JP5818559B2 - Image processing device - Google Patents

Description

  The present invention relates to an image processing apparatus including an attribute determination unit that determines whether each pixel of image data obtained by reading a document image is a pixel belonging to a halftone dot region.

  Conventionally, based on image data obtained by photoelectrically reading a manuscript image in a copying machine or the like, character areas, photo areas, halftone areas, etc. in the image data are determined, and image processing corresponding to each area is performed. There has been proposed a method for acquiring a high-quality printed image by performing the above.

  For example, character edges can be sharpened by applying relatively strong edge enhancement processing to character areas, sharp edges can be added to image areas by applying relatively weak edge enhancement processing, and halftone dots. It has been proposed to reduce the occurrence of moire by performing a smoothing process on the region.

  Here, when the resolution of the original image is close to the print resolution but slightly deviated, moire tends to occur. In other cases, for example, the resolution of the print image is low and error diffusion processing is used. In the case of a copying machine that generates output image data, it is not necessary to perform the above-described smoothing process because moire is unlikely to occur in the halftone dot area.

  However, when the discrimination between the halftone dot area and the character area is not properly performed, the halftone dot area is discriminated as the character area, which causes a problem that edge emphasis processing is performed and moire occurs. In addition, for example, when edge enhancement processing is performed on only a part of a halftone dot region, there is a problem that a density difference between the partial region and a peripheral region occurs, and the quality of a printed image is significantly deteriorated. is there.

  In order to solve such a problem, various methods for detecting a halftone dot area with high accuracy have been proposed. For example, in Patent Document 1, a peak pixel is detected based on each pixel data of image data, and as shown in FIG. 16, the number of peak pixels PP in a reference region R centered on a predetermined target pixel is counted. However, a method for discriminating the pixel of interest as a pixel belonging to a halftone dot region when the peak pixel is equal to or greater than a predetermined threshold has been proposed.

JP-A-2-219366

  However, according to the halftone dot attribute determination method of Patent Document 1, it is necessary to find a plurality of peak pixels PP as shown in FIG. 16, for example, an original image with a halftone dot of 100 lines and 45 degrees is photoelectrically generated at 600 dpi. In the case of reading the image, it is necessary to secure a reference area of at least 9 pixels × 9 pixels as shown in FIG. FIG. 17 is a schematic diagram, and the relationship between the line spacing of 100 lpi and the pixel size of 600 dpi is not accurate.

  Furthermore, if it is intended to increase the determination accuracy of the halftone dot attribute, it is necessary to detect a larger number of peak pixels, so it is necessary to secure a wider reference area, and the size corresponding to the reference area. A memory with a capacity is required, which increases costs.

  SUMMARY OF THE INVENTION In view of the above circumstances, an object of the present invention is to provide an image processing apparatus that can appropriately determine a halftone dot region without requiring a large-capacity memory as described above.

  An image processing apparatus according to the present invention includes a peak candidate pixel detection unit that detects a peak candidate pixel in image data obtained by reading a document image, and a peak candidate pixel in a reference region centered on a predetermined target pixel in the image data. And an attribute determination unit that determines that the pixel of interest is a pixel belonging to a halftone dot area when the number of non-peak candidate pixels other than is greater than or equal to a predetermined halftone dot determination threshold.

  In the image processing apparatus of the present invention described above, a line-likeness calculating unit that calculates an index indicating line-likeness is provided for a predetermined pixel of interest in image data, and the attribute determining unit is a line calculated by the line-likeness calculating unit. It can be determined that the pixel of interest is a pixel belonging to a character when the index indicating the likelihood is equal to or greater than a predetermined character determination threshold.

  In addition, an edge pixel determination unit that determines whether or not a predetermined pixel of interest in the image data is an edge pixel is provided, and the attribute determination unit has an index that determines that the pixel of interest is an edge pixel and represents lineiness When it is equal to or greater than the character determination threshold, it can be determined that the pixel of interest is a pixel belonging to the character.

