JP2006155572A - Image processor, image forming device, image processing method and computer program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processor, an image forming device, an image processing method and a computer program, performing edge emphasis processing or smoothing processing according to image data. <P>SOLUTION: This image processor has: a frequency component decision part 372a deciding whether an image is an image including an edge component or not, on the basis of a spatial frequency component of the image data; a block area decision part 372b deciding an attribute of the image of a block unit on the basis of area identification information outputted from an area separation processing part; and a processing content decision part 372c totally deciding processing to be executed in an arithmetic processing part 373 on the basis of a decision result of the block area decision part 372b and the frequency component decision part 372a. The decided processing is applied to the data having the prescribed spatial frequency component to convert the data in a frequency area into data in a density area by an inverse frequency conversion part 374. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an image processing apparatus that performs edge enhancement and smoothing of image data, an image forming apparatus, an image processing method that realizes the image processing apparatus, and a computer program.

  In the digital image processing technique, one of methods for obtaining a high-definition image is an edge enhancement process and a smoothing process of an image. For example, outlines of characters and continuous tone images (for example, photographic paper photographs) can be clearly seen by emphasizing the edges. For printed matter (printed photographs) formed with halftone dots, The occurrence of moire can be suppressed by performing a smoothing process. In these edge enhancement and smoothing processes, a spatial filter process using a digital filter is generally used (for example, see Non-Patent Document 1).

Similarly, as one of the methods for enhancing the edge component of the image, the original image is converted into the frequency domain by orthogonal transform, and the frequency data is changed using a constant prepared in advance. There has been proposed a method of obtaining a sharpened image by performing inverse orthogonal transform of the frequency data (see, for example, Patent Document 1).
The Journal of Electrophotographic Society, Vol. 36, No. 4 (1997), pages 343-352 JP-A-9-54825

  However, in the method described in Patent Document 1, the frequency data is changed by multiplying the data in the entire frequency range by a constant prepared in advance. There is a problem that unnecessary block noise is conspicuous in a gently curved region such as a contour of the.

  In addition, the frequency data is changed by multiplying the frequency data of the frequency domain subjected to orthogonal transformation by a value prepared in advance for each period of the frequency data to obtain a clearer image. There is a problem that a plurality of values must be set by increasing the value prepared in advance. When a plurality of values are set in this way, there is a problem that many registers are required in advance and the circuit scale becomes large.

  The present invention has been made in view of such circumstances, and an image processing apparatus, an image forming apparatus, the image processing apparatus, and an image forming apparatus capable of realizing highly accurate edge enhancement processing and smoothing processing without increasing the circuit scale. An object of the present invention is to provide an image processing method and a computer program for realizing the apparatus.

  An image processing apparatus according to the present invention generates image data composed of a plurality of spatial frequency components by performing orthogonal transform on image data, and performs image processing based on the generated data. Detecting means for detecting the characteristics of the plurality of spatial frequency components from data consisting of the plurality of spatial frequency components, means for applying processing corresponding to the detection result of the detecting means to data belonging to a predetermined spatial frequency region of the data, and the processing Means for performing inverse orthogonal transform on the data obtained by applying the data.

  In the present invention, it is provided with a detecting means for detecting image features from data composed of a plurality of spatial frequency components, and processing corresponding to the detection result is performed on data belonging to a predetermined spatial frequency range. Therefore, it is not necessary to perform processing for the entire spatial frequency, parameters used for processing are reduced, and as a result, the circuit scale is suppressed. Further, when the processing to be performed is edge enhancement processing, the enhancement region is limited, and the appearance of an unnecessary block pattern at the edge boundary portion on the curve can be suppressed. Furthermore, when edge enhancement processing is not necessary, smoothing can be performed by processing using, for example, two-dimensional matrix data, and a smoother image can be obtained.

  The image processing apparatus according to the present invention includes a determination unit that determines a magnitude relationship between the magnitude of the spatial frequency component and a preset threshold value, and the detection unit is characterized by the image based on a determination result of the determination unit. Is detected.

  In the present invention, since the feature of the image is detected by determining the magnitude relationship between the magnitude of the spatial frequency component and a preset threshold value, for example, the spatial frequency component in the low frequency range is compared with the threshold value. By doing so, it is easily detected whether or not an edge component is included in the target image.

  The image processing apparatus according to the present invention includes a calculation unit that performs a predetermined calculation using a direct current component and an alternating current component among the spatial frequency components, and a magnitude relationship between a value obtained by the calculation and a preset threshold value. Judgment means for judging is provided, and the detecting means detects the feature of the image based on the judgment result of the judging means.

  In the present invention, a predetermined calculation is performed using a direct current component and an alternating current component among the spatial frequency components, and a magnitude relationship between a value obtained by the calculation and a preset threshold value is determined to determine the characteristics of the image. Since detection is performed, a direct current component representing the average density of the entire image can also be taken in as a determination element, and edge block determination depending on the density of the entire image can be performed, thereby improving edge determination accuracy. In particular, a thin character on a highlight is easily determined as an edge block, and the character reproducibility is improved.

  The image processing apparatus according to the present invention is characterized in that the detection means detects whether or not the image is an edge image based on the determination result.

  In the present invention, since it is detected whether or not the image is an edge image based on the magnitude relationship between the magnitude of the spatial frequency component and a preset threshold value, for example, a character or a continuous tone image is detected. It is possible to perform processing that emphasizes edges for an image including a contour portion, and smoothing processing that suppresses the occurrence of moire for an image composed of halftone dots.

  The image processing apparatus according to the present invention is characterized in that the processing is edge enhancement processing or smoothing processing.

  In the present invention, in order to perform edge enhancement processing or smoothing processing according to the characteristics of the image, for example, edge enhancement processing that enhances the edge for an image including a character or a contour portion of a continuous tone image, An image composed of halftone dots can be subjected to a smoothing process that suppresses the occurrence of moire.

  In the image processing apparatus according to the present invention, when the detection unit detects that the image is an edge image, an edge enhancement process is performed, and the detection unit determines that the image is a non-edge image. If detected, a smoothing process is performed.

  In the present invention, edge enhancement processing is performed on edge images and smoothing processing is performed on non-edge images. Therefore, when edge enhancement is necessary, unnecessary block patterns are suppressed. When a clean edge curve is obtained and edge enhancement is not required, a smooth image can be obtained by smoothing.

  The image processing apparatus according to the present invention further includes an area determination unit that determines whether the image data belongs to a character area, a halftone dot area, or a photographic area.

  In the present invention, since the image data further includes area determination means for determining whether the image data belongs to a character area, a halftone dot area, or a photo area, the attribute of the image is comprehensively determined. The

  In the image processing apparatus according to the present invention, the detection unit detects that the image is an edge image, and the region determination unit determines that the image data belongs to a character region or a photo region. The edge determination processing is performed, the detection unit detects that the image is a non-edge image, and the region determination unit determines that the image data is data belonging to a halftone dot region or a photographic region. Is determined, a smoothing process is performed.

  In the present invention, edge enhancement processing is applied to data belonging to a character region or a photo region, and smoothing processing is performed on data belonging to a halftone dot region or a photo region. Therefore, when edge enhancement is necessary, a clean edge curve with an unnecessary block pattern suppressed can be obtained, and when edge enhancement is unnecessary, a smooth image is obtained by smoothing. can get.

  The image processing apparatus according to the present invention multiplies the data belonging to the spatial frequency range by a real number greater than 1 when performing the edge enhancement process, and divides the data by a real number greater than 1 when performing the smoothing process. It is characterized by the above.

  In the present invention, when edge enhancement processing is performed, data belonging to a predetermined spatial frequency region is multiplied by a real number larger than 1, and when smoothing processing is performed, division is performed by a real number larger than 1. Therefore, it is not necessary to prepare parameters over the entire spatial frequency region, and the circuit scale can be suppressed. In particular, when edge enhancement processing is performed, an image with more effective edges can be obtained by changing the real number to be multiplied according to the spatial frequency.

  An image forming apparatus according to the present invention includes the image processing apparatus according to any one of the above-described inventions, and a unit that forms an image on a sheet based on image data processed by the image processing apparatus. It is characterized by.

  In the present invention, when performing processing according to the characteristics of the image, in order to form an image on the sheet based on image data obtained by processing data belonging to a predetermined spatial frequency range, for example, An image including a clean edge curve in which the appearance of unnecessary block patterns at the edge boundary portion is suppressed, and a smooth image obtained by smoothing are formed on the sheet.

  An image processing method according to the present invention is an image processing method that performs orthogonal transformation on image data to generate data composed of a plurality of spatial frequency components, and performs image processing based on the generated data. A first step of detecting the characteristics of the data from data consisting of the plurality of spatial frequency components, a second step of applying a process according to the detection result to data belonging to a predetermined spatial frequency region of the data, and performing the process And a third step of performing inverse orthogonal transform on the data obtained in this manner.

