than a certain degree, and otherwise, it is determined as image information.

Therefore, if Ve <K in the case of FIG. 4, the image data conversion unit 110 determines that the character "king"
The dot data is generated from the font ROM held by the byte code. Conversely, when V ≧ K, the image data stored in the image buffer memory is sent to the image editing unit as it is.

By performing the above processing on all the meshes, the character portion is recorded in the recognized character font, and the area other than the character is recorded as a reproduced image in which both are mixed as image information similar to a normal copying machine. Display becomes possible.

When an image is reproduced by mixing image information and font data as described above, a higher-quality reproduced image can be obtained by using font data that matches the font of the characters on the document. Recognition of a font in this embodiment will be described with reference to a flowchart shown in FIG.

As described above with reference to FIG.
Then, the mesh of the character image in the input image is determined, the character image is divided by the determined mesh size, and the feature amount is obtained from the image signal in each mesh. Then, step S1
In step 1, character recognition is performed with reference to the dictionary.

The feature amount (stored in the dictionary) of the first candidate character (the character at the bottom in FIG. 4) obtained by recognition is obtained from the image in the mesh of interest. It is determined whether or not the distance (difference) from the feature value is less than a threshold value K. If it is determined that the distance is less than the threshold K, the process proceeds to step S13, and the image in the mesh of interest may be determined to be a character image. Is stored in a memory (not shown) provided along with the position information where the mesh exists (step S13).

Further, in this case, since the recognized image clearly has a high probability of the character, the character feature amount Sj is calculated. This information is the basis of this calculation. If the feature amount includes the width and direction of the character, the feature amount may be used. If there is no information about the same, but different for Mincho, etc., the calculation is performed including the image data in the mesh. Then, the calculated character features are sequentially added (step S15). Although an example of simple addition is shown in the figure, in fact, there is only one of the relationship between the Mincho style and the Gothic style, and italics fall into another category. Therefore, the character feature amount obtained for each mesh is added to the variable indicating the degree of the Mincho style (or the degree of the Gothic style) and the variable indicating the degree of the italic style. .

If the distance is equal to or greater than the threshold value K, the image data is not a character, or the character is not reliable if it is recognized as a character. Through 110,
The image is supplied to the image editing unit 107, and is treated as an image similar to a normal copy. In this case, the processing of steps S13 to S15 is skipped, and thus does not affect the font determination.

Thus, in step S16, the above processing is repeatedly performed on the images in all the meshes.

When the processing for all the meshes is completed, the process proceeds to step S17, where the font is determined based on the font feature value S obtained in step S15, and the determination is sent to the image data conversion unit 110. For example, by dividing the total amount indicating the degree of the Mincho style by the total number of meshes, the average degree of the Mincho style per character is calculated by dividing the total amount indicating the degree of the italic type by the total number of meshes. The degree of typical italic is calculated. If the value indicating the average degree of Mincho is greater than or equal to the threshold value, it is determined to be Mincho, otherwise it is determined to be Gothic.
The same applies to italics.

Upon receiving this notification, the image data conversion unit 110 selects a font memory of the corresponding font and generates a font pattern based on the character code stored in its own memory. At this time, if it is determined that the character is italic, it goes without saying that an italic character pattern is generated. Then, the generated font pattern and its development position information are output to the image editing unit 107.

Since the image editing unit 107 has a memory for developing an image of one page as described above, it stores the character pattern data supplied from the image data conversion unit 110 and its position information, Based on the image data that has not been recognized as a character and its position information, they are combined, and the recording operation is performed after the completion of image combination and development for one page.

As a result, the font is determined based on the image data in the mesh which is likely to be a character, so that the accuracy of the font determination can be improved.

In the above example, only the two types of fonts, Mincho and Gothic, and the determination of whether it is italic are shown. However, the font is determined in other or additional forms. Of course, it may be done.

Next, the setting of the threshold value K for determining as a character in the embodiment will be described.

FIG. 6 shows a screen of the display unit 5052 with a touch panel of the operation display unit shown in FIG.
Only the part related to the setting is shown.