  In addition, the attribute determination unit determines whether the number of non-peak candidate pixels in the reference area centered on a predetermined pixel of interest in the image data is less than the halftone threshold and the index indicating the line-likeness of the pixel of interest is character determination. If it is less than the threshold and greater than or equal to the predetermined threshold, it can be determined that the pixel of interest is a pixel that belongs between the character and the halftone dot.

  Further, the attribute determination unit determines that the number of non-peak candidate pixels in the reference region centered on the predetermined pixel of interest in the image data is less than the halftone threshold, and the index indicating the line-likeness of the pixel of interest is predetermined. When it is less than the threshold value, it can be determined that the pixel of interest is a pixel belonging to the photograph.

  According to the image processing apparatus of the present invention, peak candidate pixels in image data obtained by reading a document image are detected, and non-peaks other than peak candidate pixels in a reference region centered on a predetermined target pixel in image data are detected. Since the target pixel is determined to be a pixel belonging to the halftone dot area when the number of candidate pixels is equal to or greater than a predetermined halftone dot determination threshold, the halftone dot is based on the number of peak pixels as described above. Compared with the method for determining whether or not the reference value is smaller, the reference area can be made smaller, and the capacity of the memory used for attribute determination can be reduced.

  Further, in the image processing apparatus according to the present invention, an index indicating the lineiness is calculated for a predetermined pixel of interest in the image data, and the pixel of interest is equal to or greater than a predetermined character determination threshold value. Is determined to be a pixel belonging to a character, it is possible to detect a thin line or character without erroneously determining the halftone dot as a thin line or character. As a result, it is possible to prevent the moire from being generated due to the strong edge emphasis processing applied to the halftone dots, and it is possible to clarify the outline by applying the strong edge emphasis processing to the fine lines and characters.

  In addition, it is determined whether or not a predetermined target pixel in the image data is an edge pixel, and it is determined that the target pixel is an edge pixel, and the target pixel is equal to or greater than a character determination threshold value Is determined to be a pixel belonging to a character, the character can be detected more appropriately, and the character existing in the photo area can also be detected appropriately. Note that the photographic area here means an area of the entire photographic image including characters and pictures.

  In addition, the number of non-peak candidate pixels in a reference area centered on a predetermined target pixel in the image data is less than a halftone dot determination threshold, and an index indicating the line-likeness of the target pixel is lower than a character determination threshold If the pixel of interest is determined to be a pixel belonging between the character and the halftone dot when the threshold value is equal to or greater than the threshold, determine a pixel having an intermediate attribute between the character and the halftone dot, By applying such an edge emphasis process or density conversion process between characters and halftone dots to such a pixel with an attribute that is not clearly a character, without applying a strong edge emphasis process like a character, It is possible to obtain a print image with an image quality without a sense of incongruity.

  In addition, when the number of non-peak candidate pixels in the reference area centered on a predetermined target pixel in the image data is less than the halftone dot determination threshold, and the index indicating the line-likeness of the target pixel is lower than the predetermined threshold Or, if all the pixels in the reference area are non-peak candidate pixels, and if it is determined that the pixel of interest is a pixel belonging to the photograph, it is possible to appropriately detect the pixel having the attribute of the photograph. And edge enhancement processing and density conversion processing according to the attribute can be performed.

1 is a block diagram showing a schematic configuration of a printing apparatus using an embodiment of an image processing apparatus of the present invention. The block diagram which shows the structure of the image process part shown in FIG. The block diagram which shows the structure of the halftone dot likeness acquisition part shown in FIG. 6 is a flowchart for explaining the overall processing flow of a printing apparatus using an embodiment of the image processing apparatus of the present invention. Flowchart for explaining operation of halftone dot likeness acquisition unit Diagram for explaining the response of a Laplacian filter to line images and isolated points The figure which shows an example of the Laplacian filter coefficient used for a peak candidate pixel detection The figure for demonstrating the case where the Laplacian amount of the orthogonal Laplacian filter coefficient is the same sign, and a different sign The figure which shows an example of the reference area for counting the number of non-peak candidate pixels The figure for demonstrating the difference with the ratio of the peak candidate pixel of a halftone dot area | region, and a non-peak candidate pixel, and the ratio of the peak candidate pixel of a character or a line, and a non-peak candidate pixel Flow chart for explaining the operation of the linearity acquisition unit The figure which shows an example of a line detection filter coefficient The figure which shows an example of the differential filter coefficient for edge pixel determination Flow chart for explaining the operation of the attribute determination unit The figure which shows an example of the gamma curve used when performing a density conversion process with respect to the pixel of a photograph attribute, and the gamma curve used when performing a density conversion process with respect to the pixel of a character attribute The figure for demonstrating the determination method of the conventional halftone dot area | region The figure for demonstrating the magnitude | size of the reference area required in the determination method of the conventional halftone dot area | region

