JP5041141B2 - Image processing apparatus, image enlarging apparatus, image encoding apparatus, image decoding apparatus, image processing system, and program - Google Patents

Description

  The present invention relates to an image processing apparatus that enlarges an image.

  Generally, the resolution at the time of reading an image is different from the resolution at the time of printing. For example, the number of dots per inch is 200 to 300 dpi in the former case, and 600 to 1200 dpi in the latter case. For this reason, when a character document is read and printed by a scanner or the like, the character image needs to be enlarged.

Japanese Patent Application Laid-Open No. 2004-228561 discloses a technique for improving character resolution by smoothing diagonal lines of a character image.
Patent Document 2 discloses a method for enlarging an image based on an edge direction.
Japanese Patent Application Laid-Open No. 2004-228561 discloses a method of determining a stepped portion that becomes a jagged factor of an inclined portion of an image contour and performing pixel replacement (black and white inversion) based on a determination result of the jagged factor.
However, these methods perform image enlargement (resolution conversion) assuming local smoothness of the image. Therefore, there is a problem that it is not known whether the assumption is correct.

  Non-Patent Document 1 discloses a method of generating a single high-resolution image by inputting a plurality of low-resolution images. However, there is a problem (registration problem) on how to align pixel positions of a low resolution image and a high resolution image. In general, accurate image registration involves the following difficulties.

  First, when performing registration of the entire image, the amount of calculation becomes very large. Second, if the movement is too great, the difficulty increases. In general, a registration method has been studied as a motion compensation method in the moving image encoding method, but in this case, only a portion having a small motion is targeted. Furthermore, no registration method has been established when the input image format is not limited. In the case of a halftone image such as a photograph, a method of differentiating the difference and moving it in a direction in which the difference becomes smaller can be used. However, in an image having many edges such as a character, a line drawing, or CG. Such a differentiation method is difficult to use. However, if the input is limited to characters, a method such as pattern matching may be used.

  In Patent Document 4, a character area is extracted from an input image and recognized as a character, a character code is output, various fonts corresponding to the character code are searched from a font database, and these characters are displayed together with the image in the character area. In this character matching unit, the font type, the character size, and the character color that best approximate the pixel of the character region are searched, and together with these data, the start position of the character part, the character spacing, A method in which a character code is output as encoded data is disclosed.

Patent Document 5 discloses a technique for preventing erroneous recognition in position detection by pattern matching.
Patent Document 6 discloses a technique for obtaining a wide range of images for the purpose with a small amount of calculation when a plurality of images partially overlapping each other are combined to obtain a wide range of images.
Patent Document 7 discloses an image composition apparatus and an image composition method that have high matching accuracy and a small amount of processing only by roughly superimposing the divided images to be synthesized manually when the divided images are synthesized. ing.
However, these methods are for creating panoramic images and the like, and are not intended for accuracy of one pixel or less.

JP-A-6-225122 JP-A-6-313151 JP-A-8-310057 JP-A-10-178638 Japanese Patent Laid-Open No. 2002-190021 Japanese Patent Laid-Open No. 2004-334843 Japanese Patent Laid-Open No. 10-108003 S.C.Park, M.K.Park, and M.G.Kang, "Super-Resolution Image Reconstruction -A Technical Overview-", IEEE Signal Processing Magagine, May, 2003, pp.21-36.

  The present invention has been made from the above-described background, and an object of the present invention is to provide an image processing apparatus capable of enlarging an image with high image quality at a low cost compared to the case where the present configuration is not provided. And

  In order to achieve the above object, an image processing apparatus according to the present invention includes an extraction unit that extracts a first image and a second image similar to the first image at a first resolution; And generating means for generating an image of the second resolution based on each image extracted by the extracting means and the phase of each image calculated with higher accuracy than one pixel at the first resolution. According to the present invention, it is possible to enlarge an image with high image quality at a lower cost than in the case where the present configuration is not provided. Specifically, it is possible to reduce blurring of characters, figures, etc., unevenness of curves, widths of thin lines, and unevenness of intervals.

  Preferably, the image processing apparatus further includes multi-value conversion means for converting the binary image into multi-values, and the extraction means is configured to output, from the binary input image, a first image and a second image similar to the first image. The image is extracted at a first resolution, and the generation means calculates each image extracted by the extraction means and multi-valued by the multi-value conversion means with higher accuracy than one pixel at the first resolution. A character image having a second resolution is generated based on the phase of each of the images. According to the present invention, when the input image is a binary image, it is possible to generate a high-resolution image with higher image quality than when the input image is a multi-value image.

Preferably, the multi-value conversion means performs multi-value processing by applying a low-pass filter to the image.
Preferably, the multi-value conversion means multi-values each character image by applying an enlargement method.

  Preferably, the image processing apparatus further includes multi-value conversion means for converting the binary image into multi-values, and the extraction means includes the first image and the first image from the multi-value image by the multi-value conversion means. A second image similar to the image is extracted at the first resolution.

  Preferably, the image processing apparatus further includes enlargement means for enlarging each image extracted by the extraction means to a second resolution higher than the first resolution, and the generation means is extracted by the extraction means and is added by the enlargement means. Based on each enlarged image and the phase of each image calculated with higher accuracy than one pixel in the first resolution, an image having the second resolution is generated.

  Preferably, the image processing apparatus further includes a character group image storage unit that stores a character group image including a plurality of character images, and the generation unit includes a predetermined number or more of second character images extracted by the extraction unit. Generating a character image having a second resolution based on each character image extracted by the extracting means and the phase of each character image calculated with higher accuracy than one pixel in the first resolution; When the number of second character images extracted by the extracting unit is less than a predetermined number, the character image is extracted from the character group image stored in the character group image storage unit and is similar to the first character image. A character image having a second resolution is generated based on the character image. According to the present invention, even when a similar character image is not included in the input image or the number of extracted similar character images is small, the character resolution can be increased.

  Preferably, the image processing apparatus further includes a character group image storage unit that stores a character group image including a plurality of character images, and the generation unit includes a predetermined number or more of second character images extracted by the extraction unit. Generating a character image having a second resolution based on each character image extracted by the extracting means and the phase of each character image calculated with higher accuracy than one pixel in the first resolution; If the second character image extracted by the extracting means is less than the predetermined number, the first character image is a character image extracted from the character group image stored in the character group image storage means, and Replace with a character image similar to the one character image.

  Preferably, the apparatus further includes character group image acquisition means for acquiring a character group image including a plurality of character images based on the size of the first character image extracted by the extraction means, and the generation means includes: When the number of second character images extracted by the extraction unit is a predetermined number or more, each character image extracted by the extraction unit and each character image calculated with higher accuracy than one pixel at the first resolution The character group acquired by the character group image acquisition means when the character image of the second resolution is generated on the basis of the phase and the number of second character images extracted by the extraction means is less than a predetermined number. A character image having the second resolution is generated based on a character image extracted from the image and similar to the character image similar to the first character image. According to the present invention, even when a similar character image is not included in the input image or the number of extracted similar character images is small, the character resolution can be increased.

  Preferably, the character group image acquisition means acquires a plurality of character group images having different phases. According to the present invention, it is possible to increase the resolution of an image.

  Preferably, the image processing apparatus further includes corresponding character image generation means for generating a character image corresponding to the size of the first character image extracted by the extraction means, wherein the generation means is the first character image extracted by the extraction means. When the number of the two character images is equal to or greater than the predetermined number, the character images extracted by the extracting unit and the phase of each character image calculated with higher accuracy than one pixel at the first resolution are used. When a character image having a resolution of 2 is generated and the number of second character images extracted by the extracting unit is less than a predetermined number, the second resolution is further determined based on the character image generated by the corresponding character image generating unit. Generate a character image.

  The second image processing apparatus according to the present invention includes an extraction unit that extracts a first image and a second image similar to the first image at a first resolution, and each of the extraction units extracted by the extraction unit. Based on the size adjusting means for adjusting the size of the image, each image adjusted by the size adjusting means, and the phase of each image calculated with higher accuracy than one pixel in the first resolution, the second Generating means for generating a resolution image. According to the present invention, an image can be enlarged with high image quality even when the number of appearances of the image is small. In particular, since the resolution of characters with a large number of points such as titles can be increased, the visual and psychological effects are great.

  The third image processing apparatus according to the present invention includes an extraction unit that extracts a first character image and a second character image similar to the first character image at a first resolution, and the extraction unit extracts the first character image. Generating means for generating a character image having a second resolution higher than the first resolution based on each character image thus obtained and the phase of each character image calculated with higher accuracy than one pixel in the first resolution. And have. According to the present invention, it is possible to enlarge a character image with high image quality at a lower cost than when the present configuration is not provided. Particularly, since the ratio of the character area to the entire input image is small, the motion estimation area can be reduced.

  Preferably, the generation means interpolates the pixel value of the character image at the second resolution from the pixel value of each character image based on the phase of each character image at the first resolution. According to the present invention, it is possible to eliminate the need for a device for scanning in with a phase shift. Also, the more characters that are identical, the higher the resolution. In addition, it is possible to prevent a decrease in productivity due to a plurality of scans.

  According to a fourth image processing apparatus of the present invention, an extraction unit that extracts a first character image and a second character image similar to the first character image at a first resolution, and the extraction unit extracts the first character image. And removing means for removing noise of the character image based on each character image thus obtained and the phase of each character image calculated with higher accuracy than one pixel in the first resolution, and removing the noise by the removing means And generating means for generating a character image having a second resolution higher than the first resolution based on the character image thus formed and the phase of the character image. According to the present invention, a better image can be generated.

  An image enlarging apparatus according to the present invention includes an extraction unit that extracts a first character image and a second character image similar to the first character image at a first resolution, and each of the extraction units extracted by the extraction unit. Generating means for generating a character image having a second resolution higher than the first resolution based on the character image and the phase of each character image calculated with higher accuracy than one pixel at the first resolution; Arranging means for arranging the character image of the second resolution generated by the generating means at a position in the image of the second resolution calculated based on the position of the character image extracted by the extracting means.

  An image encoding apparatus according to the present invention includes a partial image extraction unit that extracts a first partial image and a second partial image similar to the first partial image at a first resolution, and the partial image extraction unit. Representative image generation means for generating a representative image of the second resolution based on each partial image extracted by the above and the phase of each partial image calculated with higher accuracy than one pixel in the first resolution; Coding means for coding the representative image generated by the representative image generating means and the phase and position of each partial image extracted by the extracting means; According to the present invention, it is possible to perform encoding at a high compression rate while suppressing deterioration in image quality.

  An image decoding apparatus according to the present invention includes a decoding unit that decodes code data in which a representative image of a second resolution and a phase and a position of a partial image of the first resolution are encoded, and the decoding A partial image generating means for generating a partial image of the first resolution based on the representative image of the second resolution decoded by the means and a phase of the partial image of the first resolution; and the decoding means And a partial image arrangement unit that arranges the partial image generated by the partial image generation unit on the output image based on the decoded position. According to the present invention, it is possible to restore an image encoded at a high compression rate while suppressing deterioration in image quality.

  Preferably, the image processing apparatus further includes magnification / position control means for controlling the magnification of the partial image generated by the partial image generation means and the position where the partial image arrangement means is arranged. According to the present invention, the quality of an encoded image can be improved.

  An image processing system according to the present invention includes a partial image extraction unit that extracts a first partial image and a second partial image similar to the first partial image at a first resolution, and the partial image extraction unit. Representative image generation means for generating a representative image of the second resolution based on each extracted partial image and the phase of each partial image calculated with higher accuracy than one pixel in the first resolution; Encoding means for encoding the representative image generated by the representative image generating means, the phase and position of each partial image extracted by the extracting means, and decoding the code data encoded by the encoding means Decoding means for generating, a representative image decoded by the decoding means, a partial image generating means for generating a partial image of the first resolution based on the phase of the partial image, and decoding by the decoding means Turned into It was based on the position, and a partial image arrangement means for arranging the partial image generated by the partial image generating means in the output image. According to the present invention, it is possible to suppress deterioration in image quality and to encode and decode at a high compression rate.

A first program according to the present invention is an image processing apparatus including a computer.
An extraction step for extracting a first image and a second image similar to the first image at a first resolution, each extracted image, and calculation with higher accuracy than one pixel at the first resolution Based on the phase of each image thus generated, the computer of the image processing apparatus executes a generation step of generating an image of the second resolution.

  The second program according to the present invention includes an extraction step of extracting a first image and a second image similar to the first image at a first resolution in an image processing apparatus including a computer, and the extraction A second resolution based on the size adjustment step for adjusting the size of each of the images, the adjusted images, and the phase of each image calculated with higher accuracy than one pixel in the first resolution. Generating the image of the image processing apparatus.

  The third program according to the present invention includes an extraction step of extracting a first character image and a second character image similar to the first character image at a first resolution in an image processing apparatus including a computer; Generating a character image having a second resolution higher than the first resolution based on each extracted character image and the phase of each character image calculated with higher accuracy than one pixel in the first resolution. A generating step to be executed by the computer of the image processing apparatus.