When normal copying is possible, a screen 420 is displayed, and a recognition copy mode selection display key 421, a character / photo mode selection display key 422, and a photo mode selection display key 423 are displayed according to the type of the document or the purpose of copying. Have been.

In the embodiment, the key 421 is touched,
When the recognition copy is selected, the above-described recognition copy mode is executed. Also, the user setting key 42
When 4 is pressed, the screen shifts to 430. This screen has a recognition copy mode setting display key 431 that can be set by the user in the conventional copying machine. .

When the recognition copy mode setting display key 431 is depressed to shift to the screen 440, the lever display 44 is displayed.
Keys 441 and 443 for moving 2 right and left are displayed. That is, the value of the constant K is set according to the position of the lever display 442. Now, when the lever 442 is moved to the left by operating the key 441, it is assumed that the document is almost composed of only characters. At this position, there is a possibility that a part of the graphic exists in the mesh. Since the number is small, the threshold value K is set to a sufficiently large value. In other words, regardless of the value of the shortest distance, the character obtained at the shortest distance is most likely to be a document character, so that it is possible to recognize and copy even characters that are difficult for humans to read with almost no restrictions imposed by the constant K. It is possible.

Conversely, the operation of the key 441 causes the lever position 44
By moving 2 to the right, it is assumed that there are few characters in the original document and images of other attributes such as graphics are mixed, and at this position, there is a high possibility that a part of the graphics exists in the mesh. Therefore, K is set to a small value.

The lever position 442 and the constant K are set in advance in correspondence with each other as described above.

The display, setting and constant K are stored and held in the RAM 115 which is held by a backup battery or the like.

Further, the screen 440 has a typeface copy setting display key 446 and a typeface conversion setting display key 445, and is normally selected and set to one of the modes. In typeface copy mode,
As described above, even the typeface of the characters in the original is recognized, and the image is reproduced with the recording data obtained by developing the font in the typeface. At this time, the typeface determination may be made in units of one page of the document.

On the other hand, in the typeface conversion mode, an image is reproduced with recording data obtained by developing a font in a predetermined typeface regardless of the typeface of the original. If this is selected, steps S15 and S17 described above become unnecessary and may be skipped.

Return key 4 to end each setting
Pressing 44 or 432 returns to the screen 420.

The setting of K is arranged on the normal screen 420,
Each operator can also set each document. In this case, after a lapse of a predetermined time from the end of the copying operation,
The mode is returned to the standard K value stored in No. 15 or the set character style setting mode.

<Explanation of Modification> The example in which the operator sets the constant K has been described. However, if the rough state of the document can be grasped, the variable K can be automatically and variably set for each document. That is, the image is read in the course of conveying the original to the original plate by the original feeding device 1 serving as the original feeding means in FIG. 1, and the presence of a halftone image such as a picture can be determined from the density histogram of the original.

FIG. 7 is an example of a typical density distribution. The density distribution in FIG. 9A is an example in which one peak is provided in each of the almost white level and the intermediate density portion. The former peak indicates the background portion of the document and occupies 90% of the number of previous pixels, and the other is a character portion Is shown. That is, the density distribution A indicates a typical document having only a typical character line image.

On the other hand, FIG. 7B shows an example of an image including a halftone in which the density distribution B includes halftones. There are many small peaks from white to black, and a light density area other than the density distribution clearly indicated by a character is continuous. It can be recognized that it exists. Usually, it is possible to identify a document including a halftone portion in the document from the density distribution.

Now, in order to recognize the ratio of characters occupying a document including a character line drawing, that is, a document similar to the density distribution A, the image is binarized, and the positions of the pixels binarized to black in the image are binarized. Are projected in two directions perpendicular to the paper surface of the manuscript, respectively.
Evaluate the arrangement of black pixels.

FIG. 8A shows an example in which a horizontally written original is projected in the horizontal direction of each pixel line, and a repeated mountain pattern corresponding to each row is obtained. In the projection in the figure, the center of each peak is the center of each row, and the valley portion indicates the space between rows. Therefore, it can be determined that the valley of the projection in the figure is a line gap, and the mountain portion is a character portion. In other words, as shown in the figure, if the peaks and valleys exist alternately at approximately the same interval, it can be determined that most of the original is composed of characters.