  Hereinafter, an embodiment of a printing apparatus using an image processing apparatus of the present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram illustrating a schematic configuration of a printing apparatus according to the present embodiment.

  As shown in FIG. 1, the printing apparatus 1 according to the present embodiment includes a document image reading unit 10 that photoelectrically reads a document image to acquire image data, and a character area in the image data read by the document image reading unit 10. A halftone dot area, a photograph area, and the like, based on the image processing section 20 that performs different image processing adapted to each area, and processed image data that has undergone image processing in the image processing section 20 And a printing unit 30 that performs a printing process on the printing paper.

  The document image reading unit 10 includes a line image sensor that photoelectrically reads image information of a document image, scans the document image with the line image sensor, reads the document image, and converts the image data representing the document image into bitmap data. Is output as In the present embodiment, each pixel data constituting the bitmap data is represented by 8 bits and can take a value of 0 to 255.

  As shown in FIG. 2, the image processing unit 20 includes a halftone dot degree obtaining unit 21 that obtains an index representing the halftone dot degree for each pixel based on the input image data, and each of the image processing unit 20 based on the input image data. An edge pixel determination unit 22 that determines whether or not a pixel is an edge pixel, a line-likeness acquisition unit 23 that acquires an index representing the line-likeness of each pixel based on the input image data, and the above-described halftone dot-likeness Are determined by the attribute determination unit 24 and the attribute determination unit 24 that determine attributes such as a character region, a halftone dot region, and a photo region for each pixel based on the index indicating the edge, the edge pixel determination result, and the index indicating the lineiness. Edge enhancement processing unit that applies edge enhancement processing to each pixel data based on the attribute and density conversion processing adapted to the attribute for each pixel data A density conversion processing section 26 to perform, and a halftone processing section 27 for performing halftoning process on the image data subjected to the edge enhancement processing and density conversion processing. The detailed operation of each of the above parts will be described in detail later.

  FIG. 3 is a block diagram showing a more specific configuration of the halftone dot likelihood obtaining unit 21 in the image processing unit 20. As shown in FIG. 3, the halftone dot likelihood acquisition unit 21 temporarily stores input image data and a peak candidate for detecting a peak candidate pixel based on the image data stored in the data memory 40. Based on the pixel detection unit 41, the peak memory 42 that stores the position information of the peak candidate pixel detected by the peak candidate pixel detection unit 41, and the peak candidate pixel detected by the peak candidate pixel detection unit 41, a predetermined reference is made. A halftone dot likelihood calculation unit 43 is provided that calculates the number of non-peak candidate pixels in the region and acquires this as an index representing the halftone dot likelihood. The detailed operation of each of the above parts will also be described in detail later.

  The printing unit 30 performs printing on printing paper based on the processed image data that has been subjected to edge enhancement processing and density conversion processing in the image processing unit 20. As the printing part 30, what performs stencil printing, inkjet printing, laser printing, etc. can be used, for example.

  Next, the operation of the printing apparatus 1 of the present embodiment will be described with reference to FIGS. 4 to 15 as appropriate. Note that the flowchart shown in FIG. 4 shows the flow of the entire process from reading an original image to printing processing on printing paper in the printing apparatus 1. Further, the flowchart shown in FIG. 5 shows the process of obtaining the halftone dot quality (S12) shown in FIG. 4 in more detail, and the flowchart shown in FIG. 11 is the acquisition of the line likeness shown in FIG. 4 (S16). ) In more detail, and the flowchart shown in FIG. 14 shows the attribute determination (S18) process shown in FIG. 4 in more detail.