  A fourth program according to the present invention includes an extraction step of extracting a first character image and a second character image similar to the first character image at a first resolution in an image processing apparatus including a computer. Removing a noise of the character image based on the extracted character image and the phase of the character image calculated with higher accuracy than one pixel in the first resolution; and removing the noise Based on the obtained character image and the phase of the character image, the computer of the image processing apparatus executes a generation step of generating a character image having a second resolution higher than the first resolution.

  The fifth program according to the present invention includes an extraction step of extracting a first character image and a second character image similar to the first character image at a first resolution in an image enlargement apparatus including a computer. Generating a character image having a second resolution higher than the first resolution based on each extracted character image and the phase of each character image calculated with higher accuracy than one pixel in the first resolution. The image enlargement step, and an arrangement step of arranging the generated character image of the second resolution at a position in the image of the second resolution calculated based on the position of the extracted character image. Let the computer of the device execute.

  A sixth program according to the present invention is a partial image for extracting a first partial image and a second partial image similar to the first partial image at a first resolution in an image encoding device including a computer. A representative image that generates a representative image of the second resolution based on the extraction step, each extracted partial image, and the phase of each partial image calculated with higher accuracy than one pixel in the first resolution. The computer of the image encoding device is caused to execute a generating step, an encoding step for encoding the generated representative image, and the phase and position of each partial image extracted by the extracting means. According to the present invention, it is possible to perform encoding at a high compression rate while suppressing deterioration in image quality.

  A seventh program according to the present invention decodes code data in which a representative image of the second resolution and a phase and a position of a partial image of the first resolution are encoded in an image decoding apparatus including a computer A partial image generation step of generating a partial image of the first resolution based on the decoded representative image of the second resolution and the phase of the partial image of the first resolution; Based on the decoded position, the computer of the image decoding apparatus executes a partial image arrangement step of arranging a partial image generated by the partial image generation means in an output image. According to the present invention, it is possible to restore an image encoded at a high compression rate while suppressing deterioration in image quality.

  According to the image processing apparatus of the present invention, it is possible to enlarge a character image with high image quality at a lower cost than when the present configuration is not provided.

First, in order to help understanding of the present invention, its background and outline will be described.
FIG. 1 is a diagram for explaining the quality of a character image when the character image is enlarged.
FIG. 1A illustrates a character image read by a scanner or the like. FIG. 1B is a desired character image obtained by enlarging the character image shown in FIG. However, when the read character image is enlarged, the fine lines constituting the character may be blurred as shown in FIG. 1C depending on the size of the character or the relationship between the resolution before and after the enlargement. . Alternatively, as shown in FIG. 1D, a smooth curve may be uneven, or the line width and line spacing may become uneven. In such a case, the quality of characters is significantly reduced.

  Specifically, when enlarging a character image, blur and curved irregularities are a problem in the multi-value data format, and unevenness of width and spacing and curved irregularities are a problem in the binary data format. Also, jaggy is a problem in the nearest neighbor interpolation method, blur is remarkable in the linear interpolation method, and unevenness in the curve is a problem in the convolution interpolation method of the sampling function.

  On the other hand, there is a technique for configuring one image having a resolution higher than the imaging resolution based on a plurality of images that are slightly shifted in imaging position. By using such a technique, the quality of enlarged characters can be improved.

FIG. 2 is a diagram for explaining a method for generating an image having a higher resolution than the imaging resolution.
FIG. 2A shows character image data 1 to 4 having pixel shift phases 1 to 4 acquired from the original image. The character image data 1 to 4 are obtained by shifting the original image in units of 1/2 pixel. FIG. 2B shows an image having a resolution four times that of the original image generated based on the character image data 1 to 4 shown in FIG.

  However, when such a technique is used, it is necessary to slightly shift the imaging position by using a plurality of cameras with different imaging directions or by shifting the imaging optical system. For example, when a document is read by a multi-function peripheral equipped with a scanner device, the carriage position of the line sensor is slightly changed, and a plurality of scans are executed. For this reason, an increase in apparatus cost and a decrease in productivity due to scanning become problems.

FIG. 3 is a diagram for explaining the outline of image processing by the image processing apparatus according to the present invention.
As shown in FIG. 3A, the image processing apparatus according to the present invention extracts characters that appear multiple times in a one-page document or a document composed of a plurality of pages. Such a plurality of identical characters are often read at different phases with respect to the character (for example, the center of gravity of the character) when read by a scanner or the like. For this reason, the phases of the extracted characters are shifted from each other.

  The image processing apparatus according to the present invention calculates the phase of each extracted character as shown in FIG. The phase of each character is calculated, for example, from the relative position between the calculated center of gravity and the sampling grid point by calculating the center of gravity of the character. Furthermore, as shown in FIG. 3C, the image processing apparatus according to the present invention generates a high-resolution character image based on the calculated phase and each extracted character image.

The target of the high resolution processing is not limited to a character image, but may be a partial image having the same shape, such as a graphic. In this case, the image processing apparatus according to the present invention extracts partial images that appear multiple times in the input image, calculates the phase of each partial image, and based on the calculated phase and each extracted partial image To generate a high-resolution partial image.
In the following, for ease of explanation, the image processing apparatus according to the present invention will be described by taking as an example a case where the target of high resolution processing is a character image.

Hereinafter, the image processing apparatus 2 according to the first embodiment of the present invention will be specifically described.
FIG. 4 is a diagram showing the hardware configuration of the image processing apparatus 2 and the image enlargement apparatus 3 described later, centering on the control apparatus 20, according to the embodiment of the present invention.
As shown in FIG. 4, the image processing apparatus 2 includes a control device 20 including a CPU 202 and a memory 204, a communication device 22 that transmits and receives data to and from an external computer via a network, a hard disk drive, and the like. Storage device 24, a display device such as a liquid crystal display, and a user interface (UI) device 26 including a keyboard and a pointing device.

  The image processing apparatus 2 or the like is, for example, a general-purpose computer in which a character image acquisition program 4 to be described later is installed, and acquires and acquires image data via the communication device 22, the storage device 24, the recording medium 240, or the like. Image processing is performed on the captured image. The image processing device 2 acquires image data optically read by a scanner function of a printer device (not shown) or the scanner device.

FIG. 5 is a diagram showing a functional configuration of the character image acquisition program 4 executed by the control device 20 of the image processing device 2 according to the first embodiment of the present invention.
As shown in FIG. 5, the character image acquisition program 4 includes a character cutout unit 40, a similar character extraction unit 42, a centroid calculation unit 44, and a high resolution character image generation unit 46. For example, the character image acquisition program 4 is supplied to the control device 20 via the communication device 22, loaded into the memory 204, and specifically uses hardware on an OS (not shown) operating on the control device 20. Executed. The character image acquisition program 4 may be stored in a recording medium 240 such as FD, CD, or DVD and supplied to the image processing apparatus 2. The following programs are also supplied and executed in the same manner. Further, all or part of the functions of the character image acquisition program 4 may be realized by an ASIC provided in the image processing apparatus 2 or the like.

  In the character image acquisition program 4, the character cutout unit 40 receives an input image having a predetermined resolution (first resolution) read by a scanner device or the like via the communication device 22, the storage device 24, or the recording medium 240. A character image included in the input image is cut out and output to the similar character extraction unit 42. The input image may be a one-page document or a document composed of a plurality of pages.

  The similar character extraction unit 42 receives the character images cut out by the character cutout unit 40, extracts similar character images, and outputs them to the centroid calculation unit 44 and the high resolution character image generation unit 46. Specifically, the similar character extraction unit 42 identifies the first character image from the plurality of clipped character images, and the first character image and the second character image similar to the first character image, To extract.

  The centroid calculation unit 44 calculates the centroid of each character image received from the similar character extraction unit 42, and outputs the calculated centroid value to the high-resolution character image generation unit 46. The centroid calculation unit 44 calculates the centroid with higher accuracy than one pixel in the first resolution. The center-of-gravity calculation unit 44 calculates the phase of each character image based on the relative position between the calculated center of gravity and the sampling grid point, and outputs the calculated phase to the high-resolution character image generation unit 46. Also good. The center-of-gravity calculation unit 44 may calculate the center of the character image or the center based on the width and height of the character image, for example. In this case, the center of gravity calculation unit 44 outputs the calculated center of the character image and the like.

  The high-resolution character image generation unit 46 determines the first character image based on each character image extracted by the similar character extraction unit 42 and the phase of each character image calculated with higher accuracy than one pixel in the first resolution. A character image having a second resolution higher than the resolution of is generated. Specifically, the high resolution character image generation unit 46 calculates the phase of the character image based on the relative position between the centroid of each character image calculated by the centroid calculation unit 44 and the sampling grid point, and extracts similar characters. A character image is generated based on the similar character image received from the unit 42 and the phase of each character image. When the phase of the character image is calculated by the gravity center calculation unit 44 and input to the high resolution character image generation unit 46, the high resolution character image generation unit 46 may use this phase.

The second resolution may be the same as the first resolution. In this case, the high-resolution character image generation unit 46 further performs resolution conversion processing on the character image having the second resolution higher than the first resolution, so that the second resolution is the same resolution as the first resolution. Generate a character image. Further, the second resolution may be lower than the first resolution. Also in this case, the high-resolution character image generation unit 46 further performs resolution conversion processing (reduction processing) on the character image having the second resolution higher than the first resolution, thereby reducing the resolution lower than the first resolution. A character image having the second resolution is generated.
The character image generation processing by the high resolution character image generation unit 46 will be described in detail later.

FIG. 6 is a diagram for explaining the character extraction process by the character extraction unit 40 and the similar character extraction process by the similar character extraction unit 42. FIG. 6A illustrates an input image, and FIG. The character image cut out from the input image by the character cutout unit 40 is illustrated, and FIG. 6C illustrates the similar character image extracted by the similar character extraction unit 42.
As illustrated in FIGS. 6A and 6B, the character cutout unit 40 cuts out a plurality of character images included in the input image. In this example, the character images “2” and “c” are each cut out from the input image.

  For example, the character cutout unit 40 may binarize the input image and cut out a block of continuous black pixels (for example, black pixels connected in the vicinity of 8) from the input image as one character image. One character image may be cut out by estimating a string and a character interval and estimating one character region. For example, the character cutout unit 40 may cut out a character image specified by an external OCR (Optical Character Reader) or the like. The character cutting method is not particularly limited.

  As illustrated in FIG. 6C, the similar character extraction unit 42 extracts a similar character image from the extracted character image. In this example, a character image similar to the character image “2” and a character image similar to the character image “c” are extracted. For example, the similar character extraction unit 42 determines whether or not these character images are similar by performing pattern matching between the first character image and the second character image. In this case, the similar character extraction unit 42 performs pattern matching between the character images.

  For example, the similar character extraction unit 42 may treat one character image data as one vector data and perform clustering to extract a character image similar to the character image. In this case, when the distance (for example, Euclidean distance) between the vector data representing the character image and the vector data representing the determination target character image is less than a predetermined value (ie, the two vector data) Is determined to be similar to the character image to be determined. A similar determination method is not particularly limited.

FIG. 7 is a diagram for explaining high-resolution character image data generation processing by the high-resolution character image generation unit 46.
FIG. 7A shows the sampling grid of the input image and the barycentric position of the character image data at the first resolution. In FIG. 7A, the rectangle indicated by the arrow a is the sampling grid of the input image, and the point where the broken line indicated by the arrow b intersects is the barycentric position of the character image data. In this example, description will be made using four sampling grids of the input image. First, as shown in FIG. 7B, the high-resolution character image generation unit 46 moves the phases of the four sampling grids based on the center of gravity of the character image data.

  FIG. 7C and FIG. 7D are diagrams illustrating a method for setting a sampling grid having a second resolution higher than the first resolution. In FIG. 7C, the circled number indicated by the arrow c exemplifies the value of the character image data at the first resolution. Here, the character image data is shown such that the circled numbers are represented on the grid points of the sampling grid at the first resolution. In FIG. 7D, a thick line indicated by an arrow d indicates a sampling grid for a high-resolution character image at the second resolution.

  When the phases of the four sampling grids at the first resolution are moved, the high-resolution character image generation unit 46 sets the sampling grid at the second resolution as shown in FIG. As shown in FIG. 7D, the phase of the sampling grid at the second resolution is moved based on the center of gravity of the character image data.

  FIG. 7E is a diagram for explaining a method for calculating the value of character image data at the second resolution. In FIG. 7E, bold circle numbers indicated by arrows e exemplify values of character image data at the second resolution. Here, the character image data at the second resolution is shown such that bold circle numbers are represented on the grid points of the sampling grid at the second resolution.

  Based on the phase of each character image at the first resolution, the high-resolution character image generation unit 46 interpolates the pixel value of the character image at the second resolution from the pixel value of each character image. In this example, the high resolution character image generation unit 46 applies the nearest neighbor interpolation method to interpolate the pixel values of the character image at the second resolution. That is, the high-resolution character image generation unit 46 selects the sampling grid point at the second resolution among the four values of the character image data at the first resolution (in this example, circled numbers “1” to “4”). Is selected as the value of the character image data at the second resolution. The interpolation method is not particularly limited, and other methods (for example, a linear interpolation method) may be applied.