In the projection shown in FIG. 9B, a black dot also exists in the valley, and it can be determined that there is a high possibility that a figure other than a character line, a line, a different character, or the like exists.

Therefore, in the present embodiment, the ratio occupied by the characters is evaluated by the “maximum value L of the valley” shown in FIG.

That is, it is determined that the ratio of characters is higher as the value of L is closer to 0, and that the ratio of mixed graphics is higher as the value of L is higher. If the projection is evaluated in two orthogonal directions, the presence of both vertical and horizontal rows can be recognized.

As described above, the character ratio can be predicted from the density distribution and the projection of the document.
Can be set automatically.

The above processing can be performed prior to the character recognition processing scan. In particular, the projection processing is processing performed when a document is divided into meshes, and can also be performed during the character recognition processing scan. It is.

The means for recognizing the character ratio is not limited to the one disclosed in the present embodiment. For example, another means such as a distribution of run lengths of continuous black pixels in a binary image may be used. It goes without saying that a similar effect can be obtained.

In the embodiment, an example in which the present invention is applied to a copying machine will be described. However, image data to be output or recognized is not read by a scanner, but is stored in a recording medium such as a floppy disk. The present invention may be applied to a case where the output is not limited to printing but is displayed on a display device, so that the present invention is not limited thereto.

Furthermore, a system was constructed with a source of image data to be recognized (scanner, storage medium, etc.), an output device of the recognition result (printer, display device, etc.), and a processing device interposed therebetween. In this case, the present invention can be applied. In this case, the processing device could be constructed with a general-purpose information processing device such as a personal computer.

Therefore, the present invention may be applied to a system including a plurality of devices (for example, a host computer, an interface device, a reader, a printer, etc.)
An object of the present invention is to provide a storage medium in which software program codes for realizing the functions of the above-described embodiments are recorded.
It is needless to say that the present invention is also achieved by supplying the data to a system or an apparatus and causing a computer (or CPU or MPU) of the system or the apparatus to read and execute the program code stored in the storage medium.

In this case, the program code itself read from the storage medium realizes the function of the above-described embodiment, and the storage medium storing the program code constitutes the present invention.

As a storage medium for supplying the program code, for example, a floppy disk, hard disk, optical disk, magneto-optical disk, CD-ROM, CD
-R, a magnetic tape, a nonvolatile memory card, a ROM, or the like can be used.

When the computer executes the readout program code, not only the functions of the above-described embodiment are realized, but also the OS (Operating System) running on the computer based on the instruction of the program code. ) May perform some or all of the actual processing, and the processing may realize the functions of the above-described embodiments.

Further, after the program code read from the storage medium is written in a memory provided in a function expansion board inserted into the computer or a function expansion unit connected to the computer, based on the instruction of the program code, It goes without saying that the CPU included in the function expansion board or the function expansion unit performs part or all of the actual processing, and the processing realizes the functions of the above-described embodiments.

As described above, according to the present embodiment,
Image degradation due to erroneous recognition can be suppressed. In addition, the character recognition accuracy can be improved. Furthermore, a good reproduced image is adaptively obtained according to the document.

[0078]

As described above, according to the present invention, a character code is output for an image portion likely to be a character in a document image, and a corresponding image in the document image is output otherwise. Thus, the reproducibility of the entire document image can be improved, and a high-quality output image of characters can be obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram of an apparatus according to an embodiment.

FIG. 2 is a sectional structural view of an image copying apparatus to which the embodiment is applied;

FIG. 3 is a top view of the operation unit 116 according to the embodiment.

FIG. 4 is a conceptual diagram of a tree-structured dictionary referred to in character recognition in the embodiment.

FIG. 5 is a flowchart showing processing contents in the embodiment.

FIG. 6 is a display 5052 with a touch panel of an operation display unit.
FIG. 6 is a diagram showing an example of the screen and its transition.

FIG. 7 is a diagram illustrating a histogram for each document type.

FIG. 8 is a diagram illustrating a histogram when a horizontally written document according to the embodiment is projected in a horizontal direction of each pixel line.

