JP4710672B2 - Character color discrimination device, character color discrimination method, and computer program - Google Patents

Description

  The present invention relates to an apparatus and a method for determining the color of a character included in an image input by an image input apparatus such as a scanner.

  A technique in which an image drawn on paper is scanned with a scanner and stored as electronic data has become widespread. Usually, in order to obtain electronic data of a high-quality image, image processing corresponding to the characteristics is often performed on the scanned image. For example, image processing for increasing the character identification is performed on a region including a character image representing a character. That is, image processing is performed so that the user can clearly understand what character the character image represents.

  As an image processing method for a scanned character area, a method as described in Patent Document 1 has been proposed. According to this method, a binary image of a character region is generated from a color image, a character rectangle is created for the binary image of the character region, single color determination of the character rectangle is performed, and a plurality of non-monochromatic character rectangles It is determined that the set is a single-color character set, and the character inside the determination rectangle is cut. As a result, information on the character color and the character region can be extracted.

  According to the method described in Patent Document 2, a document image is input as a color digital image, the background color of the document image is specified, and pixels other than the background region are extracted from the document image using the background color Then, the connected components are generated by integrating the pixels, and the connected components are classified and identified into areas such as characters / ruled lines / drawings / photographs from the shape features and the color features. Character rectangle data is acquired as a character area that is an identified area, a pixel corresponding to the background color in the character rectangle is removed, a representative color is calculated based on the pixel values of the remaining pixels, and the representative color is calculated. The character color of the character rectangle.

In addition, Patent Document 3 discloses a method of determining a photo area, a color character area, and a black character area from an input image. From the color signal of the input image, the color determination unit determines whether each pixel is a color pixel. In the color determination correction unit, the color character cut is connected by the expansion processing by the first expansion processing unit to the area determined as the color pixel by the color determination unit, and the color of the black character part by the contraction processing by the contraction processing unit The pixel is removed, and the color pixel region is expanded by the second expansion processing unit so as to include the color character portion. Thereby, the color pixel determination of the black character portion is removed, and a portion determined not to be a color pixel of the color character portion is also a color pixel. On the other hand, a photographic structure and a character structure are extracted by the structure extraction unit, and a photo / character character determination unit determines each region of the photograph, color character, and black character by logical operation together with the color determination result corrected by the color determination correction unit.
JP 2004-242075 A JP 2002-236922 A Japanese Patent Laid-Open No. 11-4344

  By the way, usually, the edge part of the scanned character image, that is, the pixel near the boundary with the background image (background image) of the character is mixed with the color of the background image, and the color of the character image is often blurred. . Pixels far from the background image are less blurred and the original color of the character appears clearly.

  When the character line represented by the character image is thick, it is possible to select many pixels that are distant from the background image, so that the character color can be accurately determined to some extent even using the conventional method as described above. be able to.

  However, when the character line represented by the character image is thin, many pixels of the character image are located near the boundary with the background image, so that the color is blurred. That is, there are fewer pixels in which the original color clearly appears compared to a character image with thick character lines. Therefore, the accuracy of the determined color is poor in the conventional method.

  In view of such problems, it is an object of the present invention to determine the color of a character represented by a character image input by a scanner or the like more accurately than in the past.

A character color discriminating device according to the present invention comprises a central region detecting means for detecting a central region consisting of a central line of a line figure of a character appearing in an input image input by an image input device and a region in the vicinity thereof, and a plurality of pixels. a Li Kui plurality of small areas a portion of the center line is located at the center, and the small area extracting means for extracting from said central region, for each of the small areas, meet a given condition from among the pixels belonging to the small region Condition matching pixel detection means for detecting a pixel as a condition matching pixel, and color determination means for determining a color of a character appearing in the input image based on color information of the condition matching pixel in each of the small regions. Features.

  Preferably, the central area detecting means detects a thin line detected by performing thinning processing on the input image or an area in the vicinity thereof as the central area, and the predetermined condition is that the brightness is the highest. The color discrimination means calculates a peak of the gradation of the primary color of the condition-matching pixel for each primary color, and calculates the color represented by the peak of each primary color of the character appearing in the input image. It is determined that the color.

  In the present invention, “character” includes not only kanji, hiragana, katakana, and alphabet, but also numerals such as numbers “0” to “9” and operators.

  According to the present invention, the color of a character represented by a character image input by a scanner or the like can be determined more accurately than before.

  1 is a diagram showing an example of a system configuration having an image forming apparatus 1 according to the present invention, FIG. 2 is a diagram showing an example of a hardware configuration of the image forming apparatus 1, and FIG. 3 is a functional configuration of the image forming apparatus 1. It is a figure which shows an example.

  An image forming apparatus 1 shown in FIG. 1 is an image processing apparatus that integrates various functions such as copying, scanning, faxing, network printing, document server, and file transfer. Sometimes called a multi-function peripheral or MFP (Multi Function Peripherals).