FIG. 8 is a flowchart of the high-resolution character image acquisition process (S10) by the character image acquisition program 4.
As shown in FIG. 8, in step 100 (S100), the character cutout unit 40 of the character image acquisition program 4 receives the input image of the first resolution, cuts out the character image from the input image, and extracts the similar character extraction unit 42. Output for.

In step 102 (S102), the similar character extraction unit 42 extracts similar character images from the character images cut out by the character cutout unit 40 at the first resolution.
In step 104 (S104), the centroid calculating unit 44 calculates the centroid of each character image extracted by the similar character extracting unit 42 with higher accuracy than one pixel in the first resolution.

  In step 106 (S106), the high resolution character image generation unit 46 calculates the phase of each character image based on the center of gravity of each character image calculated by the center of gravity calculation unit 44, and is extracted by the similar character extraction unit 42. A character image having the second resolution is generated based on each similar character image and the phase of the similar character image.

  In step 108 (S108), the high resolution character image generation unit 46 outputs the generated high resolution character image data of the second resolution. The high resolution character image generation unit 46 may transmit the high resolution character image data to an external computer (not shown) via the communication device 22 or may store it in the storage device 24 or the recording medium 240.

Next, the image enlargement apparatus 3 according to the embodiment of the present invention will be specifically described.
The image enlargement apparatus 3 is a general-purpose computer in which the image enlargement program 5 is installed.
FIG. 9 is a diagram showing a functional configuration of the image enlargement program 5 executed by the control device 20 of the image enlargement apparatus 3 according to the embodiment of the present invention.
As shown in FIG. 9, the image enlargement program 5 includes a character cutout unit 40, a similar character extraction unit 42, a centroid calculation unit 44, a high resolution character image generation unit 46, a position calculation unit 50, and a character arrangement unit 52. 9 that are substantially the same as those shown in FIG. 5 are denoted by the same reference numerals. As described above, the image enlargement program 5 includes the character image acquisition program 4.

  In the image enlargement program 5, the position calculation unit 50 receives character position data of the character image cut out from the input image of the first resolution by the character cutout unit 40, and sets the character position based on the position of the character image. The position of the character image in the enlarged image of the second resolution at the corresponding position is calculated. The position calculation unit 50 outputs the calculated position to the character placement unit 52.

  The character placement unit 52 is generated by the high-resolution character image generation unit 46 and the position in the enlarged image of the second resolution calculated by the position calculation unit 50 based on the position of the character image extracted at the first resolution. A character image corresponding to the position and a character image corresponding to the position is received, and the character image having the second resolution is arranged at the position of the enlarged image. The character arrangement unit 52 arranges character images having the second resolution corresponding to the character images included in the input image having the first resolution on the enlarged image, and generates enlarged image data.

FIG. 10 is a flowchart of image enlargement processing (S20) by the image enlargement program 5.
As shown in FIG. 10, the image enlargement program 5 performs a high-resolution character image acquisition process (FIG. 8; S10), and acquires a character image of the second resolution corresponding to each of the character images of the first resolution. To do.
In step 200 (S200), the position calculation unit 50 of the image enlargement program 5 calculates the position of each character image in the enlarged image based on the position of each character image included in the input image having the first resolution.

In step 202 (S202), the character arrangement unit 52 arranges character images of the second resolution corresponding to the positions in the enlarged image calculated based on the positions of the character images extracted at the first resolution. .
In step 204 (S204), the character arrangement unit 52 outputs the generated enlarged image data. The character arrangement unit 52 may transmit the enlarged image data via the communication device 22 or may store it in the storage device 24 or the like.

Next, an image processing apparatus 2 according to the second embodiment of the present invention will be described.
The image processing apparatus 2 according to the present embodiment is a general-purpose computer in which a character image acquisition program 6 is installed. The character image acquisition program 6 is different from the character image acquisition program 4 in that a noise removal function is added to the character image acquisition program 4 (FIG. 5).

FIG. 11 is a diagram showing a functional configuration of the character image acquisition program 6 executed by the control device 20 of the image processing apparatus 2 according to the second embodiment of the present invention.
As shown in FIG. 11, the character cutout unit 40, the similar character extraction unit 42, the centroid calculation unit 44, the phase calculation unit 60, the representative character extraction unit 62, the phase representative value calculation unit 64, the noise removal unit 66, and the high resolution character image A generation unit 68 is included. 11 that are substantially the same as those shown in FIG. 5 are denoted by the same reference numerals. Thus, the character image acquisition program 6 includes the character image acquisition program 4.

  In the character image acquisition program 6, the phase calculation unit 60 determines the phase of each character image based on the relative position between the centroid calculated by the centroid calculation unit 44 with higher accuracy than one pixel at the first resolution and the sampling grid point. Is output to the representative character extraction unit 62 and the high-resolution character image generation unit 68.

  The representative character extraction unit 62 receives a plurality of similar character images extracted by the similar character extraction unit 42 and a phase of the character image calculated by the phase calculation unit 60, and receives a plurality of character images having a small phase difference. Extracted as a representative character image. The representative character extraction unit 62 sets a character image having a phase difference smaller than a preset threshold as a representative character image. The representative character extraction unit 62 outputs the extracted representative character images to the phase representative value calculation unit 64 and the noise removal unit 66.

  The phase representative value calculating unit 64 selects one phase representative value from the phases of the plurality of character images extracted by the representative character extracting unit 62 and outputs the selected phase representative value to the high resolution character image generating unit 68. Here, the phase representative value is calculated for each similar character image.

  The noise removing unit 66 calculates the noise of the character image based on each character image extracted by the similar character extracting unit 42 and the phase of each character image calculated with higher accuracy than one pixel in the first resolution. Remove. Specifically, the noise removal unit 66 performs a noise removal process on the character image by performing a smoothing process based on the plurality of representative character images extracted by the representative character extraction unit 62. The smoothing process is performed, for example, by obtaining an average value of pixel values at corresponding positions in the extracted representative character images. The noise removing unit 66 outputs the character image from which the noise has been removed to the high resolution character image generating unit 68.

  The high-resolution character image generation unit 68 uses the first resolution based on the character image from which noise has been removed by the noise removal unit 66 and the phase representative value of the character image calculated by the phase representative value calculation unit 64. A character image having a high second resolution and from which noise has been removed is generated. The high-resolution character image generation unit 68 uses the second resolution based on the character image from which the noise extracted by the similar character extraction unit 42 is not removed and the phase of the character image calculated by the phase calculation unit 60. May be generated.

FIG. 12 is a flowchart of the high-resolution character image acquisition process (S30) including the noise removal process by the character image acquisition program 6. Of the processes shown in FIG. 12, the same reference numerals are assigned to the processes that are substantially the same as those shown in FIG.
As shown in FIG. 12, a character image is extracted from the input image by the processing of S100 to S104, and the center of gravity of the extracted character image is calculated.

In step 300 (S300), the phase calculation unit 60 calculates the phase of each extracted character image based on the calculated center of gravity.
In step 302 (S302), the representative character extracting unit 62 extracts from the similar characters as representative character images having a small phase difference.

In step 304 (S304), the noise removal unit 66 performs noise removal processing based on each extracted character image and the phase of each character image, and generates a character image from which noise has been removed.
In step 306 (S306), the phase representative value calculation unit 64 selects one phase representative value from the phases of the extracted character images.

In step 308 (S308), the high resolution character image generation unit 68 removes noise based on the character image from which the noise has been removed and the phase representative value of the character image calculated by the phase representative value calculation unit 64. The character image having the second resolution is generated.
In step 310 (S310), the high resolution character image generation unit 68 outputs the generated high resolution character image data of the second resolution.

FIG. 13 is a diagram showing a functional configuration of a character image acquisition program 400 executed by the control device 20 of the image processing device 2 according to the third embodiment of the present invention.
As shown in FIG. 13, the character image acquisition program 400 has a configuration in which a multi-value quantization unit 402 and a binarization unit 404 are added to the character image acquisition program 4 shown in FIG. Of the components shown in FIG. 13, components substantially the same as those shown in FIG. 5 are denoted by the same reference numerals.

  In the character image acquisition program 400, the multi-value conversion unit 402 inputs a binary of a predetermined resolution (first resolution) read by a scanner device or the like via the communication device 22, the storage device 24, or the recording medium 240. An image is received, the input image is subjected to multi-value processing (gradation number increasing processing), converted to a multi-value image, and the multi-value image is output to the character cutout unit 40. The multilevel processing will be described in detail later.

  Therefore, the character cutout unit 40 receives the image multivalued by the multivalue conversion unit 402 and cuts out the character image included in the multivalued image. The similar character extraction unit 42 receives the multi-value character images cut out by the character cut-out unit 40, and extracts similar multi-value character images. The center-of-gravity calculation unit 44 calculates the center of gravity of each of the multi-value character images received from the similar character extraction unit 42. The high resolution character image generation unit 46 is based on the multivalued character images extracted by the similar character extraction unit 42 and the phase of each character image calculated with higher accuracy than one pixel at the first resolution. A multi-value character image having a second resolution higher than the first resolution is generated. The high-resolution character image generation unit 46 outputs the multi-value character image having the second resolution generated in this way to the binarization unit 404. Note that the high-resolution character image generation unit 46 may store multi-value character image data in the storage device 24 or may transmit the multi-value character image data to an external computer via the communication device 22.

  The binarization unit 404 receives the multi-value character image of the second resolution generated by the high-resolution character image generation unit 46, performs binarization processing on the multi-value character image, and performs binary processing. The image is converted into a character image, and a binary character image having the second resolution is generated. The binarization unit 404 performs, for example, threshold processing on the multi-valued character image to convert it into a binary image, and outputs this binary character image. The binarization unit 404 may store the binary character image data in the storage device 24 or may transmit it to an external computer via the communication device 22.

FIG. 14 is a diagram showing an overview of the multi-value process executed by the image processing apparatus 2 according to the embodiment of the present invention.
As illustrated in FIG. 14, the image processing apparatus 2 executes any one of a plurality of multi-value conversion methods to convert a binary image into a multi-value image. Here, the image processing apparatus 2 performs multi-value processing by applying a predetermined multi-value method.

  FIG. 14A illustrates a binary input image input to the image processing apparatus 2. For example, the input image is a 3 × 3 pixel image, black pixels are represented by “1”, and white pixels are represented by “0”. FIG. 14B to FIG. 14D show a first multi-value quantization method, a second multi-value quantization method, and a third multi-value quantization method, respectively.

  As shown in FIG. 14B, in the first multi-value conversion method, the multi-value conversion unit 402 converts the binary values “0” and “1” into, for example, “0” and “255”, respectively. . In the second multi-value conversion method, the multi-value conversion unit 402 multi-values each character image by applying an enlargement method. Here, the multi-value processing is used in combination with the interpolation method of the enlargement processing. Therefore, the binary input image is enlarged, and each pixel value is converted into, for example, one of 256 gradations. In the third multi-value conversion method, each pixel value is converted into a gray font by the multi-value conversion unit 402. For example, the multilevel conversion unit 402 performs multilevel processing by applying a low-pass filter to a binary input image. For example, the multi-value quantization unit 402 performs multi-value processing by calculating an average value of pixel values of pixels included in a predetermined region.

  In the case of the first method and the third method, the high-resolution character image generation unit 46 generates multi-value high-resolution character image data in the same manner as the method shown in FIG. That is, in the present embodiment, the high-resolution character image generation unit 46 determines the character image at the second resolution from the multivalued pixel value of each character image based on the phase of each character image at the first resolution. Interpolate pixel values. The high-resolution character image generation unit 46 applies a nearest neighbor interpolation method, a linear interpolation method, or the like as an interpolation method. The second method will be described later in detail.

FIG. 15 is a flowchart of the high-resolution character image acquisition process (S40) by the character image acquisition program 400. Note that, among the processes shown in FIG. 15, substantially the same processes as those shown in FIG. 8 are denoted by the same reference numerals.
As shown in FIG. 15, in step 400 (S400), the multi-value conversion unit 402 of the character image acquisition program 400 accepts a binary input image having the first resolution, and performs multi-value processing on the input image. And output to the character cutout unit 40.

After the multi-value processing, a multi-value high-resolution character image having the second resolution is generated in the processing of S100 to S108.
In step 402 (S402), the binarizing unit 404 receives the generated multi-value character image, performs a predetermined threshold process on the character image, for example, and converts it into a binary character image. Binary high-resolution character image data with a resolution of 2 is output.

Next, the image processing apparatus 2 according to the fourth to sixth embodiments will be described with reference to FIGS. 16 to 18.
FIG. 16 is a diagram showing a functional configuration of a character image acquisition program 420 executed by the control device 20 of the image processing apparatus 2 according to the fourth embodiment of the present invention. Note that, among the components shown in FIG. 16, the same components as those shown in FIG. 13 are denoted by the same reference numerals.
As shown in FIG. 16, the character image acquisition program 420 is different from the character image acquisition program 400 shown in FIG. 13 in that the multilevel processing is executed after the character extraction processing.