  In the present invention, the feature of the image is detected from data composed of a plurality of spatial frequency components, and processing corresponding to the detection result is performed on data belonging to a predetermined spatial frequency range. There is no need to perform processing for the entire frequency range, and the parameters used for processing are reduced. Further, when the processing to be performed is edge enhancement processing, the enhancement region is limited, and the appearance of an unnecessary block pattern at the edge boundary portion on the curve can be suppressed. Furthermore, when edge enhancement processing is not necessary, smoothing can be performed by processing using, for example, two-dimensional matrix data, and a smoother image can be obtained.

  In the image processing method according to the present invention, the first step is a step of determining a magnitude relationship between the magnitude of the spatial frequency component and a preset threshold value, and detecting the feature of the image based on the determined result. It is characterized by that.

  In the present invention, since the feature of the image is detected by determining the magnitude relationship between the magnitude of the spatial frequency component and a preset threshold value, for example, the spatial frequency component in the low frequency range is compared with the threshold value. By doing so, it is easily detected whether or not an edge component is included in the target image.

  In the image processing method according to the present invention, in the first step, a predetermined calculation is performed using a DC component and an AC component of the spatial frequency components, and a value obtained by the calculation and a preset threshold value are calculated. It is a step of determining the magnitude relationship and detecting the feature of the image based on the determined result.

  In the present invention, a predetermined calculation is performed using a direct current component and an alternating current component among the spatial frequency components, and a magnitude relationship between a value obtained by the calculation and a preset threshold value is determined to determine the characteristics of the image. Since detection is performed, a direct current component representing the average density of the entire image can also be taken in as a determination element, and edge block determination depending on the density of the entire image can be performed, thereby improving edge determination accuracy. In particular, a thin character on a highlight is easily determined as an edge block, and the character reproducibility is improved.

  A computer program according to the present invention is a computer program that causes a computer to perform orthogonal transformation on image data to generate data composed of a plurality of spatial frequency components, and to perform image processing based on the generated data. Detecting a feature of the image based on the image data from the data composed of the plurality of spatial frequency components; and causing the computer to perform processing according to the detection result of the step on data belonging to a predetermined spatial frequency region. And a step of causing the computer to perform inverse orthogonal transform on the data obtained by performing the above processing.

  In the present invention, a step of detecting image features from data composed of a plurality of spatial frequency components is provided, and processing corresponding to the detection result is performed on data belonging to a predetermined spatial frequency range. Therefore, it is not necessary to perform processing for the entire spatial frequency, and the parameters used for the processing are reduced. As a result, the circuit scale is suppressed. Further, when the processing to be performed is edge enhancement processing, the enhancement region is limited, and the appearance of an unnecessary block pattern at the edge boundary portion on the curve can be suppressed. Furthermore, when edge enhancement processing is not necessary, smoothing can be performed by processing using, for example, two-dimensional matrix data, and a smoother image can be obtained.

  In the case of the present invention, a detecting means for detecting image features from data composed of a plurality of spatial frequency components is provided, and processing corresponding to the detection result is performed on data belonging to a predetermined spatial frequency range. . Therefore, it is not necessary to perform processing for the entire spatial frequency, and the parameters used in the processing can be reduced. As a result, the circuit scale can be suppressed. Further, when the processing to be performed is edge enhancement processing, the enhancement region is limited, and the appearance of an unnecessary block pattern at the edge boundary portion on the curve can be suppressed. Further, when edge enhancement processing is not necessary, smoothing can be performed by processing using, for example, two-dimensional matrix data, and a smoother image can be obtained.

  Further, according to the present invention, the feature of the image is detected by determining the magnitude relationship between the magnitude of the spatial frequency component and a preset threshold value. Therefore, for example, it is possible to easily detect whether or not an edge component is included in the target image by comparing a spatial frequency component in a low frequency range with a threshold value.

  Further, according to the present invention, a predetermined calculation is performed using a direct current component and an alternating current component among the spatial frequency components, and a magnitude relationship between a value obtained by the calculation and a preset threshold value is determined, and the feature of the image is determined. To detect. Therefore, a DC component representing the average density of the entire image can also be taken in as a determination element, edge block determination depending on the density of the entire image can be performed, and edge determination accuracy can be improved. In particular, a thin character on a highlight is easily determined as an edge block, which is effective in improving the reproducibility of characters.

  Furthermore, in the case of the present invention, it is detected whether or not the image is an edge image based on the magnitude relationship between the magnitude of the spatial frequency component and a preset threshold value. Therefore, for example, an image that includes a character or a contour portion of a continuous tone image is subjected to processing for enhancing an edge, and an image composed of halftone dots is subjected to smoothing processing to suppress the occurrence of moire. Is possible.

  Furthermore, according to the present invention, edge enhancement processing or smoothing processing is performed according to the characteristics of the image. Therefore, for example, edge emphasis processing that emphasizes edges is applied to images including outlines of characters and continuous tone images, and smoothing processing is applied to images composed of halftone dots to suppress the occurrence of moire. Is possible.

  Furthermore, according to the present invention, edge enhancement processing is performed on edge images, and smoothing processing is performed on non-edge images. Therefore, when edge enhancement is necessary, it is possible to obtain a clean edge curve that suppresses unnecessary block patterns, and when edge enhancement is unnecessary, a smooth image can be obtained by smoothing. .

  Further, according to the present invention, the image forming apparatus further includes area determination means for determining whether the image data belongs to a character area, a halftone dot area, or a photo area. Therefore, the attribute of the image can be comprehensively determined, and appropriate image quality adjustment can be realized for each image by performing processing according to the determination result.

  Further, according to the present invention, edge enhancement processing is applied to data belonging to a character area or a photographic area, and smoothing is performed on data belonging to a halftone dot area or a photographic area. Processing is performed. Therefore, when edge enhancement is necessary, it is possible to obtain a clean edge curve that suppresses unnecessary block patterns, and when edge enhancement is unnecessary, a smooth image can be obtained by smoothing. .

  Further, according to the present invention, when edge enhancement processing is performed, data belonging to a predetermined spatial frequency region is multiplied by a real number larger than 1, and when smoothing processing is performed, division by a real number larger than 1 is performed. I have to. Therefore, it is not necessary to prepare parameters over the entire spatial frequency domain, and the circuit scale can be suppressed. In particular, when edge enhancement processing is performed, an image with more effective edges can be obtained by changing the real number to be multiplied according to the spatial frequency.

  Further, according to the present invention, when processing according to the characteristics of an image is performed, an image is formed on a sheet based on image data obtained by processing data belonging to a predetermined spatial frequency range. Therefore, for example, an image including a clean edge curve in which the appearance of an unnecessary block pattern at the edge boundary portion is suppressed, and a smooth image obtained by smoothing can be formed on the sheet.

Hereinafter, the present invention will be specifically described with reference to the drawings illustrating embodiments thereof.
Embodiment 1 FIG.
FIG. 1 is a block diagram showing an internal configuration of an image processing apparatus according to the present invention. The image processing apparatus according to the present invention includes a control unit 1, an image input unit 2, an image processing unit 3, an image output unit 4, a storage unit 5, and an operation unit 6. The control unit 1 includes a CPU that controls each part of the hardware and a RAM that temporarily stores data necessary for the control. The storage unit 5 is, for example, a non-volatile semiconductor memory, and stores a control program and various data for controlling each unit of hardware. The control unit 1 loads a control program from the storage unit 5 when the apparatus is turned on, and executes the control program while referring to various data in the storage unit 5, whereby the image processing apparatus according to the present invention as a whole. To act as. The operation unit 6 includes various hardware keys for receiving operation instructions from the user.

  The image input unit 2 is a unit that optically reads an image of a document, and includes a light source that irradiates light to a document for reading, an image sensor such as a CCD (Charge Coupled Device), and the like. The image input unit 2 forms a reflected light image from the original set at a predetermined reading position on the image sensor, and outputs RGB data (R: Red, G: Green, B: Blue). The RGB data output from the image input unit 2 is input to the image processing unit 3.

  The image processing unit 3 performs various color adjustment processes on the RGB data output from the image input unit 2 to generate image data for output. The generated image data is output to the image output unit 4. In the present embodiment, CMYK data (C: Cyan, M: Magenta, Y: Yellow, K: Black) is generated as image data for output. The internal configuration and operation of the image processing unit 3 will be described in detail later.

  The image output unit 4 is a unit that forms an image on a sheet such as paper or an OHP film based on the image data output by the image processing unit 3. Therefore, the image output unit 4 is a charger that charges the photosensitive drum to a predetermined potential, and laser writing that generates an electrostatic latent image on the photosensitive drum by emitting laser light according to image data received from the outside. An apparatus, a developing device that supplies toner to the electrostatic latent image formed on the surface of the photosensitive drum to make it visible, a transfer device that transfers the toner image formed on the surface of the photosensitive drum onto paper (not shown), etc. And an image desired by the user is formed on a sheet by electrophotography. In addition to image formation by an electrophotographic method using a laser writing apparatus, an image formation may be performed by an inkjet method, a thermal transfer method, a sublimation method, or the like.