  The image forming apparatus 1 is installed in offices or offices such as businesses, public facilities such as schools or libraries, stores such as convenience stores, and other various places and can be shared by a plurality of users. Further, it can be connected to a terminal device 2 such as a personal computer or a workstation via a communication line 3. As the communication line 3, the Internet, a LAN, a public line, a dedicated line, or the like is used.

  As shown in FIG. 2, the image forming apparatus 1 includes a CPU 10a, a RAM 10b, a ROM 10c, a hard disk 10d, a control circuit 10e, an operation panel 10f, a scanner unit 10g, a printing unit 10h, a modem 10j, and a network interface 10k. The

  The scanner unit 10g is an apparatus that optically reads an image such as a photograph, a character, a picture, or a chart drawn on a document sheet (hereinafter simply referred to as “document”) and converts it into electronic data. The read image data is stored in the RAM 10b, and is subjected to various processes and filed as described later. In this embodiment, a color scanner is used.

  The printing unit 10h is an apparatus that prints an image read by the scanner unit 10g or an image transmitted from the terminal device 2 or the like on a sheet using toners of four colors of yellow, magenta, cyan, and black. .

  The operation panel 10f includes an operation unit and a display unit. A numeric keypad or the like is used as the operation unit, and a liquid crystal display or the like is used as the display unit. By operating the operation unit, the user gives an instruction to start or stop the processing to the image forming apparatus 1 and designates processing conditions such as a data destination, a scanning condition, or an image file format, Various other items can be specified. The display unit includes a screen for giving a message or an instruction to the user, a screen for inputting the type of processing and processing conditions desired by the user, and a screen showing the result of the processing executed by the image forming apparatus 1 Etc. are displayed. When a touch panel is used as the operation panel 10f, the touch panel serves as both an operation unit and a display unit. As described above, the operation panel 10 f serves as a user interface for a user who operates the image forming apparatus 1. Note that an application program and a driver for giving a command to the image forming apparatus 1 are installed in the terminal device 2. Therefore, the user can remotely operate the image forming apparatus 1 with the terminal device 2.

  The modem 10j has a built-in NCU (Network Control Unit), and is connected to another fax terminal via an analog public line to perform data control based on a facsimile protocol, fax data modulation / demodulation, and the like. The network interface 10k is a NIC (Network Interface Card), and connects to the terminal device 2 and the like via a LAN or the Internet, and transmits and receives files.

  The control circuit 10e is a circuit for controlling devices such as the hard disk 10d, the operation panel 10f, the scanner unit 10g, the printing unit 10h, the modem 10j, and the network interface 10k.

  The hard disk 10d stores programs and data for realizing functions such as the preprocessing unit 101, the non-character region processing unit 102, the character region processing unit 103, and the PDF file generation unit 104 as shown in FIG. ing. These programs are read into the RAM 10b as necessary, and the programs are executed by the CPU 10a. Some or all of these programs or data may be stored in the ROM 10c. Alternatively, part or all of the functions of FIG. 3 may be realized by the control circuit 10e.

  FIG. 4 is a diagram illustrating an example of document images GA0 and GA0 ′, and FIG. 5 is a diagram illustrating a configuration example of the non-character area processing unit 102.

  Next, functions and processing contents of each unit of the image forming apparatus 1 shown in FIG. 3 will be described.

  The preprocessing unit 101 converts a document image input by the scanner unit 10g (hereinafter referred to as “document image GA0”) to a resolution corresponding to the processing after the non-character area processing unit 102. (Resolution conversion processing) is performed, and further, processing (background removal processing) for removing the color of the background portion of the document image GA0 is performed. The “background portion of the document image GA0” means an area where there is no content (object). For example, in the case of a manuscript image GA0 as shown in FIG. 4, a rectangular area with an image of a character string “new digital MFP 255255”, a character string “The price difference of a digital machine can be expected ...”. A portion other than a total of three rectangular regions including a rectangular region with an image and a rectangular region in which a photograph of the MFP is arranged at the center is a background portion of the document image GA0.

  The document image GA0 includes images of various types of objects such as character images, photographic images, and background images. Hereinafter, a region including a character image is referred to as a “character region RM”, and a region including only an image of an object other than a character image is referred to as a “non-character region RH”. For example, the document image GA0 of FIG. 4 has two character areas RM and one non-character area RH. The document image GA0 that has been subjected to the resolution conversion process and the background removal process is referred to as “document image GA0 ′”.

  The character region RM includes a line image representing a character and a background image of the line. Hereinafter, the former is referred to as “character image”, and the latter is referred to as “character background image”. For example, when a document having red characters written on white paper is scanned, a red pixel group in the character region RM becomes a character image, and a white pixel group becomes a character background image. However, near the boundary between the character line and the background, the color may not be distinguished due to color blur. This problem will be dealt with by processing that will be described later.

  As shown in FIG. 5, the non-character area processing unit 102 includes a lightness calculation unit 121, a photographic background extraction unit 122, a gradation correction unit 123, a non-character reproduction data generation unit 124, and the like. Image processing or the like is performed on the non-character area RH in GA0 ′, that is, an area such as a photograph and a background.