  In the character image acquisition program 420, the multi-value conversion unit 422 receives the binary character image cut out by the character cut-out unit 40, performs multi-value conversion processing on the character image, converts the character image into a multi-value character image, The multi-value character image is output to the similar character extraction unit 42.

FIG. 17 is a diagram showing a functional configuration of a character image acquisition program 440 executed by the control device 20 of the image processing device 2 according to the fifth embodiment of the present invention. Of the components shown in FIG. 17, components substantially the same as those shown in FIG. 13 are denoted by the same reference numerals.
As shown in FIG. 17, the character image acquisition program 440 is different from the character image acquisition program 400 shown in FIG. 13 in that the multi-value processing is executed after the similar character extraction processing.

  In the character image acquisition program 440, the multi-value conversion unit 442 receives similar binary character images extracted by the similar character extraction unit 42, performs multi-value conversion processing on the character image, and generates a multi-value character image. And outputs the similar multi-valued character image to the centroid calculating unit 44 and the high-resolution character image generating unit 46.

FIG. 18 is a diagram showing a functional configuration of a character image acquisition program 460 executed by the control device 20 of the image processing device 2 according to the sixth embodiment of the present invention. 18 that are substantially the same as those shown in FIG. 17 are denoted by the same reference numerals.
As illustrated in FIG. 18, the character image acquisition program 460 outputs a character image similar to the similar character extraction unit 42 to the centroid calculation unit 44 and the multi-value conversion unit 462, and multi-value processing performs a high-resolution character image. It differs from the character image acquisition program 400 shown in FIG. 17 in that it is executed before the generation process.

  In the character image acquisition program 460, the multi-value conversion unit 462 accepts binary character images similar to each other extracted by the similar character extraction unit 42, and multi-values the character image by performing multi-value conversion processing on the character image. The similar multi-value character image is output to the high-resolution character image generation unit 46.

  FIG. 19 is a diagram for describing high-resolution character image data generation processing when the second multi-value quantization method shown in FIG. 14C is applied. In addition, each symbol shown in FIG. 19 shows the substantially same content as the symbol shown in FIG. The second multi-value conversion method is executed with the configuration of the character image acquisition program 460 (FIG. 18), for example.

  As shown in FIG. 19A, the multi-value quantization unit 462 performs an enlargement process for each of the plurality of character image data for each phase. Specifically, the multi-value conversion unit 462 constitutes an enlargement unit that enlarges each extracted character image to a second resolution higher than the first resolution. FIG. 19B shows the sampling grid of the input image and the barycentric positions of a plurality of (4 in this example) character images that have been subjected to the enlargement process. In FIG. 19B, each rectangle indicates a sampling grid of the input image, and a point where broken lines intersect indicates the barycentric position of each character image.

  As shown in FIG. 19C, the high resolution character image generation unit 46 moves the phase of each sampling grid based on the barycentric position of each character image. Further, as shown in FIG. 19D, the high resolution character image generation unit 46 sets a sampling grid (arrow d in the figure) at a second resolution higher than the first resolution. Thereafter, as shown in FIG. 19E, the high-resolution character image generator 46 generates the second resolution based on the value of the multi-value character image data positioned around the sampling grid of the second resolution. The pixel value of the character image at is interpolated. As the interpolation method, nearest neighbor interpolation, linear interpolation, or the like is applied.

FIG. 20 is a diagram for explaining a second embodiment of the high-resolution character image data generation processing by the high-resolution character image generation unit 46.
As shown in FIGS. 20A to 20E, the image processing apparatus 2 applies a predetermined interpolation method so that the phases of the enlarged character images coincide with each other, so that the characters of the first resolution The image is enlarged to a character image having a second resolution higher than the first resolution. Hereinafter, the method of the present embodiment will be described in detail.

  FIG. 20A shows the sampling grid of the input image and the barycentric position of the character image at the first resolution. Here, the rectangle indicated by the arrow a is a sampling grid of the input image, and the point where the broken line indicated by the arrow b intersects is the barycentric position of the character image. FIG. 20B shows a sampling grid having a second resolution higher than the first resolution, and the barycentric position of each character image. As shown in FIGS. 20A and 20B, the multi-value quantization unit 462 performs an enlargement process on the sampling grid of the second resolution based on the center of gravity of each character image at the first resolution. . Here, the high-resolution character image generation unit 46 performs an enlargement process by interpolation so that the phases of the character images coincide with each other. Therefore, the multi-value quantization unit 462 converts each character image to the first resolution based on each extracted character image and the phase of each character image calculated with higher accuracy than one pixel in the first resolution. Enlarging means for enlarging to a higher second resolution is configured.

  FIG. 20C is a diagram schematically showing pixel values of each enlarged character image. FIG. 20D is a diagram illustrating a technique for calculating the pixel value of the character image having the second resolution based on each enlarged character image. As shown in FIGS. 20C and 20D, the high-resolution character image generation unit 46 calculates an average value of pixel values at the same position in each enlarged character image, and calculates the calculated value The pixel value of the position is used. The value calculated in this way is, for example, the pixel value at the position shown in FIG.

  The high-resolution character image generation unit 46 enlarges each character image of the first resolution as it is for each character image, and the phase of each character image matches each character image of the enlarged second resolution. The averaging process may be performed after the movement. The pixel value interpolation method is not particularly limited.

Next, a process executed for characters whose input document includes less than a predetermined number will be described.
FIG. 21 is a diagram for explaining the outline of image processing executed for characters whose document contains less than a predetermined number.
As shown in FIG. 21A, a character may appear only once in a one-page document or a document composed of a plurality of pages, or may appear less than a predetermined number of times. In such a case, the image processing apparatus 2 uses a character group image prepared in advance as shown in FIG. 21B or a character image similar to a character image extracted from an input document from a newly generated character group image. As shown in FIGS. 21 (C) and 21 (D), an image is extracted, and the phase of each extracted character is calculated. Based on the calculated phase and each extracted character image, high-resolution characters are extracted. Generate an image.

  The character group image includes a plurality of character images. Such character group image data is generated by a font drawing device described later. The character group image is generated by determining the font pattern of characters included in the character group image based on the size and thickness of the character extracted from the input document.

  Each character image may be generated with a different pixel phase. Further, a character group image may be further generated when a similar character image is extracted from the input original or the character group image. In this case, the extracted character image is drawn with a different pixel phase in the generated character group image. Further, the image processing apparatus 2 may extract a similar character image from another document image that is different from the designated input image and can be accessed. For example, the image processing apparatus 2 extracts similar character images from a document image group stored in the same folder (or directory) as the input image, a document image group designated collectively by the user, or the like.

FIG. 22 is a diagram showing a functional configuration of a character image acquisition program 480 executed by the control device 20 of the image processing device 2 according to the seventh embodiment of the present invention.
As shown in FIG. 22, the character image acquisition program 480 has a configuration in which a character group image storage unit 482 and a character image number determination unit 484 are added to the character image acquisition program 4 shown in FIG. 22 that are substantially the same as those shown in FIG. 5 are denoted by the same reference numerals.

  In the character image acquisition program 480, the character group image storage unit 482 stores a character group image including a plurality of character images. The character group image includes, for example, 8000 character images (font patterns) (4 types of fonts for each of 2000 types of common Chinese characters). The character group image may further include, for example, 8000 character images (40 fonts for each of 200 alphanumeric symbols). The character group image storage unit 482 is realized by at least one of the memory 204 and the storage device 24.

  The character image number determination unit 484 determines whether the number of similar character images extracted by the similar character extraction unit 42 is less than a predetermined number. For example, the character image number determination unit 484 determines whether the number of similar character images is 0 (whether it is less than 1). When the number of similar character images is less than the predetermined number, the character image number determination unit 484 further adds similar characters from the character group image stored in the character group image storage unit 482 to the character cutout unit 40. Output instructions to cut.

  Therefore, the character cutout unit 40 reads out the character group image stored in the character group image storage unit 482, cuts out the character image included in the character group image, and outputs it to the similar character extraction unit 42. The similar character extraction unit 42 receives the character images cut out from the input image and the character group image by the character cutout unit 40, and extracts similar character images from these images.

FIG. 23 is a flowchart of the high-resolution character image acquisition process (S50) using the stored character group image by the character image acquisition program 480. Of the processes shown in FIG. 23, substantially the same processes as those shown in FIG. 8 are denoted by the same reference numerals.
As shown in FIG. 23, when character image extraction processing and similar character image extraction processing are executed in S100 and S102, in step 500 (S500), the character image number determination unit 484 extracts the extracted similarities. It is determined whether or not the number of character images to be performed is a predetermined number or less. The character image acquisition program 480 proceeds to the process of S502 when the number of similar character images is equal to or less than the predetermined number, and otherwise performs the processes of S104 to S108 to generate a high-resolution character image.

  If the number of extracted similar character images is less than or equal to the predetermined number, in step 502 (S502), the character image number determination unit 484 performs an instruction to further extract similar characters from the stored character group image. Output to the unit 40. The character cutout unit 40 reads the character group image from the character group image storage unit 482, cuts out the character image from the character group image, and extracts it to the similar character extraction unit 42.

  In step 504 (S504), the similar character extraction unit 42 extracts similar character images from the character images cut out from the character group image by the character cutout unit 40. The similar character extraction unit 42 outputs similar character images, which are character images included in the input image and the character group image, to the centroid calculation unit 44 and the high resolution character image generation unit 46. Thereafter, the processing of S104 to S108 is executed to generate a high resolution character image.

FIG. 24 is a diagram showing a functional configuration of a character image acquisition program 500 executed by the control device 20 of the image processing device 2 according to the eighth embodiment of the present invention.
As shown in FIG. 24, the character image acquisition program 500 has a configuration in which a character image number determination unit 502 and a character group image generation unit 504 are added to the character image acquisition program 4 shown in FIG. Note that, among the components shown in FIG. 24, the same reference numerals are given to the substantially same components as those shown in FIG.

  In the character image acquisition program 500, the character image number determination unit 502 determines whether or not the number of similar character images extracted by the similar character extraction unit 42 is less than a predetermined number. When the number of similar character images is less than the predetermined number, the character image number determination unit 502 outputs an instruction to generate a character group image to the character group image generation unit 504.

  When the character group image generation unit 504 receives an instruction to generate a character group image from the character image number determination unit 502, the character group image generation unit 504 receives the size of the character image extracted by the similar character extraction unit 42 from the similar character extraction unit 42, A character group image is generated based on this size and output to the character cutout unit 40. In this manner, the character group image generation unit 504 constitutes a character group image acquisition unit that acquires a character group image including a plurality of character images. The character group image generation process will be described in detail later.

  Therefore, the character cutout unit 40 receives the character group image generated by the character group image generation unit 504, cuts out the character image included in the character group image, and outputs the character image to the similar character extraction unit 42. The similar character extraction unit 42 receives the character images cut out from the input image and the character group image by the character cutout unit 40, and extracts similar character images from these images.

FIG. 25 is a diagram illustrating a detailed functional configuration of the character group image generation unit 504.
As shown in FIG. 25, the character group image generation unit 504 includes a character group image control unit 506, a font code instruction unit 508, a text code instruction unit 510, and a font pattern drawing unit 512. In the character group image generation unit 504, the character group image control unit 506 receives an instruction output from the character image number determination unit 502 and controls other components.

  The font code instruction unit 508 outputs a font code indicating the font type of the character image included in the character group image to the font pattern drawing unit 512. The font code may be determined in advance or may be determined by the control of the character group image control unit 506. The font code is, for example, Mincho or Gothic.

  The text code instruction unit 510 outputs a text code indicating the text type of the character image included in the character group image to the font pattern drawing unit 512. The text code may be determined in advance or may be determined by the control of the character group image control unit 506. The text code is, for example, numbers, alphabets, hiragana, katakana, common kanji.

  The font pattern drawing unit 512 receives the font code specified by the font code specifying unit 508, the text code specified by the text code specifying unit 510, and the size of the character image output by the similar character extracting unit 42, and the character group Under the control of the image control unit 506, a font pattern is drawn based on the received font code, text code, and character image size to generate a character group image, and the character group image is output to the character cutout unit 40. To do.

  Note that the character group image generation unit 504 may be included in a program that operates on hardware different from that of the image processing apparatus 2. In this case, the hardware functions as a font drawing device having the function of the character group image generation unit 504. The image processing device 2 may acquire character group image data by transmitting and receiving character group image data to and from such a font drawing device via a network.

FIG. 26 is a diagram illustrating a character group image generated by the character group image generation unit 504.
As illustrated in FIG. 26, the character group image includes Gothic and Mincho numbers, lowercase and uppercase alphabets, hiragana and common kanji.

FIG. 27 is a flowchart of the high-resolution character image acquisition process (S60) using the generated character group image by the character image acquisition program 500. Of the processes shown in FIG. 27, the same reference numerals are assigned to the processes substantially the same as those shown in FIGS.
As shown in FIG. 27, when the character image clipping process and the similar character image extraction process are executed in S100 and S102, the number of similar character images extracted in S500 is less than or equal to the predetermined number. It is determined whether or not there is.