  In the present embodiment, the image input unit 2 is configured as an image reading unit and the image output unit 4 is configured as an image forming unit. However, as a unit for receiving image data from the outside and a unit for transmitting image data to the outside. You may make it comprise, and the structure which combined them suitably may be sufficient. For example, when the image input unit 2 is an image reading unit and the image output unit 4 is an image data transmission unit, the present invention operates as a scanner device (image reading device). When the data receiving unit and the image output unit 4 are image forming units, the present invention operates as a printer device (image forming device). Needless to say, the present invention can also be applied to a digital multi-function peripheral including an image reading unit, an image forming unit, and an image data transmitting / receiving unit.

  FIG. 2 is a block diagram illustrating the configuration of the image processing unit 3. The image processing unit 3 includes an AD conversion unit 31, a shading correction unit 32, an input gradation correction unit 33, a region separation processing unit 34, a color correction unit 35, a black generation / under color removal unit 36, an image quality adjustment unit 37, an output floor A tone correction unit 38 and a gradation reproduction processing unit 39 are provided.

  The AD conversion unit 31 converts the RGB data input from the image input unit 2 into digital format data. The shading correction unit 32 performs processing for removing various distortions generated in the illumination system, the imaging system, and the imaging system of the image input unit 2 on the digital RGB data output from the AD conversion unit 31. In the present embodiment, AD conversion and shading correction are performed inside the image processing unit 3, but digital RGB data obtained by performing AD conversion and shading correction may be received from the outside. Of course, in that case, the AD conversion unit 31 and the shading correction unit 32 need not be mounted.

  The input tone correction unit 33 performs a process of adjusting the color balance of the RGB data received from the shading correction unit 32, converts the density data into a density signal that can be easily processed by the image processing apparatus, and outputs the density signal to the subsequent region separation processing unit 34. To do.

  The region separation processing unit 34 separates each pixel in the input image into one of a character region, a halftone dot region, and a photographic region based on RGB data. Based on the separation result, the region separation processing unit 34 outputs region identification information indicating which region each pixel belongs to to the black generation / under color removal unit 36, the image quality adjustment unit 37, and the gradation reproduction processing unit 39. In addition, the RGB data output from the input tone correction unit 33 is output to the subsequent color correction unit 35 as it is. A known method can be used as the method of region separation. For example, in “Image Electronics Society Proceedings 90-60-04”, the halftone dot region and the non-halftone dot region are utilized by utilizing the fact that the fluctuation of the image signal is large in the halftone dot region and the density is higher than the background. It is described that can be separated. Further, it is described that the character area and the photographic area can be separated by utilizing the fact that the difference between the maximum signal level and the minimum signal level is large in the character area and the density in the small area is high.

  The color correction unit 35 converts the RGB data output from the region separation processing unit 34 into CMY data, and color turbidity based on the spectral characteristics of the CMY color material including unnecessary absorption components for realizing faithful color reproduction. The process which removes is performed. After being converted into CMY data, the data subjected to the process of removing color turbidity is output to the black generation / under color removal unit 36.

  The black generation / under color removal unit 36 generates black (K) data from the CMY three-color data after color correction, and subtracts K data obtained by black generation from the original CMY data to obtain new CMY data. Is generated, and CMY three-color data is converted into CMYK four-color data. As an example of black generation processing, a method of generating black using skeleton black is known. In this method, the input characteristic of the skeleton curve is y = f (x), the input data is C, M, Y, the output data is C ′, M ′, Y ′, K ′, and the undercolor removal rate is If α (0 <α <1), the black generation / under color removal processing is K ′ = f {min (C, M, Y)}, C ′ = C−αK ′, M ′ = M−αK ′. , Y ′ = Y−αK ′.

  The image quality adjustment unit 37 performs spatial filter processing using a digital filter on the CMYK data input from the black generation / undercolor removal unit 36 based on the region identification information, and corrects the spatial frequency characteristics to blur the output image. The processing for preventing the deterioration of graininess and the like is performed. In particular, the present invention relates to the image quality adjustment unit 37. When the attribute of the image is determined for each block having a predetermined number of pixels and it is determined that the image belongs to the character region and the edge image, the black character or the color character is determined. Edge enhancement processing is performed to improve the reproducibility of the image. Further, when it is determined that the image belongs to a halftone dot region and a non-edge image, smoothing processing is performed in order to suppress the occurrence of moire or the like.

  The output gradation correction unit 38 performs an output gradation correction process for converting a signal such as a density signal into a halftone dot area ratio corresponding to the characteristics of the output destination.

  The gradation reproduction processing unit 39 performs gradation reproduction processing that finally separates an image into pixels and performs processing so that each gradation can be reproduced. For example, for a region separated into photographs by the region separation processing unit 34, binarization or multi-value processing is performed on the screen with an emphasis on gradation reproducibility, and the region separation processing unit 34 separates the characters into characters. The binarization or multi-value processing is performed on a high-resolution screen suitable for high-frequency reproducibility.

  Details of the image quality adjustment unit 37 will be described below. FIG. 3 is a block diagram showing the internal configuration of the image quality adjustment unit 37. The image quality adjustment unit 37 includes a frequency conversion unit 371, a block determination unit 372, an arithmetic processing unit 373, and an inverse frequency conversion unit 374, and image data (for example, output from the black generation / undercolor removal unit 36 in the previous stage) 256-gradation image data Pi (X, Y)) is converted into a spatial frequency component, processed according to the determination result of the spatial frequency component, and the image data (for example, 256) obtained by reconverting these data into a density region. The gradation image data Po (X, Y)) is output. Here, as shown in the conceptual diagram of FIG. 4, the image data Pi (X, Y) is configured by pixels arranged in a two-dimensional matrix in the X direction (right direction) and the Y direction (down direction). Among the pixel data, the pixel data at the Xth pixel position on the Yth line is represented. In the present embodiment, block unit image data composed of 8 × 8 pixels is input to the frequency conversion unit 371, and each process to be described later is performed in this block unit.

  The frequency conversion unit 371 converts the image data input in units of blocks into a frequency domain. In the present embodiment, a transform to the frequency domain (hereinafter referred to as a DCT transform) is performed using a discrete cosine transform (DCT), which is one of orthogonal transforms. In the present embodiment, as shown in FIG. 4, DCT conversion is performed in block units in the right direction (X direction) from the block including the upper left pixel on the two-dimensional image, and the line is changed in block units. However, DCT conversion is finally performed up to the final block including the lowermost right pixel. The expression of DCT conversion can be expressed as the following expression when the input image is Aij, the output image is Bij, and the size of the input image Aij is M rows and N columns (M and N are positive integers).

  The frequency conversion unit 371 receives image data having 8 × 8 pixels as one block, performs DCT conversion on the 8 × 8 pixel block, and uses the DCT converted spatial frequency component Qj (S, T) as a block determination unit. 372 and the arithmetic processing unit 373. FIG. 5 is a conceptual diagram showing blocks before and after conversion.

  The block determination unit 372 includes a frequency component determination unit 372a, a block region determination unit 372b, and a processing content determination unit 372c. The block determination unit 372 detects image features for each input block, and based on the detected features. A process to be performed on the block (in this embodiment, an edge enhancement process or a smoothing process) is determined. That is, the block determination unit 372 functions as a determination unit of the image processing apparatus according to the present invention.

  The block component data converted into the spatial frequency component by the frequency conversion unit 371 is input to the frequency component determination unit 372a, and the size of the spatial frequency component at a position near the DC component in each block is determined. . FIG. 6 is a conceptual diagram showing a block after being converted into a spatial frequency component. In the present embodiment, data of a spatial frequency component (DCT coefficient) having a matrix size of 8 × 8 is obtained by performing DCT conversion on image data composed of blocks of 8 × 8 pixels. Of the 8 × 8 elements, the component corresponding to the upper left position is a DC component, and the components corresponding to other positions are AC components. The frequency component determination unit 372a obtains the magnitude of the DCT coefficient representing the frequency component at a position in the vicinity of the direct current component, and determines whether or not at least one is greater than a predetermined threshold value.