  The brightness calculation unit 121 generates the brightness image GA1 by calculating the brightness of each pixel of the preprocessed document image GA0, that is, the document image GA0 '. For example, when a full-color (that is, 24-bit) document image GA0 is input, the brightness is calculated by the following method.

The gradation value of each primary color of R (red), G (green), and B (blue) of the document image GA0 ′ obtained by the preprocessing is reduced to an 8-bit gradation. These values are substituted into a known calculation formula such as the following formula (1).
Y = 0.3 × R + 0.6 × G + 0.1 × B (1)
However, Y is the brightness, R is the gradation value of the R component, G is the gradation value of the G component, and B is the gradation value of the B component.

  If necessary, smoothing processing may be performed to remove noise from the brightness image GA1 and correct the contour.

  The photographic background extraction unit 122 performs labeling processing on the brightness image GA1, thereby determining the position of an image area other than characters (non-character area RH) such as a photograph or background included in the document image GA0 ′. It discriminate | determines and extracts non-character area | region RH. In general, when a labeling process is performed on an image such as a photograph, a large white or black lump appears, and when a labeling process is performed on a character image, many irregular thin lines appear. Based on such characteristics, the position of the non-character region RH is determined. Then, non-character area position data 81 indicating the determination result is generated. In the case of FIG. 4, non-character area position data 81 indicating the position of the non-character area RHa is generated.

  The tone correction unit 123 corrects the tone of the image in the non-character region RH extracted by the photo background extraction unit 122 as necessary using a known method.

  The non-character reproduction data generation unit 124 includes image data and non-character region position of each non-character region RH corrected by the gradation etc. correction unit 123 (hereinafter referred to as “non-character image GC”). By integrating the data 81, non-character area image reproduction data 71 is generated.

  Information regarding the position of each non-character region RH and suitable image data (image data) included in the document image GA0 'can be obtained by the processing of each unit constituting the non-character region processing unit 102 as described above. Hereinafter, the image data of the non-character area RH obtained as a result of the image processing by the non-character area processing unit 102 will be referred to as “non-character area image reproduction data 71”.

  6 is a diagram illustrating an example of the configuration of the character region processing unit 103, FIG. 7 is a diagram illustrating an example of the character region RM divided into a plurality of blocks BR, and FIG. 8 illustrates an example of the configuration of the document character color determination unit 135. FIG. 9 is a diagram for explaining an example of a method for generating a thinned image GS, FIG. 10 is a diagram showing an example of the position of the center line of a character line graphic, and FIG. FIG. 12 is a diagram for explaining an example of a method of dividing a nearby region into a plurality of character color candidate regions RC, FIG. 12 is a flowchart for explaining an example of the flow of approximate color pixel selection processing, and FIG. It is a figure which shows the example of distribution of a gradation.

  Returning to FIG. 3, the character area processing unit 103, as shown in FIG. 6, includes a lightness calculation unit 131, a character area extraction unit 132, a character region division unit 133, a binarization processing unit 134, and a document character color determination unit 135. The character background processing unit 136, the character reproduction data generation unit 137, and the like, and performs image processing on an area including the character image, that is, the character area RM.

  The lightness calculation unit 131 generates the lightness image GA1 by calculating the lightness of each pixel of the document image GA0 ', similarly to the lightness calculation unit 121 of FIG.

  The character region extraction unit 132 performs a labeling process on the lightness image GA1, determines the position of the character region RM included in the document image GA0 'based on the characteristics of the processing result, and extracts the character region RM. Then, character area position data 80 indicating the determination result is generated. In the case of FIG. 4, character area position data 80 indicating each position is generated for each of the two character areas RMa and RMb.

  The brightness calculation unit 131 and the brightness calculation unit 121 may be shared, and the character area extraction unit 132 and the photographic background extraction unit 122 may be shared. That is, the brightness image obtained by one brightness calculation unit may be used in common by each unit in the subsequent processing. Further, the character area extraction unit 132 and the photographic background extraction unit 122 are shared as one area extraction unit, and the character area RM and the non-character area RH are selected from the document image GA0 ′ based on the result of one labeling process. Both may be extracted, and character area position data 80 and non-character area position data 81 indicating the respective positions may be generated.

  The character area dividing unit 133 groups a plurality of characters arranged in the horizontal direction or the vertical direction when a large number of characters appear in the character area image GM. For example, about 1 to 10 characters are grouped. In the case of English sentences, etc., they may be grouped for each word (that is, for each character group between the spaces). In horizontal writing, characters arranged in the left-right direction are grouped, and in vertical writing, characters arranged in the vertical direction are grouped. Whether it is horizontal writing or vertical writing can be determined based on the characteristics of the arrangement of the character images. Then, a portion including the character image of each grouped character is extracted as a block BR from the character region RM. Thereby, the character area RM is divided into a plurality of blocks BR. For example, the character area RMb (see FIG. 4) is divided into a plurality of blocks BR (BR1, BR2,...) As shown in FIG. For each block BR, block position data 82 indicating the position in the character area RM is generated.