  First, as shown in the flowchart of FIG. 4, a document is placed on a document table of the document image reading unit 10, scanned by a line image sensor while being pressed by a pressing plate, and image data is read (S10).

  The image data read by the document image reading unit 10 is input to the halftone dot acquisition unit 21, the edge pixel determination unit 22, and the lineiness acquisition unit 23. An index representing the halftone dot is acquired, the edge pixel determination unit 22 determines the edge pixel, and the line likelihood acquisition unit 23 acquires an index representing the line-likeness of each pixel (S12 to S16).

  Here, first, the processing of the halftone dot likelihood obtaining unit 21 will be described with reference to the flowchart shown in FIG.

  The image data output from the document image reading unit 10 is sequentially stored in units of five lines in the data memory 40 of the halftone dot acquisition unit 21 (S30).

  In the peak candidate pixel detection unit 41, detection of peak candidate pixels serving as a basis for halftone dot detection is performed based on the image data for five lines stored in the data memory 40. A Laplacian amount that is a second derivative of the density amplitude of data is calculated, and a peak candidate pixel is detected based on the Laplacian amount. The reason why the Laplacian amount is used in this way is that the Laplacian amount is more sensitive to the isolated point as shown on the right side of FIG. 6 than the original image of the line as shown on the left side of FIG.

  First, the peak candidate pixel detection unit 41 calculates a Laplacian amount for each pixel using four sets of eight types of filter coefficients shown in FIGS. 7A to 7D (S32). The two filter coefficients in each group of FIGS. 7A to 7D are created so that numerical values other than 0 are arranged in directions orthogonal to each other. Also, the four sets of filter coefficients in FIGS. 7A to 7D are created so that four different Laplacian amounts in the orthogonal direction can be calculated. In the present embodiment, the filter coefficients of 5 rows × 5 columns are used. However, the present invention is not limited to this, and a filter coefficient of 3 rows × 3 columns smaller than this may be used.

  Specifically, the peak candidate pixel detection unit 41 calculates Lap1 [a] and Lap2 [a] by calculating the following equations using the two filter coefficients shown in FIG. Note that data (j, i) indicates the pixel data value of the coordinate value (j, i) of a predetermined target pixel, and j is the vertical direction as shown in FIGS. It means a coordinate value, and i means a coordinate value in the horizontal direction.

Lap1 [a] = − data (j, i) × 2 + (j−2, i) + data (j + 2, i)
Lap2 [a] = − data (j, i) × 2 + (j, i + 2) + data (j + 2, i−2)
Further, the peak candidate pixel detection unit 41 calculates Lap1 [b] and Lap2 [b] by calculating the following equations using the two filter coefficients shown in FIG. 7B, and FIG. Using the two filter coefficients shown in Fig. 7, Lap1 [c] and Lap2 [c] are calculated by calculating the following formula, and the following formula is calculated using the two filter coefficients shown in Fig. 7 (d). Thus, Lap1 [d] and Lap2 [d] are calculated.

Lap1 [b] = − data (j, i) × 2 + (j−2, i + 1) + data (j + 2, i−1)
Lap2 [b] = − data (j, i) × 2 + (j + 1, i + 2) + data (j−1, i−2)
Lap1 [c] = − data (j, i) × 2 + (j−2, i + 2) + data (j + 2, i−2)
Lap2 [c] = − data (j, i) × 2 + (j + 2, i + 2) + data (j−2, i−2)
Lap1 [d] = − data (j, i) × 2 + (j−1, i + 2) + data (j + 1, i−2)
Lap2 [d] = − data (j, i) × 2 + (j + 2, i + 1) + data (j−2, i−1)
Next, the peak candidate pixel detection unit 41 refers to the four sets of Laplacian amounts calculated for each pixel, and detects peak candidate pixels for pixels in which the Laplacian amounts Lap1 and Lap2 have the same sign in all sets. Pixels for which the Laplacian amounts Lap1 and Lap2 of any of the four sets of Laplacian amounts have different signs are excluded from the peak candidate pixel detection targets (S34, NO, S36). S38).