  For example, in the edge block determination area surrounded by the thick line in FIG. 6 (that is, an area having a matrix size of 3 × 3), the coordinate of the position of the DC component is (s0, t0), and the spatial frequency component at that position is Q (S0, t0), the coordinates of the right next position are (s0 + 1, t0), the spatial frequency component at that position is Q (s0 + 1, t0), and the coordinates of the right next position are (s0 + 2, t0), The spatial frequency component at that position is defined as Q (s0 + 2, t0), and the spatial frequency components corresponding to the respective positions are sequentially moved down from the left side by Q (s0, t0 + 1), Q (s0 + 1, t0 + 1), Q (s0 + 2, t0 + 1), and the spatial frequency components corresponding to the respective positions in order from the left side further down one line are Q (s0, t0 + 2), Q (s0 + 1, t0 + 2), Q (s0 + 2, t0 + 2), and so on. If the determination of whether a block (edge blocks) including an edge, (the α positive constant) threshold α a predetermined performed by the following equation was used.

  | Q (s0 + n, t0 + m) |> α, where n, m = 0, 1, 2, and n = m ≠ 0

  Processing is performed when there is a set of n and m satisfying the above-described determination formula, that is, when any of the spatial frequency components (AC components) other than the DC component is determined to be larger than the threshold value α. It is determined that the target block (hereinafter referred to as a processing block) is an edge block. Further, when it is determined that any of the AC components is equal to or less than the threshold value α, it is determined that the processing block is a non-edge block. That is, the frequency component determination unit 372a functions as a detection unit that detects image characteristics (whether or not an edge component is included), and the determination result by the frequency component determination unit 372a is output to the processing content determination unit 372c.

  Here, the reason why it can be determined whether or not an edge is included in the block by the above-described method will be described. If a higher luminance is assigned to a natural image having a larger value in each block after DCT conversion, the luminance of the component corresponding to the low frequency side (upper left part of each block) is determined from the result after execution. Get higher. That is, the low frequency component has a large amount of information. As for this, it is known that the signal power concentrates in a low frequency region when the signal having a strong correlation is viewed in frequency. This means that there is a coefficient in which power is concentrated in each DCT coefficient and a coefficient in which power is not concentrated. In other words, a large deviation occurs in the size of each DCT coefficient. When DCT conversion is performed on a certain natural image and the DCT coefficient is seen, it can be seen that the DCT coefficient of the DC component has a particularly large value, and the value of the low frequency component is also large in the DCT coefficient of the AC component.

  On the other hand, when DCT conversion is performed on a solid image (an image having no edge component and a uniform density), only the DC component has a value, and the low frequency components around the DC component are zero. On the other hand, in the case of an image having an edge component or a feature amount (such as a natural image) having a density, the low frequency component around the direct current component also shows a large value. Depending on the value, it can be used as a material for determining whether or not the original image has a feature value. Further, when an edge component is conspicuous in the main scanning direction or an edge component is conspicuous in the sub-scanning direction, the value of the alternating current component to the right of the direct current component or the direct current alternating current component is affected and increases. For the above reasons, it is possible to determine whether or not an edge component is included by examining the magnitude of the absolute value of the low frequency component.

  Note that the threshold α used in the above-described determination formula can be arbitrarily set. For example, when the setting value of the threshold value α is small, it is easy to determine that the processing target block is an edge block, and when the setting value of the threshold value α is large, it is difficult to determine that the processing block is an edge block. Here, α = 64 can be set as an example. By setting α, it can be used as a material for determining whether an image in the density region includes an edge component for each processing block. Further, the threshold value α may be set according to the determination result of the block area determination unit 372b. For example, a threshold value α1 (for example, α1 = 32) that can be easily determined to be an edge block is set in the case of a character area, and an edge block is easily set in the case of a dot area. It is also possible to set a threshold value α2 (for example, α2 = 96) that is not determined as follows. Furthermore, the configuration may be such that the user himself sets the threshold value α through the operation unit 6 in accordance with the input image.

  Further, although the matrix size of the edge block determination area is 3 × 3 pixels in the present embodiment, it may be expanded to a wider range in the low frequency area if the hardware condition permits.

  The block region determination unit 372b receives region identification information for each pixel from the region separation processing unit 34, and determines a region for the entire block. The block area determination unit 372b determines the area of the entire block based on the area identification information determined in units of pixels, the pixels belonging to the character area, the pixels belonging to the halftone dot area, and the photograph within the processing block (8 × 8 pixels) Three types of pixels belonging to the region are counted, their ratios are obtained, and the entire block is determined based on the result.

  In the present embodiment, the region separation processing unit 34 separates each pixel into a character region, a halftone dot region, and a photographic region, and based on the region separation information output from the region separation processing unit 34, a block region determination unit Although the region separation in block units is performed in 372b, the region separation processing unit 34 may perform region separation in block units, and the separation result may be input to the processing content determination unit 372c. In this case, it is needless to say that the image quality adjustment unit 37 does not need to include the block area determination unit 372b.

  FIG. 7 is a table summarizing the relationship between the determination criteria used in the block area determination unit 372b and the determination results. The block area determination unit 372b determines how many pixels determined to belong to the character area, pixels determined to belong to the halftone dot area, and pixels determined to belong to the photograph area among the 64 pixels constituting the processing block. It is determined whether it is included in the ratio of. Specifically, when there are three regions, region A, region B, and region C (however, region A, region B, and region C correspond to any of a character region, a halftone dot region, and a photo region). If the area A occupies 60% of the processing block, it is determined that the processing block is an image belonging to the area A, and the ratio of each area A, B, C is less than 60%. If there is, it is determined that the processing block is another image. Further, when there are two areas, area A and area B, and area A occupies 70% of the processing block, it is determined that the processing block is an image belonging to area A. If the ratio of the areas A and B is less than 70%, it is determined that the processing block is another image. Further, when only the area A exists, it is determined that the processing block is an image belonging to the area A.

  The above-described determination criterion is an example for determining the image region to which the processing block belongs, and an appropriate value may be set so that a preferable image can be obtained.

  As described above, in the block area determination unit 372b as the area determination means, when many determination pixels calculated in the processing block are biased, the area which is determined to be biased is determined as a determination result. If the two areas exist at the same rate, it is determined not to belong to one area, but is determined to be another area, and the determination result is output to the processing content determination unit 372c in the subsequent stage.

  The processing content determination unit 372c performs comprehensive determination by combining the determination result by the frequency component determination unit 372a and the determination result by the block region determination unit 372b, and determines the processing content in the arithmetic processing unit 373 described later. FIG. 8 is a table summarizing the processing contents determined by the combination of the determination result by the frequency component determination unit 372a and the determination result by the block region determination unit 372b. As determined in this chart, when the entire block is determined to be a character region by region determination and the frequency component determination result indicates an edge block, the calculation processing unit 373 executes the determination. The power processing is determined as edge enhancement processing. Further, when it is determined by the area determination that the entire block is a character area, and the determination result of the frequency component indicates a non-edge block, processing to be performed on the spatial frequency component Qj (S, T) Is not selected, and the arithmetic processing unit 373 performs processing for sending the data of the spatial frequency component Qj (S, T) to the inverse frequency conversion unit 374 as it is.

  When the area determination result of the entire block is a character area and the frequency component determination result is a non-edge block, it is erroneously determined as a character area that includes not only the edge portion outside the character but also the inside of the character. Including cases. That is, when a character (for example, a character with a width of about 10 pixels) thicker than the block size (8 × 8 pixels) used in the present embodiment is included, the inside of the character region is also a character region. Although erroneous determination may occur, erroneous determination can be prevented by performing edge determination with frequency components.

  In addition, when it is determined by the region determination that the entire block is a halftone dot region and the determination result of the frequency component indicates a non-edge block, the processing to be executed by the arithmetic processing unit 373 is smoothed. Is determined. If it is determined by the region determination that the region is a halftone dot region and the determination result of the frequency component indicates an edge block, the processing to be performed on the spatial frequency component Qj (S, T) is not selected, The arithmetic processing unit 373 performs processing for sending the data of the spatial frequency component Qj (S, T) to the inverse frequency conversion unit 374 as it is.

  Since the halftone area does not include low frequency components, the frequency component determination result usually indicates a non-edge block. However, if there is a character in a halftone dot area, such as a character on the map, or a character composed of halftone dots, such as a halftone dot character or a color character, the area is determined to be a halftone dot area. Then, the frequency component determination result can be determined as an edge block. In this case, if the edge emphasis process is performed, moire may occur, and if the smoothing process is performed, the character portion may be blurred. Therefore, it is desirable not to perform these processes. As described above, by making a comprehensive determination using the determination result of the region determination and the determination result of the frequency component, it is possible to increase the determination accuracy and execute an optimal process.

  Further, when it is determined by the area determination that the entire block is a photographic area, and the determination result of the frequency component indicates an edge block, the process to be executed by the arithmetic processing unit 373 is determined as the edge enhancement process. To do. If it is determined by region determination that the entire block is a photographic region, and the determination result of the frequency component indicates a non-edge block, the processing to be executed by the arithmetic processing unit 373 is determined as smoothing processing. .