  If the number of extracted similar character images is less than or equal to the predetermined number, in step 600 (S600), the character image number determination unit 502 outputs an instruction to generate a character group image to the character group image generation unit 504. To do. The character group image generation unit 504 generates a character group image based on the size of the character image extracted by the similar character extraction unit 42 and outputs the character group image to the character cutout unit 40.

  Furthermore, a character image is cut out from the character group image in the process of S502, and a similar character image is extracted in the process of S504. Thereafter, the processing of S104 to S108 is executed to generate a high resolution character image.

FIG. 28 is a diagram illustrating a detailed functional configuration of a character group image generation unit 514 that is a first modification of the character group image generation unit 504.
As shown in FIG. 28, the character group image generation unit 514 includes a character group image control unit 506, a font code instruction unit 508, a text code instruction unit 510, a drawing angle instruction unit 516, and a font pattern drawing unit 518. Of the components shown in FIG. 28, the same components as those shown in FIG. 25 are denoted by the same reference numerals.

  In the character group image generation unit 514, the drawing angle instruction unit 516 outputs an angle for drawing the font pattern to the font pattern drawing unit 518. The drawing angle may be determined in advance or may be determined by control of the character group image control unit 506. The drawing angle is, for example, 0 degrees, 90 degrees, 180 degrees, 270 degrees, or the like.

  The font pattern drawing unit 518 outputs the font code specified by the font code specifying unit 508, the text code specified by the text code specifying unit 510, the drawing angle specified by the drawing angle specifying unit 516, and the similar character extracting unit 42. The character group image is received and the character group image control unit 506 controls to generate a character group image by drawing a font pattern based on the received font code, text code, drawing angle, and character image size. The character group image is output to the character cutout unit 40.

FIG. 29 is a diagram illustrating a character group image generated by the character group image generation unit 514.
As illustrated in FIG. 29, the character group image includes Gothic lowercase letters and uppercase alphabets drawn with drawing angles of 0 degrees, 90 degrees, 180 degrees, and 270 degrees.

FIG. 30 is a diagram illustrating a detailed functional configuration of a character group image generation unit 520 that is a second modification of the character group image generation unit 504.
As shown in FIG. 30, the character group image generation unit 520 has a configuration in which a simulated scan-in processing unit 522 is added to the character group image generation unit 504. 30 that are substantially the same as those shown in FIG. 25 are denoted by the same reference numerals.

  In the character group image generation unit 520, the simulated scan-in processing unit 522 is drawn by the font pattern drawing unit 512 when the input image data is data read (scanned in) by a reading device such as a scanner. A process (simulated scan-in process) simulating a scan-in process such as a spatial filter and noise superimposition is performed on the character group image. The simulated scan-in processing unit 522 outputs the character group image subjected to the simulated scan-in process to the character cutout unit 40. The simulated scan-in processing unit 522 may be added to the character group image generation unit 514.

FIG. 31 is a diagram showing a functional configuration of a character image acquisition program 540 executed by the control device 20 of the image processing device 2 according to the ninth embodiment of the present invention.
As shown in FIG. 31, the character image acquisition program 540 has a configuration in which a character image number determination unit 502 and a corresponding character image generation unit 542 are added to the character image acquisition program 4 shown in FIG. 31 that are substantially the same as those shown in FIGS. 5 and 24 are denoted by the same reference numerals.

  In the character image acquisition program 540, if the character image number determination unit 502 determines that the number of similar character images is less than the predetermined number, the corresponding character image generation unit 542 extracts the characters extracted by the similar character extraction unit 42. The size of the image is received from the similar character extraction unit 42, and a character image corresponding to this size is generated and output to the similar character extraction unit 42.

FIG. 32 is a flowchart of the high-resolution character image acquisition process (S70) using the generated character image by the character image acquisition program 540. Of the processes shown in FIG. 32, substantially the same processes as those shown in FIGS. 8 and 23 are denoted by the same reference numerals.
As shown in FIG. 32, when the character image cut-out process and the similar character image extraction process are executed in S100 and S102, the number of similar character images extracted in S500 is less than or equal to the predetermined number. It is determined whether or not there is.

  When the number of extracted similar character images is less than or equal to the predetermined number, in step 700 (S700), the character image number determination unit 502 outputs an instruction to generate a character image to the corresponding character image generation unit 542. . The corresponding character image generation unit 542 generates a character image corresponding to the size of the character image extracted by the similar character extraction unit 42 and outputs the character image to the similar character extraction unit 42.

  In step 702 (S702), the similar character extraction unit 42 receives the character image generated by the corresponding character image generation unit 542, and further extracts a similar character image from the received character image. Thereafter, the processing of S104 to S108 is executed to generate a high resolution character image.

FIG. 33 is a diagram showing a functional configuration of a character image acquisition program 560 executed by the control device 20 of the image processing device 2 according to the tenth embodiment of the present invention.
As shown in FIG. 33, the character image acquisition program 560 includes a character image number determination unit 502, a generation phase instruction unit 562, and a multi-phase character group image generation unit 564 added to the character image acquisition program 4 shown in FIG. It has the structure made. 33, substantially the same components as those shown in FIGS. 5 and 24 are denoted by the same reference numerals.

In the character image acquisition program 560, the generation phase instruction unit 562 outputs a plurality of phases of the generated character group image to the multi-phase character group image generation unit 564.
When the multi-phase character group image generation unit 564 receives an instruction to generate a character group image from the character image number determination unit 502, the multi-phase character group image generation unit 564 determines the size of the character image extracted by the similar character extraction unit 42 from the similar character extraction unit 42. A plurality of phases are received from the generation phase instruction unit 562, and a plurality of character group images having different phases are generated based on the size and the plurality of phases, and output to the character cutout unit 40. In this way, the multi-phase character group image generation unit 564 acquires a plurality of character group images having different phases.

  Therefore, the character cutout unit 40 receives a plurality of character group images with different phases generated by the character group image generation unit 504, cuts out character images included in the character group image, and outputs them to the similar character extraction unit 42. Output. The similar character extraction unit 42 receives the character images cut out from the input image and the character group image by the character cutout unit 40, and extracts similar character images from these images.

FIG. 34 is a flowchart of the high-resolution character image acquisition process (S80) using the plurality of generated character group images by the character image acquisition program 560. Note that, among the processes shown in FIG. 34, the same reference numerals are given to the substantially same processes as those shown in FIGS.
As shown in FIG. 34, when a character image cut-out process and a similar character image extraction process are executed in S100 and S102, the number of similar character images extracted in S500 is less than or equal to a predetermined number. It is determined whether or not there is.

  If the number of extracted similar character images is equal to or less than the predetermined number, the generation phase instruction unit 562 outputs a plurality of phases to the multi-phase character group image generation unit 564 in step 800 (S800). The multi-phase character group image generation unit 564 generates a plurality of character group images having different phases based on the size of the character image extracted by the similar character extraction unit 42 and the plurality of phases instructed by the generation phase instruction unit 562. And output to the character cutout unit 40.

  Further, the character image is cut out from the plurality of character group images having different phases in the process of S502, and similar character images are extracted in the process of S504. Thereafter, the processing of S104 to S108 is executed to generate a high resolution character image.

FIG. 35 is a diagram showing a functional configuration of a character image acquisition program 580 executed by the control device 20 of the image processing device 2 according to the eleventh embodiment of the present invention.
As shown in FIG. 35, the character image acquisition program 480 has a configuration in which a character image replacement unit 582 is added to the character image acquisition program 480 shown in FIG. Note that, among the components shown in FIG. 35, those substantially the same as those shown in FIGS. 5 and 22 are denoted by the same reference numerals.

  In the character image acquisition program 580, when the number of similar character images is less than the predetermined number, the character image replacement unit 582 extracts the character images from the character group images stored in the character group image storage unit 482. The character image is replaced with a character image similar to the character image. The character image replacement unit 582 outputs the replaced character image to the high resolution character image generation unit 46.

FIG. 36 is a flowchart of the high-resolution character image acquisition process (S90) accompanied by the character image replacement process by the character image acquisition program 580. Of the processes shown in FIG. 36, the same reference numerals are assigned to the processes substantially the same as those shown in FIGS.
As shown in FIG. 36, when a character image cut-out process and a similar character image extraction process are executed in S100 and S102, the number of similar character images extracted in S500 is less than or equal to a predetermined number. It is determined whether or not there is.

  If the number of extracted similar character images is less than or equal to the predetermined number, in step 900 (S900), the character image number determination unit 502 outputs an instruction to replace the character image to the character image replacement unit 582. The character image replacement unit replaces a character image whose number of characters is equal to or less than a predetermined number with a character image similar to the character image and extracted from the character group image. Output.

  Thereafter, the processing of S104 to S108 is executed to generate a high resolution character image. In this example, the character group image is generated and stored in advance, but may be generated each time.

Next, a process executed for the same character with different character sizes will be described.
FIG. 37 is a diagram exemplifying a document including characters that are the same but different in size.
As illustrated in FIG. 37, the character image “Den” is included in the document in different sizes. For example, the character image in the text indicated by the arrow B is smaller than the character image of the title indicated by the arrow A and larger than the character image of the footnote indicated by the arrow C, for example. Here, a plurality of character images “Den” indicated by arrows B and C are included in the document, while only one character image “Den” indicated by arrow A is included in the document. Similarly, for example, the character image “machine” is used twice in the text, while it is used once in the title with larger characters in the text.

  As described above, the document may include characters having different sizes even if they are the same character. The character size is also referred to as the number of character points.

FIG. 38 is a diagram for explaining the outline of the image processing executed for characters having the same character but different sizes.
As shown in FIG. 38A, images such as characters and graphics may appear in different sizes on a one-page document or a document composed of a plurality of pages. In such a case, the image processing apparatus 2 adjusts the size of an image such as characters cut out from the input document image as shown in FIG. 38B, and as shown in FIG. The phase of each character whose size has been adjusted is calculated, and a high-resolution character image is generated based on the calculated phase and each character image whose size has been adjusted. Preferably, the image processing device 2 aligns the extracted character size to a predetermined size.

FIG. 39 is a diagram showing a functional configuration of a character image acquisition program 600 executed by the control device 20 of the image processing device 2 according to the twelfth embodiment of the present invention.
As shown in FIG. 39, the character image acquisition program 600 has a configuration in which a size adjustment unit 602 is added to the character image acquisition program 4 shown in FIG. Note that, among the components shown in FIG. 39, the same components as those shown in FIG.

  In the character image acquisition program 600, the size adjustment unit 602 receives the character image cut out by the character cutout unit 40, adjusts the size of the character image, and outputs it to the similar character extraction unit 42. More specifically, the size adjustment unit 602 aligns the character image to a predetermined size.

For example, the size adjustment unit 602 selects any one of the plurality of cut out character images, and adjusts the size of the character image so that the size of the plurality of character images becomes the size of the selected character image. To do.
FIG. 40 is a diagram showing an outline of image processing executed so that the size of the cut-out character image is aligned with the size of the selected character image, and FIG. A document image is illustrated, FIG. 40B illustrates a character image whose size has been adjusted, and FIG. 40C illustrates a generated high-resolution image.

  As shown in FIGS. 40 (A) and 40 (B), the image processing apparatus 2 according to the present embodiment selects one of a plurality of extracted character images, and selects a character image larger than the selected character image. The size of the selected character image is reduced, and a character image smaller than the selected character image is enlarged to the size of the selected character image. As shown in FIG. 40C, the image processing device 2 generates a high-resolution character image based on a plurality of character images aligned with the size of the selected character image. For example, the image processing apparatus 2 calculates the phase of each character image, superimposes each character image based on the calculated phase, and calculates the average value of each character image. In this case, the image processing apparatus 2 estimates the phase of the character image to be sampled based on at least one of the center of gravity of the character image and the position where the cross-correlation coefficient is the largest.

  The image processing apparatus 2 may replace the selected character image with the generated character image. In this case, the character image acquisition program 600 (FIG. 39) includes functions equivalent to the position calculation unit 50 and the character arrangement unit 52 of the image enlargement program 5 (FIG. 9).

  In the character image acquisition program 600, the size adjustment unit 602 selects any one of the plurality of cut out character images, calculates the size when the selected character image is enlarged at a predetermined magnification, The size of the character image may be adjusted so that the sizes of the character images are equal to the calculated size.

FIG. 41 is a diagram showing a flowchart of high-resolution character image acquisition processing (S1000) accompanied by character size adjustment processing by the character image acquisition program 600. Of the processes shown in FIG. 41, the same reference numerals are assigned to the processes substantially the same as those shown in FIG.
As shown in FIG. 41, when the character image clipping process is executed in S100, in step 1002 (S1002), the size adjusting unit 602 selects one of the clipped character images, Converts a character image of a different size into the size of the character image. Thereafter, the processing of S102 to S108 is executed to generate a high-resolution character image.