  Here, the signs that the Laplacian amounts Lap1 and LaP2 in the directions orthogonal to each other have different signs mean that the pixels have a density distribution as shown in FIG. 8A, and the Laplacian amounts Lap1 and LaP2 are the same. The sign is a pixel having a density distribution as shown in FIG. A pixel having a density distribution as shown in FIG. 8A is likely to be a line or a character, and a pixel having a density distribution as shown in FIG. 8B is an isolated point such as a halftone dot. There is a high possibility.

  Next, the peak candidate pixel detection unit 41 adds the respective Laplacian amounts Lap1 and Lap2 for the pixels that are the candidates for peak candidate pixel detection in S36, and acquires the added value as the peak value Pval (S40). ). When the four peak values Pval acquired for each pixel are all positive, the maximum value among them is acquired as the maximum value Plus_Max, and when the four peak values Pval are all negative, The one having the maximum absolute value is acquired as the minimum value Minus_Max (S42).

  Here, the fact that the four peak values Pval acquired for each pixel are all positive means that the positive peak component greatly contributes, that is, there is a high possibility that the original image is a valley. Become. On the other hand, if the four peak values Pval are all negative, the negative peak component greatly contributes, that is, the possibility that the original image is a mountain is high.

  Next, the absolute value of the maximum value Plus_Max or the minimum value Minus_Max acquired as described above is compared with a preset peak threshold value (S44), and the absolute value of the maximum value Plus_Max or the minimum value Minus_Max is the peak threshold value. If it is larger, the pixel is determined as a peak candidate pixel (S46). On the other hand, when the absolute value of the maximum value Plus_Max or the minimum value Minus_Max is equal to or less than the peak threshold value, the pixel is determined as a non-peak candidate pixel (S48). Note that the peak threshold value used at this time is preferably set to a value relatively lower than the peak value in order to detect peak candidate pixels around the peak with respect to the peak value obtained experimentally, for example. Further, the pixels excluded from the peak candidate pixel detection targets in S38 are also determined as non-peak candidate pixels.

  As described above, the peak candidate pixel detection unit 41 detects the peak candidate pixel, and information for specifying the position of the peak candidate pixel is stored in the peak memory 42.

  Next, the halftone dot likelihood calculation unit 43 refers to the peak memory 42 and, as shown in FIG. 9, the number of peak candidate pixels existing in a reference region of 5 pixels × 5 pixels centered on a predetermined target pixel P. The number of non-peak candidate pixels is calculated by subtracting the number of peak candidate pixels from the total number of 25 pixels in the reference region, and the number of non-peak candidate pixels is obtained as an index representing the dot-likeness of the target pixel. (S50). Then, an index representing the dot-likeness of each pixel is obtained by performing the same calculation as above with all the pixels constituting the image data sequentially as the target pixel. At this time, if all the pixels in the reference area are non-peak candidate pixels, the target pixel is determined to be an attribute of a photo as described later, with an index indicating the likelihood of halftone dots being 255. To do. Note that paper margins other than the so-called photo area are all non-peak candidate pixels, and are judged as photo attributes.

  In the present embodiment, as described above, the number of non-peak candidate pixels, not the number of peak candidate pixels, is used as an index representing the likelihood of halftone dots. When the number of pixels is compared with the number of non-peak candidate pixels, for example, in the case of a halftone dot region, the number of non-peak candidate pixels is larger than the number of peak candidate pixels as shown in FIG. If the region is a line region, as shown in FIG. 10B, the ratio of the number of peak candidate pixels increases and the number of non-peak candidate pixels decreases compared to the halftone dot region. This is because, by using the number of pixels as an index representing the likelihood of halftone dots, it is possible to appropriately determine a halftone dot region in distinction from characters and lines.

  In addition, since the number of peak candidate pixels and the number of non-peak candidate pixels in such a relatively narrow region are detected, the capacity of the data memory 40 and the peak memory 42 described above can be reduced.

  However, for characters including thin lines or simple thin lines, the number of non-peak candidate pixels increases even within a relatively narrow area, and these areas may be erroneously determined as halftone dot areas. There is.