  Furthermore, when it is determined by the area determination that the entire block does not correspond to any of the above-described areas, that is, when it is determined that it is another area, processing to be applied to the spatial frequency component Qj (S, T) is selected. Instead, the arithmetic processing unit 373 performs processing for sending the data of the spatial frequency component Qj (S, T) to the inverse frequency conversion unit 374 as it is.

  For example, if the text area and halftone dot area are included to the same extent in the block, the edge emphasis process will also emphasize the halftone dot area, and there will be a sense of incongruity at the boundary with the smoothed dot area. Occurs. On the other hand, when the smoothing process is performed, a sense of incongruity at the boundary portion occurs with respect to the character area. Therefore, in such a case, image quality deterioration can be prevented by not performing both edge enhancement processing and smoothing processing.

  The arithmetic processing unit 373 receives the determination result output from the block determination unit 372, performs edge enhancement processing or smoothing processing, or data of the spatial frequency component Qj (S, T) output from the frequency conversion unit 371. Is sent to the inverse frequency converter 374 as it is.

  First, the case where the edge enhancement process is performed by the arithmetic processing unit 373 will be described. FIG. 9 is an explanatory diagram for explaining a region to which edge enhancement processing is performed. In the present embodiment, arithmetic processing is performed on a DCT coefficient having a component only in the vertical direction, a DCT coefficient having a component only in the horizontal direction, and a DCT coefficient corresponding to the position of the intersection of the 8 × 8 pixel processing blocks. As a result, the edge enhancement process is executed. For example, the area on which the calculation process is performed is set in advance in the storage unit 5.

  In the arithmetic processing in the case of performing edge emphasis processing, a DCT coefficient Qk (() after conversion is obtained by multiplying the DCT coefficient Qj (S, T) by a value obtained by adding the row number and column number of the processing block and a predetermined constant. S, T). The conceptual diagram of FIG. 10 shows a value obtained by adding the row number and column number of the processing block in the area where the arithmetic processing is performed. The DCT coefficient Qk (S, T) can be expressed as follows in the region where the arithmetic processing is performed.

  Qk (S, T) = Qj (S, T) × (added value of row number and column number) × (constant)

  In the region where the arithmetic processing is not performed, the input Qj (S, T) is used as the value after the arithmetic processing as it is. Here, for example, 0.35 can be used as the constant of the above-described equation. However, this value is only an example and may be a value that is emphasized by a small value. In addition, since it is emphasized by this value, it may be determined while evaluating the image quality using an actual output image in consideration of the overall balance such as whether or not the edge is overemphasized. Even when such processing is performed, since the DC component is not converted, the average density of the entire block is maintained.

  Next, the case where the arithmetic processing unit 373 performs the smoothing process will be described. FIG. 11 is an explanatory diagram for explaining a region on which the smoothing process is performed. In the present embodiment, among the 8 × 8 pixel processing blocks, each DCT coefficient is subjected to arithmetic processing in a region excluding the DC component and its periphery, and smoothing processing is performed. More specifically, arithmetic processing is performed on the DCT coefficient Qj (S, T) corresponding to the position satisfying 3 <S + T ≦ 14 in the DCT coefficient Qj (S, T). For example, the area on which the calculation process is performed is set in advance in the storage unit 5.

  FIG. 12 is a conceptual diagram showing an example of two-dimensional matrix data used in the smoothing process. The matrix data shown in FIG. 12 has a value M (S, T) corresponding to each of 8 × 8 DCT coefficients, the value of M (S, T) in 0 ≦ S + T ≦ 3 is 1, and 3 < M (S, T) in S + T ≦ 14 takes a value of 1 or more.

  The two-dimensional matrix data shown in FIG. 12 generally has a contrast sensitivity function (hereinafter referred to as CSF: CSF: Contrast Sensitive Function) characteristic, that is, a characteristic reflecting human visual characteristics. In general, the sensitivity to human contrast depends on the spatial frequency, and the human perception system is considered as a kind of band-pass filter. For example, when considering a black and white striped pattern, the sensitivity to a human striped pattern varies depending on the interval between successive striped patterns. When the interval between the stripes is very small, it becomes difficult for a human to perceive the stripe pattern. The value of M (S, T) is, for example, a value that changes concentrically according to the sensitivity to human contrast, centering on the frequency component indicated by hatching in FIG.

  When the arithmetic processing unit 373 performs arithmetic processing on the DCT coefficient Qj (S, T) and performs smoothing processing, the arithmetic processing unit 373 corresponds the frequency component data Sj (S, T) to the two-dimensional matrix data for each unit block. The DCT coefficient Qk (S, T) after the calculation is obtained by dividing by the element M (S, T). That is, the smoothing process is realized by executing the following calculation.

  Qk (S, T) = Sj (S, T) / M (S, T)

  When the DCT coefficient Qj (S, T) is divided by CSF, the DCT coefficient of the frequency component having high sensitivity to contrast is divided by a larger value than the frequency component having low sensitivity to contrast, and thus smoothed effectively. It will be. Note that Qk (S, T) may be obtained up to the decimal point.

According to the present embodiment, when a unit block does not have an edge portion, smoothing processing is performed using a two-dimensional matrix so that the unit block is smooth, and when a unit block has an edge portion, an edge The part is emphasized. Therefore, in the entire image, the edge portion is emphasized, and a flat image portion having no edge portion can obtain a smoother image,
The characteristic part of the original image can be kept good.

  Note that the two-dimensional matrix data used for the smoothing process is not limited to that shown in FIG. 12, and a small value is set for a frequency component having low sensitivity to human contrast, and a large value is set for a frequency component having high sensitivity to contrast. A set two-dimensional matrix, in other words, a low-frequency component may be maintained, and a two-dimensional matrix in which the outer DCT coefficients are more strongly suppressed concentrically within an 8 × 8 pixel block may be used. For example, two-dimensional matrix data having a Gaussian distribution can be used.

  In the embodiment, the frequency component of 3 <S + T ≦ 14 is changed. However, the present invention is not limited to this, and the frequency component of 1 <S + T ≦ 14 may be changed.

  When the block determination unit 372 determines that neither the edge enhancement process nor the smoothing process described above is performed, the DCT coefficient Qj (S, T) is not subjected to the calculation process, and the subsequent frequency conversion unit 374 is performed. Output data.

  The inverse frequency conversion unit 374 performs inverse frequency conversion on the DCT coefficient Qk (S, T) output from the arithmetic processing unit 373 and converts it to density region data (image data). Inverse frequency transformation is performed using the inverse transformation formula of the two-dimensional DCT transformation. In the input image, blocks that contain edge components are edge-enhanced, so a sharper and clearer image can be obtained. For images that do not contain edge components, smoothing that takes into account visual characteristics Processing is performed. Since the emphasis process does not emphasize the entire frequency range other than the DC component, but limits the emphasis region, it suppresses the appearance of unnecessary block patterns at the edge boundary on the curve and creates a clean edge curve. Obtainable.

Embodiment 2. FIG.
In the first embodiment, the frequency component determination unit 372a checks the size of the spatial frequency component around the DC component and determines whether or not the processing block is an edge block, but represents the average density of the entire block. The magnitude of the DC component may be added to the determination element. For example, when the block to be processed includes many edge components, the components around the DC component usually have a large value. Therefore, in determining the edge block, the magnitudes of the values around the DC component are examined, and if these values are equal to or greater than a predetermined value, it can be determined that the block is an edge block. In addition, the edge portion of the image has a difference in recognition due to the density difference from the background. For example, there is a case where a thin character or the like on a highlight is recognized as a character so that it can be clearly seen. In this case, since it is a thin character, it is difficult to determine it as an edge block, the edge emphasis process cannot be performed, and it may be difficult to see the character as a character. Therefore, in the present embodiment, determination is performed in consideration of the average density of the entire block so that even a thin character on the highlight can be determined as an edge block.

  The hardware configuration of the image processing apparatus in this embodiment is the same as that shown in the first embodiment. That is, the block determination unit 372 in the image quality adjustment unit 37 includes a frequency component determination unit 372a, a block region determination unit 372b, and a processing content determination unit 372c. The feature of the image for each block to be finally processed is detected based on the determination result by these determination units, and the processing to be performed on the block based on the detected feature (in this embodiment, edge enhancement processing) Or smoothing processing). Note that when the area identification information is not used, the block area determination unit 372b may be omitted.

  The block component data converted into the spatial frequency component by the frequency conversion unit 371 is input to the frequency component determination unit 372a. The frequency component determination unit 372a obtains the magnitude of the DCT coefficient representing the frequency component at a position near the DC component in each block, multiplies these values by the value of the DC component of the processing block, and at least one of them is It is determined whether or not it is larger than a predetermined threshold value β1 (β1 is a positive integer).