FIG. 42 is a diagram for explaining the outline of the first modification of the image processing executed for characters having the same character but different sizes.
As shown in FIG. 42A, an image such as a character may appear in a document image with a different size. In such a case, as shown in FIG. 42B, the image processing apparatus 2 adjusts the size of the character image so that the size of the image such as the extracted character is aligned with a predetermined size. As shown in 42 (C), similar character images are extracted from the character images having the same size.

  Further, as shown in FIG. 42D, the image processing apparatus 2 selects the largest character image in the document image from the extracted character images, and the size of the extracted character image is the largest. The size of the character image is adjusted to match the size of the character image. As shown in FIG. 42E, the image processing apparatus 2 generates a high-resolution character image based on a plurality of character images having the same size.

  The image processing apparatus 2 may replace the character image included in the document image with the generated character image. In this case, the image processing device 2 replaces the character image after reducing the generated character image based on the size of the character image included in the document image. For example, the image processing apparatus 2 replaces the largest character image in the document image with the generated character image, and replaces the character image other than the largest character image with the character image obtained by reducing the generated character image.

FIG. 43 is a diagram showing a functional configuration of a character image acquisition program 610 executed by the control device 20 of the image processing device 2 according to the thirteenth embodiment of the present invention.
As shown in FIG. 43, the character image acquisition program 610 has a configuration in which a second size adjustment unit 612 is added to the character image acquisition program 600 shown in FIG. 43, substantially the same components as those shown in FIG. 39 are denoted by the same reference numerals.

  In the character image acquisition program 610, the second size adjustment unit 612 searches for the largest character image in the original document image from the plurality of character images received from the similar character extraction unit 42. The second size adjustment unit 612 adjusts the size of the character image so that the plurality of character images extracted by the similar character extraction unit 42 are aligned with the size of the largest character image. The second size adjustment unit 612 outputs the character images having the same size to the centroid calculation unit 44 and the high resolution character image generation unit 46.

  The image processing apparatus 2 may replace each of the cut character images with the generated character image. In this case, the character image acquisition program 610 includes functions equivalent to the position calculation unit 50 and the character arrangement unit 52 of the image enlargement program 5 (FIG. 9). For example, the image processing apparatus 2 replaces the largest character image among similar character images in the document image with the generated character image, and replaces the generated character image with the character image other than the largest character image. The character image generated by reducing the size based on the size of the image is rewritten.

  FIG. 44 is a flowchart of a high-resolution character image acquisition process (S1010) involving two character size adjustment processes by the character image acquisition program 610. Note that among the processes shown in FIG. 44, the same reference numerals are assigned to the processes substantially the same as those shown in FIG.

As shown in FIG. 44, in the processes of S100 to S102, a character image is cut out from the input image, the size of the character image is adjusted, and a similar character image is extracted.
In step 1012 (S1012), the second size adjustment unit 612 aligns similar character images to the size of the largest character image in the input image among these character images. Thereafter, the processing of S104 to S108 is executed to generate a high resolution character image.

FIG. 45 is a diagram for explaining an outline of a second modification of the image processing executed for characters having the same character but different sizes.
The second modified example is different from the first modified example in that the size of the character image is adjusted for a character image that is close to a certain size among the cut character images. For example, when the difference in the size of the character images is less than a predetermined value, the image processing apparatus 2 determines that these character images are character images whose sizes are to be adjusted.

  As shown in FIG. 45A, when a character image appears in a document image with a different size, the image processing apparatus 2 has a certain size among the cut out character images as shown in FIG. The size of the character image is adjusted so that the size of the near character image (the character image surrounded by the broken line) is equal to a predetermined size, and as shown in FIG. From the above, similar character images are extracted.

  Further, as shown in FIG. 45D, the image processing apparatus 2 adjusts the size of the extracted character image so that the size of the extracted character image is aligned with a predetermined size, and FIG. As shown in FIG. 4, a high-resolution character image is generated based on a plurality of character images having the same size. Further, as shown in FIGS. 45E and 45F, the image processing apparatus 2 adds the generated character image to the extracted character image group, and recursively repeats the procedure shown in FIG. . In this way, the image processing apparatus 2 generates a high resolution character image.

FIG. 46 is a diagram for explaining the outline of the third modification of the image processing executed for characters having the same character but different sizes.
As shown in FIG. 46A, when the character image appears in the original image with a different size, as shown in FIG. 46B, the image processing apparatus 2 has a size of the image such as the cut out character. The size of the character image is adjusted so as to be aligned with a predetermined size.

  Further, as shown in FIG. 46C, the image processing apparatus 2 weights the character image whose size has been adjusted based on the size of the character image in the original document image. Here, W1, W2 and W3 in the figure represent weights. The image processing device 2 superimposes each character image based on the phase of each character image, and generates a high resolution character image based on the calculated weight.

FIG. 47 is a diagram showing a functional configuration of a character image acquisition program 620 executed by the control device 20 of the image processing device 2 according to the fourteenth embodiment of the present invention.
As shown in FIG. 47, in the character image acquisition program 620, the character image acquisition program 600 shown in FIG. 39 is loaded with the weight calculation unit 622, and the high resolution character image generation unit 46 is added to the high resolution character image generation unit 624. It has a substituted configuration. 47, substantially the same components as those shown in FIG. 39 are denoted by the same reference numerals.

  In the character image acquisition program 620, the weight calculation unit 622 calculates the weight of each character image based on the size of the character image cut out by the character cutout unit 40. For example, the weight calculation unit 622 increases the weight as the character image increases in the input document image. The high-resolution character image generation unit 624 calculates the phase of the character image based on the relative position between the centroid of each character image calculated by the centroid calculation unit 44 and the sampling grid point, and the character image based on the calculated phase. Are superimposed, and the weighted average value of the character image is calculated.

FIG. 48 is a flowchart of a high-resolution character image acquisition process (S1020) for calculating a weighted average value by the character image acquisition program 620. Of the processes shown in FIG. 48, substantially the same processes as those shown in FIG. 41 are denoted by the same reference numerals.
As shown in FIG. 48, in the processes of S100 to S104, a character image is cut out from the input image, the size of the character image is adjusted, a similar character image is extracted, and the center of gravity of the extracted character image is calculated. The

In step 1022 (S1022), the weight calculation unit 622 calculates a weight based on the size of the extracted character image.
In step 1024 (S1024), the high resolution character image generation unit 624 calculates a weighted average value of the character image based on the calculated weight.
Thereafter, a high-resolution character image is output in S108.

FIG. 49 is a diagram for explaining the outline of the fourth modification of the image processing executed for the same character but different sizes.
As shown in FIG. 49A, when a character image appears in a document image with a different size, as shown in FIG. 49B, the image processing apparatus 2 has a size of an image such as a cut out character. The size of the character image is adjusted so as to be aligned with a predetermined size. Here, when reducing the character image, the image processing apparatus 2 generates a plurality of reduced images having different phases. Then, as shown in FIG. 49C, the image processing apparatus 2 generates a high-resolution character image based on the plurality of character images whose sizes have been adjusted. Note that such image processing is executed by a program having a configuration equivalent to the character image acquisition program 600 shown in FIG.

FIG. 50 is a diagram illustrating a flowchart of high-resolution character image acquisition processing (S1030) involving generation processing of reduced images having different phases. Of the processes shown in FIG. 50, the same reference numerals are assigned to substantially the same processes as those shown in FIG.
As shown in FIG. 50, in the processes of S100 and S1002, a character image is cut out from the input image, and the size of the character image is adjusted.

  When the character image is reduced, in step 1032 (S1032), a plurality of character images obtained by reducing the character image at a plurality of different phases are generated. Thereafter, the processing of S102 to S108 is executed to generate a high-resolution character image.

Next, an image encoding process and a decoding process performed using the generated high-resolution character image will be described.
In the JBIG2 text region (Text Region), partial images such as characters and graphics are extracted from binary image data (hereinafter also referred to as input partial images), and the partial images already registered in the dictionary Processing for comparison with data (hereinafter also referred to as a registered partial image) is executed. As a result of the comparison, if a registered partial image similar to the input partial image is not found, the input partial image is registered in the dictionary. When a registered partial image similar to the input partial image is found, the input partial image is encoded based on the index indicating the partial image in the dictionary and the position of the input partial image in the binary image data.

  When the same plurality of partial images are read by a scanner or the like and encoded in the JBIG2 text area, the phases at the time of reading may be shifted for the plurality of partial images. In such a case, it is necessary to consider the compression rate in the encoding process and the image quality of the output image when decoded.

FIG. 51 is a diagram for explaining the encoding process and the decoding process in the JBIG2 text region. FIG. 51A shows a method for encoding each of a plurality of partial images extracted from an input image. FIG. 51B shows a method of encoding one partial image representing a plurality of partial images extracted from the input image.
In the method shown in FIG. 51A, when a plurality of partial images are out of phase, the plurality of partial images are encoded as different partial image data. In this case, since the original partial image data is decoded as it is, there is no deterioration in image quality in the restored output image. However, the compression rate in the encoding process is low, and the amount of code data is large.

  In the method shown in FIG. 51B, even when the phases of a plurality of partial images are shifted, any one of the plurality of partial images is encoded as a representative image. Therefore, in such a method, the compression rate is high and the amount of code data is small. However, in the output image to be decoded, each of the partial images is replaced with a partial image having a different phase, so that the image quality may be deteriorated. For example, the deterioration of the image quality appears as character string disturbance.

  The image encoding device according to the embodiment of the present invention encodes a high-resolution partial image generated as described above as a representative image. Hereinafter, encoding processing and decoding processing will be described using a character image as an example of a partial image.

  FIG. 52 is a diagram showing a functional configuration of an image encoding program 70 executed by the control device 20 of the image encoding device 7 according to the embodiment of the present invention. 52, substantially the same components as those shown in FIG. 5 are denoted by the same reference numerals. The image encoding device 7 has a hardware configuration equivalent to that of the image processing device 2 shown in FIG.

  As shown in FIG. 52, the image encoding program 70 is the same as the image encoding program 70 shown in FIG. 5 except that the character extracting unit 40, the similar character extracting unit 42, the centroid calculating unit 44, the high resolution character image generating unit 46, and the like. An encoding unit 700 is included. In the image encoding program 70, the high resolution character image generated by the high resolution character image generation unit 46 is also referred to as a representative image. Therefore, the high-resolution image generation unit 46 extracts each partial image (character image in this example) extracted by the similar character extraction unit 42 and each partial image calculated with higher accuracy than one pixel at the first resolution. The representative image generating means for generating the representative image of the second resolution is configured based on the phase of the second image.

  The encoding unit 700 generates code data by encoding the representative image generated by the high resolution character image generation unit 46 and the phase and position of each partial image extracted by the similar character extraction unit 42. More specifically, the encoding unit 700 determines the position of the character image input from the character cutout unit 40 in the input image, the phase of the character image estimated by the centroid calculation unit 44, and the identifier ( The representative image data generated by the high resolution character image generation unit 46 with the index) is encoded for each character image data cut out by the character cutout unit 40.

  The encoding unit 700 may store the encoded data in the storage device 24 or may transmit the encoded data to an external computer such as the image decoding device 8 described later via the communication device 22. Also, the code data may be stored in the recording medium 240 and supplied to the image decoding device 8 or the like.

53A and 53B are diagrams for describing input image encoding processing by the encoding unit 700. FIG. 53A illustrates an input image, and FIG. 53B is generated based on the input image. A representative image dictionary is illustrated, and FIG. 53C illustrates information (position and phase) related to a partial image cut out from the input image.
As illustrated in FIGS. 53A and 53B, each of the partial images (in this example, the character image “A”) included in the input image is extracted by the image encoding program 70, and the representative image is Generated. The representative image is registered in the representative image dictionary, and an index is assigned to the representative image. The shape and index of the representative image are encoded by the encoding unit 700 as information about the representative image included in the representative image dictionary.

  As illustrated in FIG. 53C, the position of the partial image in the input image, the phase when the image is cut out, and the index of the corresponding representative image are encoded as information about the partial image cut out by the character cutout unit 40. The encoding unit 700 encodes the data. As described above, the encoding unit 700 generates code data by encoding the representative image dictionary and information about the partial image cut out from the input image.

FIG. 54 is a diagram showing a flowchart of the image encoding process (S1040) by the image encoding program 70. Of the processes shown in FIG. 54, substantially the same processes as those shown in FIG. 8 are denoted by the same reference numerals.
As shown in FIG. 54, in the processing of S100 to S106, a character image is cut out from the input image, a similar character image is extracted, the center of gravity (phase) is calculated, and a high-resolution character image (representative image) is generated. The

In step 1042 (S1042), the encoding unit 700 encodes information (position, phase, and index) related to the extracted character image.
In step 1044 (S1044), the encoding unit 700 encodes the representative image that has been generated and assigned an index. Specifically, the encoding unit 700 encodes the representative image dictionary.
In step 1046 (S1046), the encoding unit 700 generates and outputs code data.