  Therefore, in the present embodiment, as described above, the edge pixel determination unit 22 and the linearity acquisition unit 23 are provided, and these processing results are also used to increase the accuracy of region determination.

  Although details will be described later, in this embodiment, an area that belongs between the halftone dot and the character is set instead of being determined as one of the halftone dot area and the character area. In other words, an intermediate image process between characters and halftone dots is performed to obtain a print image with an uncomfortable image quality. The processing result of the above-described linearity acquisition unit 23 is also used to determine the region that belongs between the halftone dot and the character.

  Hereinafter, the processing (S16 in FIG. 4) of the line likeness acquisition unit 23 will be described with reference to the flowchart shown in FIG.

  First, the line likelihood acquisition unit 23 calculates four line component amounts line1 to line4 for the predetermined target pixel using the four line detection filter coefficients shown in FIG. 12 (S60). Since the calculation method of the line component amount is the same as that of the normal filter calculation, the detailed calculation formula will not be described.

  Then, the line component amounts of the line detection filter coefficients orthogonal to each other are set as one set, that is, line 1 and line 2 are set as one set, line 3 and line 4 are set as one set, and the absolute value difference of each line component amount is set for each set. It is calculated as line_value (S62).

  Next, the larger value of the two line_values calculated as described above is acquired as an index representing the linearity (S64). At this time, for convenience of attribute determination to be described later, since the index indicating the linearity needs to have 128 as the maximum value, the value obtained by dividing line_value by 12 is acquired as the index indicating the linearity. Then, an index representing the linearity of each pixel is obtained by performing the same calculation as above with all the pixels constituting the image data as the target pixel in order.

  Next, the process of the edge pixel determination unit 22 (S14 in FIG. 4) will be described. The edge pixel determination in the edge pixel determination unit 22 is performed in order to extract a contour portion of a character.

  Specifically, the edge pixel determination unit 22 calculates the edge amounts edge1 to edge4 using the four differential filter coefficients shown in FIG. 13 for a predetermined target pixel, and the edge pixel edge1 is the largest of the absolute values of the edge1 to edge4. A large value is acquired as Max_edge.

  Then, this Max_edge is compared with a preset edge threshold value. When Max_edge is larger than the edge threshold value, the target pixel is determined as an edge pixel. On the other hand, when Max_edge is equal to or smaller than the edge threshold, the target pixel is determined as a non-edge pixel. Then, a determination result indicating whether or not the pixel is an edge pixel is obtained by performing the same calculation as above with all the pixels constituting the image data being sequentially set as the target pixel.

  As described above, for all the pixels constituting the image data, the index indicating the halftone dot, the index indicating the lineiness, and the determination result indicating whether the pixel is an edge pixel are acquired.

  Then, the index indicating the dot-likeness of each pixel, the index indicating the lineiness, and the edge pixel determination result are input to the attribute determination unit 24, and the attribute determination unit 24 determines the attribute of each pixel based on these ( S18 of FIG. In the present embodiment, five attributes of fine line, character, halftone dot, between character and halftone dot, and photograph are set, and any one attribute is assigned to each pixel. Hereinafter, the attribute determination method of the attribute determination unit 24 will be described with reference to the flowchart shown in FIG. Here, an index representing the likelihood of a line is referred to as line_like, and an index representing the likelihood of a halftone dot is referred to as ami.

  The attribute determination unit 24 first refers to the line_like of each pixel. If the line_like is 50 or more (S70, YES), this is all rounded to 50 (S72). To do. Here, the reason why the line_like is rounded to 50 is because it has been found through experiments that the line_like of the character takes a value of 50 or less. However, this is an experimental example and is not limited thereto.

  Next, the attribute determining unit 24 determines whether or not the line_like of each pixel is larger than 38 (S74, YES), and determines that the attribute of the pixel is a thin line (S76). . The reason why 38 is used as the threshold value here is that it has been found from experiments that the line_like of the halftone dot does not take a value larger than 38. However, this is an experimental example and is not limited thereto.