  As in the first embodiment, when the data input to the frequency component determination unit 372a is data composed of 8 × 8 pixel blocks, and the 3 × 3 pixel region including the DC component is the edge block determination region, the edge block Is determined by the following equation.

| Q (s0 + n, t0 + m) | × | Q (s0, t0) |> β1
Where n, m = 0, 1, 2, and n = m ≠ 0

  In this determination formula, Q (s0, t0) represents a direct current component in the block, and Q (s0 + n, t0 + m) represents an alternating current component at coordinates (s0 + n, t0 + m). When there is a set of n and m satisfying the determination formula, that is, any one of the values obtained by multiplying the spatial frequency component (AC component) other than the DC component by the size of the DC component is larger than the threshold value β1. If it is determined that the block to be processed (processing block) is an edge block. If it is determined that any of the AC components multiplied by the DC component is equal to or less than the threshold value β1, it is determined that the block is not an edge block. That is, the frequency component determination unit 372a functions as a calculation unit that performs a predetermined calculation using a DC component and an AC component, and also functions as a detection unit that detects image characteristics (whether an edge component is included). .

  By using a value obtained by multiplying the value of a predetermined spatial frequency component by a DC component, a value to be compared is emphasized in a block having a large DC component, and it is easier to determine an edge block. On the other hand, in a block having a small DC component, the value to be compared is not emphasized so much that it is difficult to determine that the block is an edge block.

  For example, when the image data is represented by 8 bits, that is, 256 gradations and 255 represents the white of the image, the direct current component has a relatively large value in the whitish image. In a block in which a thin character is written on the highlight, although there is an edge component, the density difference between the highlight portion of the background and the thin character is small. In general, it is difficult to determine an edge block by only using the determination. However, by using a value obtained by multiplying the DC component representing the average density of the entire block for the determination, the DC component is low because the average density of the processing block is low even if the density difference between the highlight portion of the background and the light character is small. If the threshold value is determined using a value obtained by multiplying the predetermined spatial frequency component by this direct current component, the threshold value to be referred to is finally reduced, and the edge becomes more edged. It becomes easy to be determined as a block. As a result, processing such as edge emphasis processing can be performed, and an effect is seen when it is desired to express a thin character on a highlight more clearly.

  Further, when it is determined that the processing block is not an edge block, that is, when it is determined that any of the AC components multiplied by the DC component is equal to or less than the threshold value β1, whether or not the processing block is a non-edge block. May be determined. For example, when the magnitude of the AC component near the DC component is compared with the threshold value β2 (β2 is a positive integer) as in the first embodiment, and the magnitude of the AC component near the DC component is equal to or less than the threshold value β2. When the processing block is determined to be a non-edge block and is larger than the threshold value β2, it can be determined that the processing block is an intermediate block.

  Here, the intermediate block represents a block that does not belong to an edge block or a non-edge block in the block determination. In the above-described determination formula, whether or not the block is an edge block is determined by the size of the threshold. Therefore, there is no problem with blocks that are clearly judged as edge blocks and non-edge blocks, but they are just blocks in between, that is, they become edge blocks or non-edge blocks depending on the threshold setting. Since such a block is sensitively affected by the threshold value, if one of the blocks is determined and subjected to enhancement or smoothing processing, the image quality may be adversely affected. For this reason, a block that is not an edge block or a non-edge block that is difficult to determine is determined to be an intermediate block according to the criteria described above, and neither process is performed as a whole block. Is provided.

  As a specific example of setting the threshold value, the threshold value β1 used to determine whether or not the edge block is a DC component is 34000, and only the spatial frequency component near the DC component is used and the non-edge block is not used. The threshold value β2 used for determining whether or not is 16. Note that these values are merely examples, and appropriate values may be set so that a preferable image can be obtained.

  The determination result by the frequency component determination unit 372a is output to the processing content determination unit 372c. The processing content determination unit 372c performs overall determination by combining the determination result by the frequency component determination unit 372a and the determination result by the block region determination unit 372b, and determines processing to be performed on the entire processing block. FIG. 13 is a chart summarizing the processing contents determined by the combination of the determination result by the frequency component determination unit 372a and the determination result by the block region determination unit 372b. As defined in this chart, the processing block determined to be an intermediate block by the frequency component determination, and the processing determined not to belong to any of the character area, the dot area, and the photograph area by the area determination Neither edge emphasis processing nor smoothing processing is performed for the block. In addition, when the entire block is determined to be a character region by region determination and the frequency component determination result indicates an edge block, it is specified that edge enhancement processing is performed on the entire block. Further, when it is determined by the region determination that the entire block is a halftone dot region, and the determination result of the frequency component indicates a non-edge block, it is prescribed that the entire block is subjected to smoothing processing, When it is determined that the entire block is a photographic region by region determination, and the frequency component determination result indicates an edge block or non-edge block, edge enhancement processing and smoothing processing are performed on the entire block, respectively. It is prescribed.

  In this way, in block determination using frequency data, by providing intermediate block determination in addition to edge blocks and non-edge blocks, neither processing that requires enhancement processing nor block that requires smoothing processing is required. Blocks can be selected, and processing that captures image features becomes possible.

Embodiment 3 FIG.
The first embodiment has been described as an image processing apparatus that implements each process by hardware. However, the image processing method of the present invention may be implemented by software processing.

  FIG. 14 is a block diagram illustrating an internal configuration of an image processing apparatus in which a computer program according to the present invention is installed. In the figure, reference numeral 10 denotes an image processing apparatus according to the present embodiment, specifically a personal computer, a workstation, or the like. The image processing apparatus 10 includes a CPU 11. The CPU 11 loads various control programs from the ROM 13 connected to the bus 12 to the RAM 14 and executes them. The operation unit 15, auxiliary storage unit 16, storage unit 17, image input The hardware such as the IF 18 and the image output IF 19 is controlled.

  The operation unit 15 includes a keyboard, a mouse, and the like in order to select image data to be processed, input parameters necessary for image processing, an instruction to start image processing, and the like. The auxiliary storage unit 16 includes a reading device for reading the computer program from the recording medium 20 on which the computer program of the present invention is recorded. As the recording medium 20, an FD (Flexible Disk), a CD-ROM, or the like can be used. The storage unit 17 includes a hard disk device having a disk-shaped magnetic recording medium, and stores a computer program read through the auxiliary storage unit 16, image data input through the image input IF 18, and the like.

  The image input IF 18 is an interface for connecting the image input device 21. An interface such as SCSI or USB can be used as the image input IF 18, and the image input device 21 connected to the image input IF 18 is a scanner device, a digital still camera, or the like. The image output IF 19 is an interface for connecting the image output device 22. An interface such as a USB can be used as the image output IF 19, and the image output device 22 connected to the image output IF 19 is a printer device or the like.

  As the recording medium 20 for recording the computer program according to the present invention, in addition to the above-mentioned FD and CD-ROM, optical recording medium such as MO, MD, DVD, magnetic recording medium such as hard disk, IC card, memory card Further, it is also possible to use a semiconductor memory such as a card-type recording medium such as an optical card, a mask ROM, an EPROM (Erasable Programmable Read Only Memory), an EEPROM (Electrically Erasable Programmable Read Only Memory), a flash ROM, or the like.

  Hereinafter, the operation of the image processing apparatus 10 according to the present embodiment will be described. FIG. 15 is a flowchart for explaining a procedure of processing executed by the image processing apparatus 10. First, the CPU 11 of the image processing apparatus 10 executes a region separation process for each pixel of the image data input through the image input IF 18 (step S11). In the region separation process, it is determined whether each pixel belongs to a character region, a halftone dot region, or a photograph region, and the determination result is output as region separation information.

  Next, the CPU 11 reads a processing target block (processing block) from the input image data (step S12). This processing block is constituted by, for example, a set of 8 × 8 pixel data, and a two-dimensional image is subjected to reading processing in units of blocks from the block including the upper left corner pixel in the right direction. While changing the line, the reading process is finally performed up to the final block including the pixel in the lower right corner.

  Next, the CPU 11 performs DCT processing on the read processing block (step S13). In this DCT process, discrete cosine transform (DCT transform), which is one of orthogonal transforms, is performed on image data read in units of blocks, and converted to frequency domain data.

  Next, the CPU 11 executes DCT coefficient determination processing based on the frequency domain data acquired in step S13 (step S14). In this DCT coefficient determination process, a block (edge block) including an edge is determined by determining whether or not the size of the spatial frequency component in the vicinity of the DC component is larger than a preset threshold value α in the processing block. Whether or not is detected. FIG. 16 is a flowchart for explaining the procedure of DCT coefficient determination processing. First, the CPU 11 determines whether or not the magnitude of the predetermined spatial frequency component is larger than a predetermined threshold value α (step S141). The determination need not be performed for the entire processing block, but may be performed for several spatial frequency components in the vicinity of the DC component, as in the first embodiment. Further, the threshold value α may be described in the computer program as a parameter used in the above-described computer program.