  The binarization processing unit 134 generates a binary image GN of each block BR of the character area RM obtained by the character area dividing unit 133. In the present embodiment, the binary image GN of the block BR is generated as follows. The density of each pixel of the image of the block BR is calculated. In the present embodiment, the density is represented by 8 bits. If the calculated density is equal to or higher than the density threshold β, the value of the pixel is “1”, and if the calculated density is less than the density threshold β, the value of the pixel is “0”. Thus, by converting the value of each pixel into either “0” or “1”, a binary image GN is generated.

  According to the generated binary image, the character image portion and the character background image portion of the character region RM respectively included in the document image GA0 ', the brightness image GA1, and the like can be distinguished. That is, it is assumed that a pixel “1” indicates a character image portion and a pixel “0” indicates a character background image portion.

  As shown in FIG. 8, the document character color determination unit 135 includes a low resolution image generation unit 301, a thin line image generation unit 302, a character image division unit 303, an approximate color pixel selection unit 304, a character color component acquisition unit 305, and A character color determining unit 306 and the like are used to determine the color of characters drawn on the original (hereinafter referred to as “original character color”). In the present embodiment, the document character color is determined for each block BR. Here, the content of the process of determining the document character color will be described by taking as an example the case where a block BRx including a character image representing the character group “255255” as shown in FIG.

  FIG. 9A shows a portion (hereinafter referred to as “block document image GA0x”) corresponding to the block BRx in the original document image GA0 ′. FIG. 9B shows a portion corresponding to the block BRx in the lightness image GA1 of the original image GA0 ′ (hereinafter referred to as “block lightness image GA1x”). FIG. 9C is a binary image GN of the block BRx.

  The low resolution image generation unit 301 in FIG. 8 generates a low resolution image GT as shown in FIG. 9D by reducing the resolution of the binary image GN of the block BRx. In the present embodiment, it is assumed that the document images GA0 and GA0 ′, the block document image GA0x, and the binary image GN have a resolution of 300 dpi, and the low resolution image generation unit 301 generates a 75 dpi low resolution image GT. .

  The thinned image generation unit 302 generates a thinned image GS as shown in FIG. 9E by performing a known thinning process on the low resolution image GT. That is, the low resolution image GT is generated by detecting the center line from a line figure having a thickness of 1 pixel (pixel) or more appearing in the low resolution image GT. For example, a center line as shown in FIG. 10 is detected from the line figure of the character “2”. In the present embodiment, a thinned image GS representing a thin line having a thickness of 1 pixel (pixel) is generated. Note that the resolution of the thinned image GS is the same as the resolution of the low-resolution image GT and is 75 dpi. By the processing of the thinned image generating unit 302, it is possible to find a pixel group (region) located at the center line of the character line figure and its vicinity.

  Based on the thinned image GS, the character image dividing unit 303 divides the center line of the character line figure represented by the character image included in the block document image GA0x and the vicinity thereof into a plurality of character color candidate regions RC. The process of dividing the character color candidate area RC is performed by the following method, for example.

  One pixel on the thin line of the thinned image GS corresponds to a plurality of pixels in the character image in the block document image GA0x. That is, for example, one pixel PX on the thin line in the thinned image GS indicated by the circular dotted line frame in FIG. 11A corresponds to the original block original image GA0x indicated by the rectangular dotted line frame in FIG. It corresponds to a plurality of pixels on a character line. The character image dividing unit 303 extracts these pixel clusters as the character color candidate region RC. In the present embodiment, a character color candidate region RC of 4 × 4 pixels is extracted from the relationship between the resolutions of the thinned image GS and the block document image GA0x, as shown in FIG. Similarly, by extracting a block of pixels corresponding to other pixels on the thin line, the center line of the character line figure represented by the character image of the block original image GA0x and its vicinity are divided into a plurality of character color candidate regions RC. can do. A plurality of character color candidate regions RC are extracted by the processing of the character image dividing unit 303.

  The approximate color pixel selection unit 304 selects a pixel (hereinafter referred to as an “approximate color pixel PK”) that is considered to represent a color closest to the original character color from each character color candidate region RC obtained by the character image division unit 303. Is selected in accordance with the procedure shown in FIG.

  Paying attention to one character color candidate region RC, out of the 4 × 4 pixels constituting the character color candidate region RC, pixels located at the top, bottom, left and right edges (for example, outside the thick line frame in FIG. 11C) 12 pixels located) are excluded from the candidates (# 402 in FIG. 12). This is because these pixels are closer to the character background image than the pixels located inside the thick line frame, and therefore there is a high possibility that color blur has occurred and the reliability is low.

  The brightness of each pixel that remains without being excluded, that is, 2 × 2 pixels located inside the thick line frame in FIG. 11C is calculated. Then, the pixel having the lowest brightness (that is, the pixel having the highest density) is selected as the pixel representing the color closest to the original character color, that is, the approximate color pixel PK (# 403). For the lightness, the lightness image GA1 generated by the lightness calculation unit 131 in FIG. 6 may be referred to.