  Next, it is determined whether or not the line_like of each pixel is greater than 27, and the value of ami is referred to. The line_like is a pixel greater than 27 or ami = 255, and the pixel is an edge pixel. In that case (S78, YES), the attribute of the pixel is determined to be a character (S80). The reason why 27 is used as the threshold value here is that it is known from experiments that line_like of some halftone dots can take the vicinity of 27, but to include more pixels around the character. is there. However, this is an experimental example and is not limited thereto. Further, here, when ami = 255 and an edge pixel, the attribute is determined to be a character in order to detect the edge portion in the photographic region here. In other words, in the present embodiment, edge portions and characters are determined as character attributes even in the photo area.

  Next, the ami of each pixel is referred to, and when the ami is greater than 8 (S82, YES), the attribute of the pixel is determined to be a halftone dot (S84). Here, the reason why 8 is used as the threshold value is that it has been found by experiments that the halftone dot ami is larger than 8, but this is only an experimental example and is not limited thereto.

  Next, the line_like of each pixel is referred to, and when the line_like is larger than 15 (S86, YES), it is determined that the attribute of the pixel is between a character and a halftone dot (S88). On the other hand, when line_like is 15 or less (S86, NO), the attribute of the pixel is determined to be a photograph (S90). Here, even when line_like is 38 or less and takes a relatively large value, the attribute is intermediate between a halftone dot and a character, and it is preferable to perform intermediate processing. Considering the value that can be said to be a photograph, 15 was adopted. However, this is an example and the present invention is not limited to this.

  As described above, the attribute determination unit 24 assigns attributes to all the pixels constituting the image data, and assigns an attribute value to each pixel as information indicating which attribute. Specifically, 0 as an attribute value indicating a thin line attribute, 16 as an attribute value indicating a character attribute, 128 as an attribute value indicating a halftone dot attribute, and 196 as an attribute value indicating an attribute between a character and a halftone dot. ˜230, 255 is assigned as an attribute value indicating the attribute of the photograph. The attribute value indicating the attribute between the character and the halftone dot changes from 196 to 230 depending on the size of line_like, and line_like changes from 16 to 50 according to the determination in S86 of FIG. Therefore, when line_like is 16, it is 230, and when line_like is 50, it is 196. That is, the larger the line_like, the smaller the attribute value.

  Then, based on the attribute value assigned to each pixel as described above, the edge enhancement processing unit 25 performs edge enhancement processing on each pixel data of the image data (S20 in FIG. 4). . The edge enhancement processing unit 25 of the present embodiment performs edge enhancement using a Laplacian filter. As for the Laplacian filter coefficient, a 3 × 3 or 5 × 5 filter which is usually used may be used.

  Specifically, the edge enhancement processing unit 25 calculates the following expression based on each pixel data data (j, i) constituting the image data and a Laplacian amount lap obtained using a Laplacian filter. As a result, the processed pixel sdata (j, i) is calculated.

sdata (j, i) = data (j, i) −lap × weight
The weight in the above equation changes according to the attribute value assigned to each pixel data data (j, i). For example, when the attribute value is 0 (thin line), weight = 1.5, When the attribute value is 16 (character), weight = 1 can be set, and when the attribute value is 128 (halftone dot) or 255 (photograph), weight = 0 can be set.

  When the attribute value is 196 to 230 (between characters and halftone dots), the weight is calculated by the following equation, for example. Note that line_like ′ = attribute value−196, and 35 in the following expression is a range of attribute values.

weight = (35−line_like ′) / 35
As described above, the weight is determined and the edge emphasis process is performed, whereby the fine line pixel data is subjected to the strong edge emphasis process, and the character pixel data is subjected to the normal edge emphasis process, and the character The pixel enhancement data between the halftone dots and the halftone dots is subjected to edge enhancement processing at a ratio corresponding to the index of lineiness. Then, it is possible to prevent the edge enhancement process from being applied to the pixel data of the halftone dots and the photograph.

  Next, after edge enhancement processing is performed in the edge enhancement processing unit 25, density conversion processing is performed on the edge enhanced pixel data in the density conversion processing unit 26 (S22 in FIG. 4).