  When it is determined that the size of the predetermined spatial frequency component is larger than the threshold value α (S141: YES), it is determined that the processing block is an edge block (step S142), and the CPU 11 performs the process shown in the flowchart of FIG. Return to step S14. If it is determined that the magnitudes of the predetermined spatial frequency components are all equal to or less than the threshold value α (S141: NO), it is determined that the processing block is a non-edge block (step S143), and the CPU 11 performs the process. Return to step S14 in the flowchart shown in FIG.

  Next, the CPU 11 executes a block area determination process based on the area identification information acquired in step S11 (step S15). That is, for each processing block, three types of pixels belonging to the character area, pixels belonging to the halftone dot area, and pixels belonging to the photographic area are counted, and the respective ratios are obtained. Alternatively, it is determined which block belongs to which other area. FIG. 17 is a flowchart for explaining the procedure of the block area determination process. First, the CPU 11 counts the number of pixels belonging to each area for the target processing block, and determines whether there are three areas, a character area, a halftone area, and a photograph area (step S151). If it is determined that there are three areas (S151: YES), it is determined whether there is an area that occupies 60% or more of the processing block (step S152). When it is determined that there is an area that occupies 60% or more of the processing block (S152: YES), the area is determined as the processing block area (step S153). That is, when pixels belonging to the character area occupy 60% or more of the processing block, it is determined that the processing block is a block belonging to the character area. The same applies when the pixels belonging to the halftone dot region and the pixels belonging to the photo region occupy 60% or more. On the other hand, when it is determined that there is no region that occupies 60% or more of the processing block (S152: NO), the processing block is determined as another region (step S154).

If it is determined that three areas do not exist in the target processing block (S151: NO), it is determined whether two areas of the character area, the halftone dot area, and the photograph area exist (step S155). If it is determined that there are two areas (S155: YES), it is determined whether one of those areas occupies 70% or more of the processing block (step S156). When it is determined that there is an area that occupies 70% or more (S156: YES), the area is determined as an area of the processing block (step S157). On the other hand, when it is determined that there is no area that occupies 70% or more (S156). : NO), the processing block is determined as another area (S154). If it is determined in step S155 that two areas do not exist (S155: NO), that is, if it is determined that only one area exists in the processing block, the existing area is determined as the area of the processing block. (Step S158).
After determining the area of the processing block, the CPU 11 returns the process to step S15 of the flowchart shown in FIG.

  Next, the CPU 11 determines whether or not the processing block belongs to the character area based on the determination result of the block area determination processing (step S16). If it is determined that it belongs to the character region (S16: YES), the CPU 11 determines whether or not the processing block is an edge block based on the determination result of the DCT coefficient determination processing (step S17). If it is determined that the processing block is an edge block (S17: YES), the CPU 11 executes edge enhancement processing on a predetermined area in the processing block (step S18). That is, when the processing block belongs to the character area and is an edge block, edge enhancement processing is performed on the processing block. Then, the CPU 11 performs an inverse DCT process (step S19), and converts the frequency domain data into density domain data.

  If it is determined in step S16 that the processing block does not belong to the character area (S16: NO), the CPU 11 determines whether or not the processing block belongs to the halftone area based on the determination result of the block area determination processing ( Step S20). If it is determined that it belongs to the halftone area (S20: YES), the CPU 11 determines whether or not the processing block is an edge block based on the determination result of the DCT coefficient determination process (step S21). When it is determined that the processing block is not an edge block (S21: NO), the CPU 11 executes a smoothing process on a predetermined area in the processing block (step S24). That is, if the processing block belongs to the halftone dot area and is a non-edge block, the processing block is subjected to smoothing processing. Then, the CPU 11 performs inverse DCT processing (S19), and converts the frequency domain data into density domain data.

  If it is determined in step S20 that the processing block does not belong to the halftone dot area (S20: NO), the CPU 11 determines whether or not the processing block belongs to the photographic area based on the determination result of the block area determination process (step S20: NO). Step S22). If it is determined that the processing block belongs to the photo area (S22: YES), the CPU 11 determines whether or not the processing block is an edge block based on the determination result of the DCT coefficient determination processing (step S23). When it is determined that the processing block is an edge block (S23: YES), the CPU 11 performs edge enhancement processing on a predetermined area in the processing block (S18), and when it is determined that the processing block is not an edge block (S23: NO), a smoothing process is executed for a predetermined area in the processing block (step S24). That is, when the processing block belongs to a photographic area and is an edge block, edge enhancement processing is performed, and when it is a non-edge block, smoothing processing is performed. After executing the edge enhancement process (S18) or the smoothing process (S24), the CPU 11 performs an inverse DCT process (S19), and converts the frequency domain data into the density domain data.

  If it is determined in step S17 that it is not an edge block (S17: NO), if it is determined in step S21 that it is an edge block (S21: YES), it is determined in step S22 that the processing block does not belong to a photographic area. If so (S22: NO), inverse DCT processing is performed without executing both edge enhancement processing and smoothing processing (S19), and frequency domain data is converted to density domain data.

  Next, the CPU 11 determines whether or not the processing has been completed for all blocks (step S25). If the CPU 11 determines that the processing for all blocks has not been completed (S25: NO), the CPU 11 advances the processing to step S12. If it is determined that the processing of all blocks has been completed (S25: YES), the processing by this routine is terminated.

  In the present embodiment, a configuration has been described in which processing for a block is performed based on a region determination result and a frequency component determination result for the entire block. However, a region determination result for a block is not used, and only a frequency component determination result is used. Processing may be performed. In this case, it is effective when the present invention is used as application software. For example, in the case of a photo read by an image input device such as a scanner or an image downloaded via a network, the object is known, so there is no need to perform region separation, and if edge enhancement or smoothing is performed as necessary Good.

Embodiment 4 FIG.
In the third embodiment, the size of the spatial frequency component in the vicinity of the DC component is compared with a threshold value α to determine whether the processing block is an edge block or a non-edge block by software processing. However, it is determined whether or not it is an edge block in consideration of the DC component. Further, without using the DC component, only the spatial frequency component in the vicinity of the DC component is an intermediate block or a non-edge block. It is good also as a structure which determines. That is, in the present embodiment, a configuration for realizing the image processing apparatus described in the second embodiment by software processing will be described. Note that the internal configuration of the image processing apparatus 10 in which the computer program of the present invention is installed is exactly the same as that of the third embodiment, and thus the description thereof is omitted.

  FIG. 18 is a flowchart for explaining a procedure of processing executed by the image processing apparatus 10. The CPU 11 of the image processing apparatus 10 executes processing from region separation processing to DCT processing in the same procedure as in the third embodiment (steps S31 to S33).

  Next, the CPU 11 executes a DCT coefficient determination process based on the frequency domain data acquired by the DCT process (step S34). In this DCT coefficient determination process, the spatial frequency component in the vicinity of the DC component is multiplied by the magnitude of the DC component in the processing block, and the magnitude relationship between the obtained value and the preset threshold value β1 is determined. It is determined whether or not the block is an edge block. Processing blocks that are determined not to be edge blocks are non-edge blocks or intermediate blocks that do not belong to any of them by comparing the magnitude relationship between the size of a predetermined spatial frequency component and a preset threshold value. It is determined whether or not there is. FIG. 19 is a flowchart for explaining the procedure of DCT coefficient determination processing. First, the CPU 11 determines whether or not a value obtained by multiplying a predetermined spatial frequency component by a DC component is larger than a threshold value β1 (step S341). The determination need not be performed for the entire processing block, and may be performed for several spatial frequency components in the vicinity of the DC component, as in the second embodiment. The threshold value β1 may be described as a parameter in the computer program to be used.

If it is determined that the value obtained by multiplying the predetermined spatial frequency component by the DC component is greater than the threshold β1 (S341: YES), the CPU 11 determines that the processing block is an edge block (step S342). When it is determined that the value obtained by multiplying the predetermined spatial frequency component by the DC component is equal to or less than the threshold β1 (S341: NO), it is determined whether the predetermined spatial frequency component in the vicinity of the DC component is greater than the threshold β2. (Step S343). Here, the threshold value β2 may be described as a parameter in the computer program. When it is determined that the magnitudes of the predetermined spatial frequency components are all equal to or smaller than the threshold β2 (S343: NO), the CPU 11 determines that the processing block is a non-edge block (step S345) and is larger than the threshold β2. If it is determined that there is a spatial frequency component (S343: YES), it is determined that the processing block is an intermediate block that does not belong to either an edge block or a non-edge block (step S344).
After determining the processing block based on the DCT coefficient, the CPU 11 returns the processing to the flowchart shown in FIG.

  Next, the CPU 11 executes a block area determination process based on the area identification information obtained by the area separation process in step S31 (step S35). That is, for each processing block, three types of pixels belonging to the character area, pixels belonging to the halftone dot area, and pixels belonging to the photographic area are counted, and the respective ratios are obtained, so that the processing block is the character area, halftone dot area, photographic area, Alternatively, it is determined whether the block belongs to any other area. A specific processing procedure is exactly the same as the flowchart of FIG. 17 described in the third embodiment.