  Similarly, the processing of steps # 402 and # 403 is executed for other character color candidate regions RC, and an approximate color pixel PK is selected.

  Through the above processing, a large number of approximate color pixels PK are selected for one block BR. Specifically, the approximate color pixels PK corresponding to the number of pixels constituting the thin line of the thin line image GS (see FIG. 9E) of the block BR are selected.

  For each approximate color pixel PK selected by the approximate color pixel selection unit 304, the character color component acquisition unit 305 in FIG. 8 performs gradations of RGB primary colors of the approximate color pixel PK in the character image of the block BR. Get the value of. The pixel value can be obtained by searching a portion corresponding to the block BR in the original document image GA0 ′ and referring to the pixel value at the position of the approximate color pixel PK in the block BR. . The acquired gradation value of each primary color has a distribution like a histogram shown in FIG. 13, for example.

  The character color determination unit 306 determines the color of the character represented by the character image of the block BR based on, for example, the following method based on the distribution of gradation values of the primary colors acquired by the character color component acquisition unit 305. To do.

  A peak value is obtained for each distribution of gradation values of each primary color acquired by the character color component acquisition unit 305. This is determined by a known method. For example, the pixel value having the largest number may be selected as the peak value, or may be obtained by a known calculation method. The color reproduced by the respective peak values of R, G, and B thus determined is determined as the character color represented by the character image of the block BR. Then, it is determined that this is the color of the character represented by the character image of the block BR.

  The color of the character represented by the character image of each block BR is determined by the processing of each part constituting the document character color discriminating unit 135 as described above. Then, the determined color is determined as the original character color in the block BR. In addition, the document character color determination unit 135 generates character color data 83 indicating the determined document character color for each block BR.

  Returning to FIG. 6, the character background processing unit 136 deletes the character image from the image of each block BR of each character region RM in the document image GA0 ′, so that an image of only the character background portion, that is, the character background image. Image data is generated. Hereinafter, the character background image obtained by the character background processing unit 136 is referred to as “character background image GH”. The deleted pixel portion of the character background image GH is blank. In addition, what is necessary is just to refer the binary image GN for the position of a character image. That is, the character background image GH can be generated by deleting from the image of the block BR the part corresponding to the position of the pixel whose pixel value is “1” in the binary image GN.

  The character reproduction data generation unit 137 generates character area reproduction data 70 for reproducing an image of each character area RM. The character area reproduction data 70 includes block position data 82 of each block BR of each character area RM, binary image GN, that is, image data (image data) of a line figure of the character image, character color data 83, and character background image. It is generated by integrating GH image data (image data) and character region position data 80 of each character region RM.

  The image processing is performed on the character region RM by the processing of each part constituting the character region processing unit 103 as described above, and the character region reproduction data 70 is generated.

  FIG. 14 is a diagram showing an example of the structure of the document reproduction file FL. Returning to FIG. 3, the PDF file generation unit 104 uses the non-character region position data 81 generated by the non-character region processing unit 102, the character region position data 80 generated by the character region processing unit 103, and the like. A document reproduction file FL for reproducing an image of a document input by the unit 10g is generated by the following procedure, for example.

  The image data of the binary image GN (line figure of the character image) and the image data of the character background image GH included in the character region reproduction data 70 are compressed. Since the former is binary data, it is desirable to compress it by G4 compression method (MMR compression method). Since the latter is image data of color or pattern, it is desirable to compress it by a compression method such as GIF or JPEG.

  In parallel with or before or after the compression processing of the binary image GN and the character background image GH, the image data of the non-character image GC included in the non-character area image reproduction data 71 is compressed. Since an image of a large size such as a photograph is arranged in the non-character area RH, the image is compressed so as to be reduced in size by an irreversible compression method such as JPEG. Alternatively, the image type may be determined for each non-character image GC, and a compression method such as JPEG or GIF may be used properly.

  Further, file attribute data 72 indicating the attributes (file name, size, creator, etc.) of the document reproduction file FL to be generated is generated. A part of the attribute (for example, a file name) may be specified by the user.

  Then, the character area reproduction data 70 of each character area RM subjected to the compression process, the non-character area image reproduction data 71 of each non-character area RH, and the generated file attribute data 72 are integrated into a file. In this way, the document reproduction file FL is generated. For example, when an original image GA0 as shown in FIG. 4 is input by the scanner unit 10g and each part of FIG. 3 executes processing on the original image GA0, an original reproduction file FL having a data structure as shown in FIG. Generated.

  If there are a plurality of areas of the same type of object, these areas may be integrated into one and processing for reducing the total number of objects may be performed.

  The generated document reproduction file FL is stored in the hard disk 10d. Alternatively, it is transferred to the terminal device 2 or the like via the communication line 3 by e-mail or FTP. The document reproduction file FL may be generated in accordance with an existing format such as Acrobat PDF. The PDF is described in, for example, the following known technical document 1.

[Prior Art Document 1] “What is Adobe PDF?-PDF Beginner's Edition”, Adobe Systems, May 14, 2005 search, Internet <URL: http://www.adobe.co.jp/products /acrobat/adobepdf13.html>
By using the generated document reproduction file FL, an image of the document can be reproduced. Here, a procedure for reproducing an image of a document will be described.