Next, the image decoding device 8 according to the first embodiment of the present invention will be described.
FIG. 55 is a diagram illustrating a functional configuration of an image decoding program 80 executed by the control device 20 of the image decoding device 8.
As shown in FIG. 55, the image decoding program 80 includes a decoding unit 800, a representative image storage unit 802, a low resolution character image generation unit 804, and a character image arrangement unit 806. Hereinafter, the image decoding program 80 will be described using a character image as an example of a partial image.

  In the image decoding program 80, the decoding unit 800 receives code data generated by the image encoding device 7 and decodes the code data. More specifically, the decoding unit 800 decodes a representative image dictionary including a representative image to which an index is assigned and information (position, phase, index) regarding the partial image based on the code data. The decoding unit 800 outputs the representative image data and the index assigned to the representative image to the representative image storage unit 802. The decoding unit 800 outputs the phase of the character image to the low resolution character image generation unit 804 and outputs the position of the character image to the character image arrangement unit 806.

  The representative image storage unit 802 stores the decoded representative image dictionary. The representative image storage unit 802 is realized by at least one of the memory 204 and the storage device 24.

  The low resolution character image generation unit 804 generates a partial image of the first resolution based on the representative image stored in the representative image storage unit 802 and the decoded phase. Specifically, the low-resolution character image generation unit 804 reads the stored representative image, reduces the representative image based on the phase, and generates low-resolution character image data having the first resolution. In this way, the low-resolution character image generation unit 804 constitutes a partial image generation unit.

  Based on the position of the partial image decoded by the decoding unit 800, the character image arrangement unit 806 arranges the low resolution character image generated by the low resolution character image generation unit 804 in the output image and generates output image data. Thus, the character image arrangement unit 806 constitutes a partial image arrangement unit.

FIG. 56 is a diagram showing a flowchart of image decoding processing (S1050) by the image decoding program 80.
As shown in FIG. 56, in step 1052 (S1052), the decoding unit 800 of the image decoding program 80 decodes the encoded data, and the phase of the representative image of the second resolution and the partial image of the first resolution. And position.
In step 1054 (S1054), the decoding unit 800 stores the representative image data and the index assigned to the representative image in the representative image storage unit 802, and generates a representative image dictionary.

In step 1056 (S1056), the low resolution character image generation unit 804 generates a low resolution character image having a first resolution based on the representative image in the representative image dictionary and the phase of the decoded partial image.
In step 1058 (S1058), the character image arrangement unit 806 arranges the generated low-resolution character image in the output image based on the decoded position.

In step 1060 (S1060), the character image placement unit 806 determines whether or not placement of all partial images on the output image has been completed. The image decoding program 80 proceeds to the process of S1062 when the arrangement process is completed for all the partial images, and returns to the process of S1056 otherwise.
In step 1062 (S1062), the character image arrangement unit 806 generates and outputs output image data.

Next, an image decoding device 8 according to the second embodiment of the present invention will be described.
FIG. 57 is a diagram illustrating a functional configuration of the image decoding program 82 executed by the control device 20 of the image decoding device 8.
As shown in FIG. 57, in the image decoding program 80, an enlargement control unit 820 and a position conversion unit 824 are added to the image decoding program 80 (FIG. 55), and the low resolution character image generation unit 804 generates a low resolution character image. The configuration is replaced with the portion 822.

  In the image decoding program 82, the enlargement control unit 820 controls the low-resolution character image generation unit 822 and the position conversion unit 824 based on the scaling factor. More specifically, the enlargement control unit 820 reads a scaling factor stored in advance in at least one of the memory 204 and the storage device 24, and outputs it to the low resolution character image generation unit 822 and the position conversion unit 824. To do. The scaling factor control unit 820 may output the scaling factor set via the UI device 26 and the communication device 22 to the low resolution character image generation unit 822 and the like.

  The low resolution character image generation unit 822 generates a partial image having the first resolution based further on the scaling factor input from the enlargement control unit 820. Specifically, the low resolution character image generation unit 822 changes the magnification for generating the low resolution character image data based on the scaling factor.

  The position conversion unit 824 converts the position of the partial image decoded by the decoding unit 800 based on the scaling factor input from the enlargement control unit 820, and outputs the converted position to the character image arrangement unit 806. . Therefore, the character image arrangement unit 806 arranges the low-resolution character image in the output image based on the position converted by the position conversion unit 824.

FIG. 58 is a flowchart of the image decoding process (S1070) by the image decoding program 82. Of the processes shown in FIG. 58, those substantially the same as the processes shown in FIG. 56 are denoted by the same reference numerals.
As shown in FIG. 58, in S1052 and S1054, the code data is decoded and a representative image dictionary is generated.

In step 1072 (S1072), the low resolution character image generation unit 822 generates a low resolution character image of the first resolution based on the representative image of the representative image dictionary, the phase of the decoded partial image, and the scaling factor. Generate.
In step 1074 (S1074), the position conversion unit 824 converts the position of the partial image based on the scaling factor.

  Also, in the process of S1058, the character image arrangement unit 806 arranges the generated low-resolution character image in the output image based on the converted position. Further, it is determined in step S1060 whether or not the arrangement processing of all the partial images has been completed, and output image data is generated in step S1062.

It is a figure explaining the quality of the character image at the time of enlarging a character image. It is a figure explaining the method of producing | generating the image of a resolution higher than imaging resolution. It is a figure explaining the outline of the image processing by the image processing apparatus which concerns on this invention. 2 is a diagram illustrating a hardware configuration of an image processing apparatus 2 and an image enlargement apparatus 3 according to an embodiment of the present invention, centering on a control apparatus 20. FIG. It is a figure which shows the function structure of the character image acquisition program 4 performed by the control apparatus 20 of the image processing apparatus 2 which concerns on the 1st Embodiment of this invention. FIG. 6A illustrates an input image, and FIG. 6B illustrates a character cutout unit. FIG. 6A illustrates a character cutout process performed by the character cutout unit 40 and a similar character extraction process performed by the similar character extraction unit 42. The character image cut out from the input image by 40 is illustrated, and FIG. 6C illustrates the similar character image extracted by the similar character extraction unit 42. It is a figure explaining the production | generation process of the high resolution character image data by the high resolution character image generation part. It is a figure which shows the flowchart of the high resolution character image acquisition process (S10) by the character image acquisition program 4. FIG. It is a figure which shows the function structure of the image expansion program 5 performed by the control apparatus 20 of the image expansion apparatus 3 which concerns on embodiment of this invention. It is a figure which shows the flowchart of the image expansion process (S20) by the image expansion program 5. FIG. It is a figure which shows the function structure of the character image acquisition program 6 performed by the control apparatus 20 of the image processing apparatus 2 which concerns on the 2nd Embodiment of this invention. It is a figure which shows the flowchart of the high resolution character image acquisition process (S30) including the noise removal process by the character image acquisition program 6. FIG. It is a figure which shows the function structure of the character image acquisition program 400 performed by the control apparatus 20 of the image processing apparatus 2 which concerns on the 3rd Embodiment of this invention. It is a figure which shows the outline | summary of the multi-value quantization process performed by the image processing apparatus 2 which concerns on embodiment of this invention. It is a figure which shows the flowchart of the high resolution character image acquisition process (S40) by the character image acquisition program 400. FIG. It is a figure which shows the function structure of the character image acquisition program 420 performed by the control apparatus 20 of the image processing apparatus 2 which concerns on the 4th Embodiment of this invention. It is a figure which shows the function structure of the character image acquisition program 440 performed by the control apparatus 20 of the image processing apparatus 2 which concerns on the 5th Embodiment of this invention. It is a figure which shows the function structure of the character image acquisition program 460 performed by the control apparatus 20 of the image processing apparatus 2 which concerns on the 6th Embodiment of this invention. It is a figure explaining the production | generation process of the high resolution character image data by the high resolution character image generation part 46 in case the 2nd multi-value-izing method is applied. It is a figure explaining the 2nd Example of the production | generation process of the high resolution character image data by the high resolution character image generation part. It is a figure explaining the outline of the image processing performed with respect to the character in which only less than the predetermined number is contained in the document. It is a figure which shows the function structure of the character image acquisition program 480 performed by the control apparatus 20 of the image processing apparatus 2 which concerns on the 7th Embodiment of this invention. It is a figure which shows the flowchart of the high resolution character image acquisition process (S50) using the character group image memorize | stored by the character image acquisition program 480. FIG. It is a figure which shows the function structure of the character image acquisition program 500 performed by the control apparatus 20 of the image processing apparatus 2 which concerns on the 8th Embodiment of this invention. It is a figure which shows the detailed functional structure of the character group image generation part. It is a figure which illustrates the character group image produced | generated by the character group image production | generation part 504. It is a figure which shows the flowchart of the high resolution character image acquisition process (S60) using the produced | generated character group image by the character image acquisition program 500. FIG. It is a figure which shows the detailed functional structure of the character group image generation part 514 which is the 1st modification of the character group image generation part 504. FIG. It is a figure which illustrates the character group image produced | generated by the character group image production | generation part 514. FIG. It is a figure which shows the detailed functional structure of the character group image generation part 520 which is the 2nd modification of the character group image generation part 504. FIG. It is a figure which shows the function structure of the character image acquisition program 540 performed by the control apparatus 20 of the image processing apparatus 2 which concerns on the 9th Embodiment of this invention. It is a figure which shows the flowchart of the high resolution character image acquisition process (S70) using the produced | generated character image by the character image acquisition program 540. FIG. It is a figure which shows the function structure of the character image acquisition program 560 performed by the control apparatus 20 of the image processing apparatus 2 which concerns on the 10th Embodiment of this invention. It is a figure which shows the flowchart of the high resolution character image acquisition process (S80) using the produced | generated several character group image by the character image acquisition program 560. FIG. It is a figure which shows the function structure of the character image acquisition program 580 performed by the control apparatus 20 of the image processing apparatus 2 which concerns on the 11th Embodiment of this invention. It is a figure which shows the flowchart of the high resolution character image acquisition process (S90) accompanying the replacement process of a character image by the character image acquisition program 580. FIG. FIG. 4 is a diagram illustrating an example of a manuscript including characters that are the same but different in size. It is a figure explaining the outline of the image processing performed with respect to the character from which a magnitude | size differs even if it is the same character. It is a figure which shows the function structure of the character image acquisition program 600 performed by the control apparatus 20 of the image processing apparatus 2 which concerns on the 12th Embodiment of this invention. FIG. 40A is a diagram illustrating an outline of image processing executed so that the size of a cut-out character image is aligned with the size of a selected character image, and FIG. 40A illustrates an input document image. FIG. 40B illustrates a character image whose size has been adjusted, and FIG. 40C illustrates a generated high-resolution image. It is a figure which shows the flowchart of the high resolution character image acquisition process (S1000) accompanying a character size adjustment process by the character image acquisition program 600. FIG. It is a figure explaining the outline of the 1st modification of the image processing performed with respect to the character from which a size differs even if it is the same character. It is a figure which shows the function structure of the character image acquisition program 610 performed by the control apparatus 20 of the image processing apparatus 2 which concerns on the 13th Embodiment of this invention. It is a figure which shows the flowchart of the high resolution character image acquisition process (S1010) accompanied by the character size adjustment process of 2 times by the character image acquisition program 610. FIG. It is a figure explaining the outline of the 2nd modification of the image processing performed with respect to the character from which a size differs even if it is the same character. It is a figure explaining the outline of the 3rd modification of the image processing performed with respect to the character from which a magnitude | size differs even if it is the same character. It is a figure which shows the function structure of the character image acquisition program 620 performed by the control apparatus 20 of the image processing apparatus 2 which concerns on the 14th Embodiment of this invention. It is a figure which shows the flowchart of the high resolution character image acquisition process (S1020) which calculates the weighted average value by the character image acquisition program 620. It is a figure explaining the outline of the 4th modification of the image processing performed with respect to the character from which a magnitude | size differs even if it is the same character. It is a figure which shows the flowchart of the high resolution character image acquisition process (S1030) accompanied by the production | generation process of the reduced image from which a phase differs. FIG. 51A is a diagram illustrating an encoding process and a decoding process in a JBIG2 text region, and FIG. 51A illustrates a method of encoding each of a plurality of partial images extracted from an input image, and FIG. ) Shows a method of encoding one partial image representing a plurality of partial images extracted from the input image. It is a figure which shows the function structure of the image coding program 70 performed by the control apparatus 20 of the image coding apparatus 7 which concerns on embodiment of this invention. FIGS. 53A and 53B illustrate an input image encoding process performed by the encoding unit 700. FIG. 53A illustrates an input image, and FIG. 53B illustrates a representative image dictionary generated based on the input image. For example, FIG. 53C illustrates information (position and phase) regarding a partial image cut out from the input image. It is a figure which shows the flowchart of the image encoding process (S1040) by the image encoding program 70. FIG. It is a figure which shows the function structure of the image decoding program 80 performed by the control apparatus 20 of the image decoding apparatus 8. FIG. It is a figure which shows the flowchart of the image decoding process (S1050) by the image decoding program 80. FIG. It is a figure which shows the function structure of the image decoding program 82 performed by the control apparatus 20 of the image decoding apparatus 8. FIG. It is a figure which shows the flowchart of the image decoding process (S1070) by the image decoding program 82. FIG.