  In the density conversion process, if the conversion is made so that the inside of the character is dark, the shadow portion of the photograph is crushed, resulting in an overall dark impression image.

  Therefore, in order to solve this problem, the density conversion processing unit 26 of the present embodiment performs density conversion processing on a character using a gamma curve having a strong contrast, and the photograph becomes brighter than the gamma curve of the character. The density conversion process is performed using such a gamma curve. Further, the density conversion processing is performed using the same gamma curve as that of characters for thin lines and the same gamma curve as that of photographs for halftone dots. FIG. 15 shows a gamma curve used when density conversion processing is performed on pixels having character and thin line attributes, and a gamma curve used when density conversion processing is performed on pixels having photo and halftone attributes. An example is shown.

  When the density conversion processing unit 26 performs the density conversion process on the attribute pixel between the character and the halftone dot, the gamma curve used for the pixel of the photograph shown in FIG. Both the gamma curve used for the pixel is used, and density conversion processing is performed by changing the ratio of these gamma curves based on the size of the target pixel's ami (an index of dot-likeness). Specifically, the density-converted pixel data outdata (j, i) is calculated based on the following equation.

outdata (j, i) =
{(35-line_like ′) * gamma1 [data (j, i)] + line_like ′ × gamma2 [data (j, i)]} / 35
gamma1: Expression showing a gamma curve for photographs and halftone dots gamma2: Expression showing a gamma curve for characters and fine lines As described above, density conversion processing corresponding to the attribute of each pixel data of image data is performed. The density converted image data is input to the halftoning processing unit 27. Then, in the halftoning processing unit 27, the halftoning process is performed on the density-converted image data (S24 in FIG. 4). Examples of the halftoning process include an error diffusion method and a dither method.

  The processed image data that has been subjected to the halftoning process in the halftoning processing unit 27 is output from the image processing unit 20 and input to the printing unit 30.

  Then, the printing unit 30 performs a printing process on the printing paper based on the input processed image data (S26 in FIG. 4).

DESCRIPTION OF SYMBOLS 1 Printing apparatus 10 Original image reading part 20 Image processing part 21 Halftone dot acquisition part 22 Edge pixel determination part 23 Linearity acquisition part 24 Attribute determination part 25 Edge emphasis processing part 26 Density conversion processing part 27 Halftoning process part 30 Printing part 40 Data memory 41 Peak candidate pixel detection unit 42 Peak memory 43 Halftone dot calculation unit

Claims (3)

A peak candidate pixel detection unit for detecting peak candidate pixels in image data obtained by reading a document image;
A linearity calculation unit for calculating an index representing the linearity for a predetermined pixel of interest in the image data;
When the number of non-peak candidate pixels other than the peak candidate pixel in the reference area centered on the predetermined target pixel in the image data is equal to or greater than a predetermined halftone dot determination threshold, the target pixel is included in the halftone dot area. The pixel of interest is determined to be a pixel belonging to a character when the index representing the linearity of the pixel of interest is greater than or equal to a predetermined character determination threshold, and the pixel of interest is centered When the number of the non-peak candidate pixels in the reference region is less than the halftone dot determination threshold and the index indicating the line-likeness of the target pixel is less than the character determination threshold and equal to or greater than a predetermined threshold An image processing apparatus comprising: an attribute determination unit that determines that the target pixel is a pixel that belongs between a character and a halftone dot . The attribute determination unit represents the number of the non-peak candidate pixels in a reference region centered on a predetermined pixel of interest in the image data that is less than the halftone dot determination threshold and represents the linearity of the pixel of interest. If the index is less than the predetermined threshold value, the image processing apparatus according to claim 1, wherein the pixel of interest is to determine that the pixel belongs to the photographic. An edge pixel determination unit that determines whether or not a predetermined pixel of interest in the image data is an edge pixel;
The attribute determining unit determines that the pixel of interest is a pixel belonging to a character when the pixel of interest is determined to be an edge pixel and the index indicating the linearity is equal to or greater than the character determination threshold. The image processing apparatus according to claim 1 , wherein the image processing apparatus is an image processing apparatus.