  The image forming apparatus 1 or the terminal device 2 activates application software corresponding to the document reproduction file FL (for example, Acrobat Reader of Adobe Systems if the document reproduction file FL is a PDF file) and reproduces the document. Load file FL into RAM. The compressed data is decompressed as appropriate.

  The image of the character region RM is reproduced in the following procedure based on the character region reproduction data 70 related to the character region RM. For each block BR constituting the character region RM, the character background image GH is reproduced and the line figure of the character image is reproduced. The line figure of the character image is colored with the original character color indicated by the character color data 83 related to the block BR. The image of the entire character region RM is reproduced by fitting the line figure of the colored character image into the character background image GH to reproduce the image of the entire block BR and disposing it at the position indicated by the block position data 82. . When there are a plurality of character areas RM, that is, when a plurality of character area reproduction data 70 are included in the document reproduction file FL, this processing is executed for each character area RM.

  Based on the non-character area image reproduction data 71 related to the non-character area RH, the image of the entire non-character area RH, that is, the non-character image GC, is processed before or after or in parallel with the process of reproducing the image of the entire character area RM. To reproduce. When there are a plurality of non-character areas RH, that is, when a plurality of non-character area image reproduction data 71 are included in the document reproduction file FL, this processing is executed for each non-character area RH.

  Then, the reproduced image of the entire character area RM is arranged at a position indicated by the corresponding character area position data 80. Further, the reproduced image of the entire non-character area RH is arranged at a position indicated by the corresponding non-character area position data 81. As a result, the image related to the document reproduction file FL is reproduced.

  The reproduced image can be output to the user by printing it on paper or displaying it on a display.

  15 is a flowchart for explaining an example of the overall processing flow of the image forming apparatus 1, FIG. 16 is a flowchart for explaining an example of the flow of non-character area processing, and FIG. 17 is a flow of character area processing. It is a flowchart for demonstrating an example.

  Next, a flow of a series of processing for scanning a document image and generating a document reproduction file FL will be described with reference to a flowchart.

  In FIG. 15, the image forming apparatus 1 scans an image of a document to be filed (# 1), and performs preprocessing such as resolution conversion processing and background removal processing on the obtained image (# 2). ).

  A non-character area image of the non-character area RH is obtained by performing processing as shown in FIG. 16 on an area in which an object (photograph or the like) other than a character is drawn in the original image, that is, the non-character area RH. Reproduction data 71 is generated (# 3).

  That is, the brightness of each pixel of the document image is calculated (# 11 in FIG. 16), and the non-character region RH is extracted by executing a labeling process (# 12). At this time, position information (non-character area position data 81) of each non-character area RH is also acquired.

  Then, correction processing such as the gradation of the image of the non-character region RH is performed as necessary (# 13), and the non-character region image reproduction data 71 is obtained based on the image data, the non-character region position data 81, and the like. Generate (# 14).

  Returning to FIG. 15, before or after or in parallel with the process of step # 3, the process shown in FIG. 17 is performed on the area where the character is drawn, that is, the character area RM in the document image. Thus, the character area reproduction data 70 of the character area RM is generated.

  That is, the brightness of each pixel of the document image is calculated (# 21 in FIG. 17), and the character region RM is extracted by executing a labeling process (# 22). At that time, position information (character area position data 80) of each character area RM is also extracted. The lightness calculation process and the labeling process may be performed in common with the processes of steps # 11 and # 12 in FIG.

  When many characters are shown in the character region RM, it is divided into a plurality of blocks BR (# 23), and a binary image GN is generated for each block BR (# 24). With this binary image GN, the character image and the character background image in the block BR can be separated.

  A process of determining the color of the character represented by the character image of each block BR is performed. That is, a pixel (approximate color pixel PK) close to the original color of the character written on the original is selected (# 25), and the original character color is determined based on these pixel values (# 26). Note that the method of selecting the approximate color pixel PK is as described above with reference to FIG. Data of the character background image GH is generated before or after or in parallel with the process of step # 25 (# 27).

  Then, based on the data of the binary image GN, the character color data 83, the character background image GH, and the block position data 82 and the character area position data 80 of the character area RM, the character area reproduction data 70 is obtained. Is generated (# 28).

  Returning to FIG. 15, file attribute data 72 indicating attributes relating to the document reproduction file FL is generated (# 5). Then, the non-character area image reproduction data 71 of each non-character area RH and the character area reproduction data 70 of each character area RM in the original image obtained by the processes of steps # 3 and # 4 are compressed (# 6) The original reproduction file FL is generated by integrating the compressed data and the file attribute data 72 (# 7).

  According to this embodiment, the character color of the original is determined by referring to a pixel as close as possible to the center line of the character line figure. Therefore, the influence of color blur can be reduced and the original character color can be determined more accurately than in the past.

  Further, according to the present embodiment, since the process for determining the document character color is performed using the binary image GN and the thinned image GS with reduced resolution, the time required for the process can be shortened.