Explanation of symbols

2 Image processing device 3 Image enlargement device 4 Character image acquisition program 5 Image enlargement program 6 Character image acquisition program 7 Image encoding device 8 Image decoding device 20 Control device 22 Communication device 24 Storage device 26 UI device 40 Character extraction unit 42 Similar Character extraction unit 44 Center of gravity calculation unit 46 High resolution character image generation unit 50 Position calculation unit 52 Character placement unit 60 Phase calculation unit 62 Representative character extraction unit 64 Phase representative value calculation unit 66 Noise removal unit 68 High resolution character image generation unit 70 Image Encoding program 80 Image decoding program 82 Image decoding program 202 CPU
204 Memory 240 Recording medium 400 Character image acquisition program 402 Multi-value conversion unit 404 Binarization unit 420 Character image acquisition program 422 Multi-value conversion unit 440 Character image acquisition program 442 Multi-value conversion unit 460 Character image acquisition program 462 Multi-value conversion unit 480 Character image acquisition program 482 Character group image storage unit 484 Character image number determination unit 500 Character image acquisition program 502 Character image number determination unit 504 Character group image generation unit 506 Character group image control unit 508 Font code instruction unit 510 Text code instruction unit 512 Font pattern drawing unit 514 Character group image generation unit 516 Drawing angle instruction unit 518 Font pattern drawing unit 520 Character group image generation unit 522 Simulated scan-in processing unit 540 Character image acquisition program 542 Corresponding character image generation unit 560 Character image acquisition program Grams 562 generation phase instruction portion 564 plural phase character group image generator 580 character image acquisition program 582 character image replacing unit 700 coding unit 800 decoding unit 802 representative-image storage unit 804 the low resolution character generator 806 character image arrangement unit

Claims (19)

Cutting means for cutting out a character image included in a scanned image obtained by scanning a document;
Wherein a first character image specified from among the plurality of character images extracted by the cutout section, of the first vector data representing the character image distance second Ru der than the predetermined value and character image from the cutout section by cut out said plurality of character image extraction means for extracting at a first resolution,
A character image extracted first character image and the second by the extraction means, the first of said first calculated at a higher than one pixel accuracy in the resolution of the character image and the second character image And generating means for generating a character image having a second resolution higher than the first resolution using the phase of the image processing apparatus. A character group image storage means for storing a character group image including a plurality of character images;
The generating means, when said extracted by said extraction means second character image is not less than the predetermined number, said first character image and the second character images extracted by the extraction means, said first using a 1 the calculated higher than one pixel accuracy in the resolution of the first character image and the second character image phase, of the second resolution is a higher resolution than the first resolution to generate a character image, when the extracted by the extracting means second character image is less than the predetermined number, there the character images extracted from the character group image stored in the character group image storage means the image processing apparatus according to claim 1 for generating a further character images having the second resolution by using a distance predetermined value less than der Rubun shaped image with vector data representing the first character image Te . A character group image storage means for storing a character group image including a plurality of character images;
The generating means, when said extracted by said extraction means second character image is not less than the predetermined number, said first character image and the second character images extracted by the extraction means, said first using a 1 the calculated higher than one pixel accuracy in the resolution of the first character image and the second character image phase, of the second resolution is a higher resolution than the first resolution to generate a character image, when the extracted by the extracting means second character image is less than the predetermined number, the first character image from the character group image stored in the character group image storage means the image processing apparatus according to claim 1, the distance between the vector data representing the first character image a extracted character image is replaced by a predetermined value less than der Rubun shaped image. Based on the size of the extracted first character image by the extraction unit further includes a character group image acquisition means for acquiring a character group image including a plurality of character images,
The generating means, when said extracted by said extraction means second character image is not less than the predetermined number, said first character image and the second character images extracted by the extraction means, said first using a 1 the calculated higher than one pixel accuracy in the resolution of the first character image and the second character image phase, of the second resolution is a higher resolution than the first resolution to generate a character image, when the extracted by the extracting means second character image is less than the predetermined number, a character image extracted from the acquired character group image by the character group image acquisition unit the image processing apparatus according to claim 1 for generating a further character images having the second resolution by using a distance predetermined value less than der Rubun shaped image with vector data representing the first character image. The image processing apparatus according to claim 4, wherein the character group image acquisition unit acquires a plurality of character group images having different phases. Further comprising a corresponding character image generating means for generating a character image corresponding to the size of the first character image extracted by said extraction means,
The generating means, when said extracted by said extraction means second character image is not less than the predetermined number, said first character image and the second character images extracted by the extraction means, said first using a 1 the calculated higher than one pixel accuracy in the resolution of the first character image and the second character image phase, of the second resolution is a higher resolution than the first resolution generating a character image, said extracted by the extraction means second character image is less than the predetermined number, the corresponding character image generating means and the second resolution further using the character image generated by The image processing apparatus according to claim 1, wherein a character image is generated. Cutting means for cutting out a character image included in a scanned image obtained by scanning a document;
Wherein a first character image specified from among the plurality of character images extracted by the cutout section, of the first vector data representing the character image distance second Ru der than the predetermined value a character image from the plurality of character images extracted by said cutout means and extracting means for extracting at a first resolution,
A character image extracted first character image and the second by the extraction means, the first of said first calculated at a higher than one pixel accuracy in the resolution of the character image and the second character image And generating means for generating a character image having a second resolution that is higher than the first resolution using the phase of
It said generating means, based on the first character image and the second character image phase in the first resolution, the pixel value of the first character image and the second character image, the first An image processing apparatus that interpolates pixel values of a character image at a resolution of 2. Cutting means for cutting out a character image included in a scanned image obtained by scanning a document;
Wherein a first character image specified from among the plurality of character images extracted by the cutout section, of the first vector data representing the character image distance second Ru der than the predetermined value and character image from the cutout section by cut out said plurality of character image extraction means for extracting at a first resolution,
A character image extracted first character image and the second by the extraction means, the first of said first calculated at a higher than one pixel accuracy in the resolution of the character image and the second character image Removing means for removing noise of the first character image and the second character image based on the phase of
Higher than the first resolution using the first character image, the second character image, and the phase of the first character image and the second character image from which noise has been removed by the removing means An image processing apparatus comprising: generating means for generating a character image having a second resolution which is a resolution. Cutting means for cutting out a character image included in a scanned image obtained by scanning a document;
Wherein a first character image specified from among the plurality of character images extracted by the cutout section, of the first vector data representing the character image distance second Ru der than the predetermined value and character image from the cutout section by cut out said plurality of character image extraction means for extracting at a first resolution,
A character image extracted first character image and the second by the extraction means, the first of said first calculated at a higher than one pixel accuracy in the resolution of the character image and the second character image Generating means for generating a character image having a second resolution higher than the first resolution using the phase of
The position in the second resolution of the character image is calculated based on the position of the extracted first character image and the second character image to the extraction means, the second generated by the generating means An image enlarging apparatus comprising: arrangement means for arranging a character image having a resolution of 2. A cutout means for cutting out a statement strokes image included in the read image obtained by reading the document,
Wherein a first character image specified from among the plurality of character images extracted by the cutout section, of the first vector data representing the character image distance second Ru der than the predetermined value a character image extraction means for extracting a character image from the cutout section by cut out said plurality of character images at a first resolution,
Wherein the character image extracted is extracted by means of the first character image and the second character image, the first of said first calculated at a higher than one pixel accuracy in resolution character image and the second by using the character images phase, a character image generating means for generating a character image of the second resolution is a higher resolution than the first resolution,
Encoding for encoding the character image of the second resolution generated by the character image generation means, and the phase and position of the first character image and the second character image extracted by the extraction means. An image encoding apparatus comprising: means. The claims 10 image encoding apparatus, comprising: a decoding means for decoding said second resolution of the character image, the code data and the phase and position is encoded in the first resolution of the character image,
A character image decoded in the second resolution by the decoding means, based on the phase of the first resolution of the character image, decoding the character image generating means for generating a character image of the first resolution When,
Based on the decoded position by said decoding means, the image decoding apparatus and a character image arrangement means for arranging the output image a character image generated by the decoded character image generating means. And the magnification of a character image generated by the decoded character image generating means, the image decoding apparatus of claim 11, further comprising a magnification-position control means for controlling the position which is located by the character image arrangement unit. A cutout means for cutting out a statement strokes image included in the read image obtained by reading the document,
Wherein a first character image specified from among the plurality of character images extracted by the cutout section, of the first vector data representing the character image distance second Ru der than the predetermined value a character image extraction means for extracting a character image from the cutout section by cut out said plurality of character images at a first resolution,
Wherein the character image extracted is extracted by means of the first character image and the second character image, the first of said first calculated at a higher than one pixel accuracy in resolution character image and the second by using the character images phase, a character image generating means for generating a character image of the second resolution is a higher resolution than the first resolution,
Encoding for encoding the character image of the second resolution generated by the character image generation means, and the phase and position of the first character image and the second character image extracted by the extraction means. Means,
Decoding means for decoding the code data encoded by the encoding means;
A character image decoded in the second resolution by the decoding means, based on the phase of the first resolution of the character image, decoding the character image generating means for generating a character image of the first resolution When,
The image processing system and a character image arranging means based on said decoded position by the decoding means, to place the character image generated by the decoded character image generating means in the output image. In an image processing apparatus including a computer,
A cutting step for cutting out a character image included in a scanned image obtained by scanning a document;
A first character image specified from among the plurality of character images the cut out, and the second character image Ru value less der which distance is predetermined between the vector data representing the character image of the first an extraction step of extracting at a first resolution from said plurality of character images cut out said,
Said first character image and the second character image the extracted, and the first the calculated higher than one pixel accuracy in the resolution of the first character image and the second character image phase A program for causing the computer of the image processing apparatus to execute a generating step of generating a character image having a second resolution that is higher than the first resolution using In an image processing apparatus including a computer,
A cutting step for cutting out a character image included in a scanned image obtained by scanning a document;
A first character image specified from among the plurality of character images the cut out, and the second character image Ru value less der which distance is predetermined between the vector data representing the character image of the first Extracting at a first resolution from the cut out character images;
Said first character image and the second character image the extracted, and the first the calculated higher than one pixel accuracy in the resolution of the first character image and the second character image phase Generating a character image having a second resolution that is higher than the first resolution using the computer, and causing the computer of the image processing apparatus to execute
It said generating step, based on said first character image and the second character image phase in the first resolution, the pixel value of the first character image and the second character image, the first A program that interpolates pixel values of character images at 2 resolutions. In an image processing apparatus including a computer,
A cutting step for cutting out a character image included in a scanned image obtained by scanning a document;
A first character image specified from among the plurality of character images the cut out, and the second character image Ru value less der which distance is predetermined between the vector data representing the character image of the first Extracting at a first resolution from the extracted plurality of character images;
Said first character image and the second character image the extracted, and the first the calculated higher than one pixel accuracy in the resolution of the first character image and the second character image phase A removing step of removing noise of the first character image and the second character image, based on
The resolution is higher than the first resolution using the first character image and the second character image from which the noise has been removed and the phase of the first character image and the second character image. A program for causing a computer of the image processing apparatus to execute a generation step of generating a character image having a second resolution. In an image enlarging apparatus including a computer,
A cutting step for cutting out a character image included in a scanned image obtained by scanning a document;
First character image and a second character image Ru value less der which distance is determined in advance as vector data representing the character image of the first identified from among the plurality of character images the cut Extracting at a first resolution from the clipped character images; and
Said first character image and the second character image the extracted, and the first the calculated higher than one pixel accuracy in the resolution of the first character image and the second character image phase Generating a character image of a second resolution that is higher than the first resolution using
The position in the extracted first character image and the second of said calculated based on the position of the character image a second resolution of the character image, a character image of the generated second resolution A program for causing a computer of the image enlargement apparatus to execute an arrangement step of arranging. In an image encoding device including a computer,
A cutout step of cutting out a statement strokes image included in the read image obtained by reading the document,
A first character image specified from among the plurality of character images the cut out, and the second character image Ru value less der which distance is predetermined between the vector data representing the character image of the first , a character image extraction step of extracting at a first resolution from a plurality of character images cut out said,
Said first character image and the second character image the extracted, and the first the calculated higher than one pixel accuracy in the resolution of the first character image and the second character image phase using a character image generating step of generating a character image of the second resolution is a higher resolution than the first resolution,
An encoding step for encoding the generated character image of the second resolution and the phase and position of the extracted first character image and the second character image; To be executed by other computers. In an image decoding apparatus including a computer,
By claim 18 programs, the decoding step of decoding said second resolution of the character image, the code data and the phase and position is encoded in the first resolution of the character image,
A character image of the second resolution, which is the decoded, based in the phase of the first resolution of the character image, and the decoded character image generating step of generating a character image of the first resolution,
A program that causes a computer of the image decoding apparatus to execute a character image arrangement step of arranging the generated character image in an output image based on the decoded position.

Images