[Explanation of symbols]

 1002 Image reading unit 1003 Image processing unit 1004 Image recording unit 4000 Image recognition unit

(72) Inventor Junnosuke Kataoka 3-30-2 Shimomaruko, Ota-ku, Tokyo Canon Inc. (72) Inventor Makoto Kobayashi 3-30-2 Shimomaruko, Ota-ku, Tokyo Canon Inc. (72) Inventor Eiwa Honda 3-30-2 Shimomaruko, Ota-ku, Tokyo F-term in Canon Inc. (reference) 5B029 AA01 BB02 CC22 CC29 EE11

Claims (10)

[Claims] 1. An image processing apparatus for recognizing a character in a character candidate image portion in a document image, wherein a character having a small difference between a feature amount of the character candidate image portion and a standard feature amount stored in a recognition dictionary is determined. A search means for searching; a comparison means for comparing the difference finally obtained by the search means with a predetermined threshold; and a comparison result of the comparison means, wherein the difference is smaller than the threshold. Is a comparison result between the character code output unit that outputs the corresponding character code based on the recognition dictionary and the comparison unit. If the difference is greater than or equal to the threshold, the target character candidate image portion Is a non-character
An image output unit that outputs the character candidate image portion as image data. 2. A character pattern generating means for generating a character pattern based on the character code output by the character code output means, a character pattern generated by the character pattern generating means, and a character pattern output by the image output means. 2. The image processing apparatus according to claim 1, further comprising: an image forming unit configured to form a visible image by combining the image data with the image data. 3. The apparatus according to claim 1, further comprising a font determining unit configured to determine a font based on a plurality of character candidate image portions for which the difference is determined to be smaller than the threshold by the comparing unit. 2. The image processing device according to claim 1. 4. A character pattern determining means for determining a character form based on a plurality of character candidate image portions for which the difference is determined to be smaller than the threshold value by the comparing means, 3. The image processing apparatus according to claim 2, wherein a character pattern corresponding to the determined font is generated. 5. The image processing apparatus according to claim 1, further comprising means for setting the value of the threshold value by a manual operation. 6. The image processing apparatus according to claim 1, further comprising: a detecting unit configured to detect a ratio of characters occupied in the document image; and an adjusting unit configured to adjust the threshold value according to the ratio detected by the detecting unit. 2. The image processing device according to claim 1. 7. The detecting means includes a histogram creating means for creating a histogram of dots in horizontal and / or vertical directions, and the adjusting means based on a minimum frequency value in the histogram created by the histogram creating means. The image processing apparatus according to claim 5, wherein the threshold is adjusted. 8. The apparatus according to claim 5, wherein said detecting means includes means for detecting a density distribution in a document image, and said adjusting means adjusts said threshold value according to said density distribution. The image processing apparatus according to any one of the preceding claims. 9. An image processing method for recognizing a character in a character candidate image portion in a document image, wherein a character having a small difference between a feature amount of the character candidate image portion and a standard feature amount stored in a recognition dictionary is determined. A search step of searching; a comparison step of comparing the difference finally obtained by the search means with a predetermined threshold; and a comparison result of the comparison step, wherein the difference is smaller than the threshold. Is a comparison result between a character code output step of outputting a corresponding character code based on the recognition dictionary and the comparison step. If the difference is equal to or greater than the threshold value, the target character candidate image part Is a non-character
An image output step of outputting the character candidate image portion as image data. 10. A storage medium storing a program code that functions as an image processing device that recognizes a character candidate image portion in a document image when the computer reads and executes the character candidate image portion. Means for searching for a character having a small difference between the character string and the standard feature value stored in the recognition dictionary; and comparing means for comparing the difference finally obtained by the search means with a predetermined threshold value When the comparison result of the comparison means is that the difference is smaller than the threshold value, the target character candidate image portion is a character, and a character code output means for outputting a corresponding character code based on the recognition dictionary; As a result of the comparison, if the difference is equal to or larger than the threshold, the target character candidate image portion is regarded as a non-character,
A storage medium storing a program code that functions as an image output unit that outputs the character candidate image portion as image data.