  FIG. 18 is a flowchart for explaining a modified example of the flow of the approximate color pixel selection process, and FIG. 19 is a diagram showing an example of the template TMP and the reference pixel position frame SGP.

  In the present embodiment, the process for determining the color represented by the character image (step # 25 in FIG. 17) is performed by the method described with reference to FIGS. That is, a character color candidate region RC corresponding to each pixel on the thin line of the thinned image GS as shown in FIG. 11A is extracted from the document image GA0 ′ and approximated from each extracted character color candidate region RC. The color pixels PK are selected one by one, and the original character color is determined based on the gradation values of the RGB primary colors of the approximate color pixels PK. However, the approximate color pixel PK may be selected from other than the character color candidate region RC. For example, the approximate color pixel PK may be selected by a method as shown in FIG.

  Templates TMP (TMP1 to TMP4) as shown in FIGS. 19A to 19D are prepared in advance. Each of these templates TMP is associated with a reference pixel position frame SGP (SGP1 to SGP4) for obtaining the position of the reference pixel as shown on the right side of the drawing. Note that one cell of the templates TMP1 to TMP4 corresponds to one pixel in the resolution (75 dpi) of the thinned image GS. On the other hand, one cell of the reference pixel position frames SGP1 to SGP4 corresponds to one pixel in the resolution (300 dpi) of the document image GA0 '.

  First, the approximate color pixels PK are selected one by one from the character color candidate regions RC corresponding to the respective pixels on the thin line of the thinned image GS (# 501 to # 503 in FIG. 18). The contents of this process are the same as the contents of the process described above with reference to FIG. That is, the pixel with the lowest brightness is selected as the approximate color pixel PK from the central portion excluding the end of the character color candidate region RC.

  Templates TMP1 to TMP4 are used as follows to further select approximate color pixels PK. In the thinned image GS, attention is paid to a pixel adjacent to the thin line, and the cell indicated by the center of the template TMP1, that is, the thick line frame in FIG. The template TMP1 is overlapped so as to match (# 506).

  At this time, if the thin line of the thinned image GS overlaps any one of the predetermined cells indicated by gray of the template TMP (Yes in # 507), the approximate color pixel PK is used using the reference pixel position frame SGP1 corresponding to the template TMP1. The pixel to be referred to as a candidate for is determined as follows (# 508). That is, the reference pixel position frame SGP1 is superimposed on 4 × 4 pixels corresponding to the target pixel in the document image GA0 ′. The pixel that overlaps the cell of the thick line frame in the reference pixel position frame SGP1 is determined as a pixel to be referred to.

  Then, the pixel with the lowest brightness is selected as the approximate color pixel PK from these pixels to be referred to (# 509).

  The processes of steps # 506 to # 509 are also executed for other pixels adjacent to the thin line of the thinned image GS. Further, similar processing is executed using the templates TMP2 to TMP4. As described above, the approximate color pixel PK is added.

  The processing to be executed after selecting the approximate color pixel PK is as described in this embodiment. That is, the original character color is determined by obtaining a histogram for each RGB component of the selected approximate color pixel PK (see FIG. 13) and obtaining a peak value.

  In the present embodiment, only one approximate color pixel PK is selected from the center four pixels of the character color candidate region RC (see FIG. 11C), but a plurality of approximate color pixels PK may be selected. For example, all the four central pixels may be selected. Similarly, in the selection method described with reference to FIG. 18, only one of the three pixels overlapping the thick line frame cell of the reference pixel position frame SGP (see FIG. 19) is selected as the approximate color pixel PK. Alternatively, two or three may be selected.

  In the present embodiment, the case where the image forming apparatus 1 has all of the functions for reading an image of a document, the function for correcting an image, the function for filing image data, and the like has been described as an example. It can also be realized by being distributed to a plurality of devices. For example, a function for reading an image of a document may be realized by a scanner for a personal computer, and other functions may be realized by a personal computer. In this case, a program or the like for realizing the functions of the respective units in FIG. 3 is installed in the personal computer. Alternatively, the image data of the document image GA0 read by the image forming apparatus 1 may be transferred to the terminal device 2 so that the terminal device 2 executes the processing of each unit in FIG.

  In this embodiment, the case where an image of a document read by a scanner is converted into a file has been described as an example. However, the present invention can also be applied to a case where an image acquired by another type of image reading apparatus is converted into a file. is there. For example, the present invention can be applied to file an image of a document shot with a digital camera.

  Rich text format data may be generated as data for reproducing the character region RM.

  When the color of the character background image is darker than the character image in the character region RM, that is, when a so-called white character is drawn, each part of the image forming apparatus 1 changes the processing content as follows. The binarization processing unit 134 in FIG. 6 calculates the density of each pixel of the block BR. If the calculated density is less than the density threshold β, the pixel value is set to “1”, and if the calculated density is equal to or greater than the density threshold β. The value of the pixel is “0”. That is, the pixel with the lower density is a pixel constituting the character image. The approximate color pixel selection unit 304 in FIG. 8 uses, as the approximate color pixel PK, a pixel having the highest lightness, that is, the pixel having the lowest density among the four pixels at the center of the character color candidate region RC (see FIG. 11C). elect.

  In this embodiment, the center line of the character line figure and the area in the vicinity thereof are extracted by generating a thinned image GS having a lower resolution than the image of the block BR, but a thinned image is generated without reducing the resolution. However, it may be extracted based on this. In this case, the center line and the area in the vicinity thereof can be known by superimposing the thinned image and the image of the block BR. Then, a plurality of 4 × 4 pixel regions as shown in FIG. 11C may be selected from the region, and the subsequent processing may be performed.

  In addition, the configuration of the entire image forming apparatus 1 or each unit, processing contents, processing order, table contents, and the like can be appropriately changed in accordance with the spirit of the present invention.

1 is a diagram illustrating an example of a system configuration having an image forming apparatus according to the present invention. 2 is a diagram illustrating an example of a hardware configuration of an image forming apparatus. FIG. 2 is a diagram illustrating an example of a functional configuration of an image forming apparatus. FIG. It is a figure which shows the example of a manuscript image. It is a figure which shows the example of a structure of a non-character area | region process part. It is a figure which shows the example of a structure of a character area process part. It is a figure which shows the example of the character area | region RM divided | segmented into the some block. It is a figure which shows the example of a structure of a character color discrimination | determination part. It is a figure for demonstrating the example of the production | generation method of a thin line image. It is a figure which shows the example of the position of the center line of the line figure of a character. It is a figure for demonstrating the example of the method of dividing | segmenting the centerline of a character line figure, and the area | region of the vicinity into a some character color candidate area | region. It is a flowchart for demonstrating the example of the flow of an approximate color pixel selection process. It is a figure which shows the example of distribution of the gradation of each primary color. 3 is a diagram illustrating an example of a structure of a document reproduction file. FIG. 5 is a flowchart for explaining an example of the overall processing flow of the image forming apparatus. It is a flowchart for demonstrating the example of the flow of a non-character area | region process. It is a flowchart for demonstrating the example of the flow of a character area process. It is a flowchart for demonstrating the modification of the flow of an approximate color pixel selection process. It is a figure which shows the example of a template and a reference pixel position frame.

Explanation of symbols

1 Image forming device (character color discrimination device)
302 Thinned image generating unit (central area detecting means)
303 Character image segmentation unit (small area extraction means)
304 Approximate color pixel selection unit (condition matching pixel detection means)
306 Character color determination unit (color discrimination means)
RC character color candidate area (small area)
PK approximate color pixel (condition matching pixel)

Claims (5)

A center area detecting means for detecting a center area consisting of a center line of a line figure of a character appearing in an input image input by the image input device and an area in the vicinity thereof;
A small area extracting means for extracting a small area portion of Li Kui said center line, such a plurality of pixels located at the center plurality, from the central region,
Condition matching pixel detection means for detecting, as a condition matching pixel, a pixel that meets a predetermined condition from the pixels belonging to the small area for each of the small areas;
Color discriminating means for discriminating the color of the character appearing in the input image based on the color information of the condition-matching pixels of the small areas;
A character color discriminating apparatus characterized by comprising: The central area detecting means detects a thin line detected by performing a thinning process on the input image or an area in the vicinity thereof as the central area,
The predetermined condition is that the brightness is the lowest,
The color determining means calculates a peak of the gradation of the primary color of the condition-matching pixel for each primary color, and determines that the color represented by the peak of each primary color is the color of a character appearing in the input image. To
The character color discrimination device according to claim 1. Center line detecting means for detecting the center line of the line figure of the character appearing in the input image input by the image input device;
Condition-matching pixel detection means for detecting a pixel that meets a predetermined condition as a condition-matching pixel from each of the surrounding areas in each direction of the centerline pixel,
Color discriminating means for discriminating the color of the character appearing in the input image based on the color information of the condition-matching pixels of the small areas;
A character color discriminating apparatus characterized by comprising: Detect the center area consisting of the center line of the line figure of the character appearing in the input image input by the image input device and the area near it,
Extracting a small area portion of Li Kui said center line, such a plurality of pixels located at the center plurality, from the central region,
For each small area, a pixel that meets a predetermined condition is detected as a condition-matching pixel from among the pixels belonging to the small area.
Determining the color of characters appearing in the input image based on the color information of the condition-matching pixels of the small areas;
Character color discrimination method characterized by the above. A computer program used in a computer that handles an input image input by an image input device,
A process for detecting the central line and the central region consisting of a region near the line graphic characters appearing in the input image,
A plurality of small areas a portion of Li Kui said center line, such a plurality of pixels located at the center, the process of extracting from said central region,
For each of the small areas, a process of detecting a pixel that satisfies a predetermined condition from among pixels belonging to the small area as a condition matching pixel;
A process of determining the color of a character appearing in the input image based on color information of the condition-matching pixels of the small areas;
A computer program for causing a computer to execute.