  Next, the CPU 11 determines whether or not the target processing block is an intermediate block that does not belong to either an edge block or a non-edge block based on the determination result of the DCT coefficient determination processing (step S36). When it is determined that the target processing block is an intermediate block (S36: YES), since neither the edge emphasis processing nor the smoothing processing is performed for the processing block, inverse DCT processing is executed (step S40), and the frequency The area data is converted into density area data.

  When determining that the block to be processed is not an intermediate block (S36: NO), the CPU 11 determines whether or not the processing block is a character area based on the determination result of the block area determination process (step S37). When determining that it belongs to the character area (S37: YES), the CPU 11 uses the determination result of the DCT coefficient determination process again to determine whether or not the target processing block is an edge block (step S38). If it is determined that the processing block is an edge block (S38: YES), the CPU 11 executes edge enhancement processing on a predetermined area in the processing block (step S39). That is, when the processing block belongs to the character area and is an edge block, edge enhancement processing is performed on the processing block. After executing the edge enhancement process, the CPU 11 performs an inverse DCT process (S40), and converts the frequency domain data into density domain data.

  If it is determined in step S37 that the processing block does not belong to the character area (S37: NO), the CPU 11 determines whether the processing block belongs to the halftone area based on the determination result of the block area determination processing (step S37). S41). When determining that it belongs to the halftone dot region (S41: YES), the CPU 11 determines whether or not the processing block is an edge block based on the determination result of the DCT coefficient determination processing (step S42). When determining that the processing block is not an edge block (S42: NO), the CPU 11 executes a smoothing process on a predetermined area in the processing block (step S45). That is, if the processing block belongs to the halftone dot area and is a non-edge block, the processing block is subjected to smoothing processing. Then, the CPU 11 performs inverse DCT processing (S40), and converts the frequency domain data into density domain data.

  When it is determined in step S41 that the processing block does not belong to the halftone dot area (S41: NO), the CPU 11 determines whether or not the processing block belongs to the photo area based on the determination result of the block area determination process (step S41). S43). If it is determined that the processing block belongs to the photographic area (S43: YES), the CPU 11 determines whether or not the processing block is an edge block based on the determination result of the DCT coefficient determination processing (step S44). When it is determined that the processing block is an edge block (S44: YES), the CPU 11 performs edge enhancement processing on a predetermined area in the processing block (S39), and when it is determined that the processing block is not an edge block (S44: NO), a smoothing process is executed for a predetermined area in the processing block. That is, when the processing block belongs to a photographic area and is an edge block, edge enhancement processing is performed, and when it is a non-edge block, smoothing processing is performed. After executing the edge enhancement process (S39) or the smoothing process (S45), the CPU 11 performs an inverse DCT process (S40) to convert the frequency domain data into density domain data.

  If it is determined in step S38 that it is not an edge block (S38: NO), if it is determined in step S42 that it is an edge block (S42: YES), it is determined in step S43 that the processing block does not belong to a photographic area. If so (S43: NO), inverse DCT processing is performed without executing both edge enhancement processing and smoothing processing (S40), and frequency domain data is converted to density domain data.

  Next, the CPU 11 determines whether or not the processing has been completed for all blocks (step S46). If the CPU 11 determines that the processing has not been completed for all blocks (S46: NO), the CPU 11 advances the processing to step S32. If it is determined that the processing of all blocks has been completed (S46: YES), the processing by this routine is terminated.

  As described above, in this embodiment, since the direct current component is taken into consideration, the edge block determination depending on the density of the entire block can be performed, and the edge determination accuracy is improved. In addition, in block determination based on frequency data, by providing intermediate block determination in addition to edge blocks and non-edge blocks, blocks that require enhancement processing, blocks that require smoothing processing, and blocks that do not require any processing. Sorting can be performed, and processing that captures the characteristics of the image becomes possible.

Embodiment 5. FIG.
In the fourth embodiment, the processing for the block is determined based on the region determination result and the frequency determination result of the entire block, but should be executed using the determination result of only the frequency component without using the region determination result. It is good also as a structure which determines a process.

  FIG. 20 is a flowchart for explaining a procedure of processing executed by the image processing apparatus 10. First, the CPU 11 of the image processing apparatus 10 reads a processing block from the image data input through the image input IF 18 (step S51). The processing block is constituted by a set of 8 × 8 pixel data, for example. Next, the CPU 11 performs DCT processing on the read processing block (step S52). In this DCT process, discrete cosine transform (DCT transform), which is one of orthogonal transforms, is performed on image data read in units of blocks, and converted to frequency domain data.

  Next, the CPU 11 determines whether or not a value obtained by multiplying a predetermined spatial frequency component in the processing block by a DC component is larger than a threshold value β1 (step S53). The determination need not be performed for the entire processing block, and may be performed for several spatial frequency components in the vicinity of the DC component, as described above. The threshold value β1 may be described as a parameter in the computer program to be used. If it is determined that the value obtained by multiplying the predetermined spatial frequency component by the DC component is greater than the threshold value β1 (S53: YES), it is determined that the processing block is an edge block (step S54), and edge enhancement processing is executed. (Step S55). After performing the edge enhancement process, the CPU 11 executes an inverse DCT process (step S56), and converts the frequency domain data into the density domain data.

  If it is determined that the value obtained by multiplying the DC component by the predetermined spatial frequency component is not more than the threshold value β1 (S53: NO), it is determined whether or not the predetermined spatial frequency component in the vicinity of the DC component is greater than the threshold value β2. Judgment is made (step S58). When it is determined that the magnitudes of the predetermined spatial frequency components are all equal to or smaller than the threshold β2 (S58: NO), the CPU 11 determines that the processing block is a non-edge block (step S60), and the processing block Smoothing processing is executed (step S61). After performing the smoothing process, the CPU 11 executes an inverse DCT process (S56), and converts the frequency domain data into the density domain data. If it is determined that the size of the predetermined spatial frequency component is larger than the threshold β2 (S58: YES), it is determined that the processing block is an intermediate block that does not belong to an edge block or a non-edge block (step S59). ), The inverse DCT process is performed without executing any of the edge enhancement process and the smoothing process (S56), and the frequency domain data is converted into the density domain data.

  After executing the inverse DCT process in step S56, the CPU 11 determines whether or not the process has been completed for all blocks (step S57). When it is determined that the processing for all blocks has not been completed (S57: NO), the CPU 11 returns the processing to step S51, and when it is determined that the processing for all blocks has been completed (S57: YES), the processing according to this routine is performed. Exit.

  As described above, in the present embodiment, since the process to be executed is determined using the determination result of only the frequency component, it is effective when the processing target is known in advance. For example, in the case of a photograph read by an image input device such as a scanner or an image downloaded via a network, processing to be performed on the image can be determined using the determination result of only the frequency component without performing region separation.

It is a block diagram which shows the internal structure of the image processing apparatus which concerns on this invention. It is a block diagram explaining the structure of an image process part. It is a block diagram which shows the internal structure of an image quality adjustment part. It is a conceptual diagram which shows image data Pi (X, Y). It is a conceptual diagram which shows the block before and behind conversion. It is a conceptual diagram which shows the block after converting into a spatial frequency component. It is the table | surface which put together the relationship between the criterion used by a block area | region determination part, and a determination result. It is the table | surface which put together the processing content determined by the combination of the determination result by a frequency component determination part, and the determination result by a block area | region determination part. It is explanatory drawing explaining the area | region which performs an edge emphasis process. It is a conceptual diagram which shows the value which added the row number and column number of the process block in the area | region which performs arithmetic processing. It is explanatory drawing explaining the area | region which performs a smoothing process. It is a conceptual diagram which shows an example of the two-dimensional matrix data used in the case of a smoothing process. It is the table | surface which put together the processing content determined by the combination of the determination result by a frequency component determination part, and the determination result by a block area | region determination part. It is a block diagram explaining the internal structure of the image processing apparatus in which the computer program based on this invention was installed. It is a flowchart explaining the procedure of the process which an image processing apparatus performs. It is a flowchart explaining the procedure of a DCT coefficient determination process. It is a flowchart explaining the procedure of a block area | region determination process. It is a flowchart explaining the procedure of the process which an image processing apparatus performs. It is a flowchart explaining the procedure of a DCT coefficient determination process. It is a flowchart explaining the procedure of the process which an image processing apparatus performs.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Control part 2 Image input part 3 Image processing part 4 Image output part 5 Memory | storage part 6 Operation part 37 Image adjustment part 371 Frequency conversion part 372 Determination part 373 Calculation processing part 374 Inverse frequency conversion part 372a Frequency component determination part 372b Block area determination Unit 372c Processing content determination unit 10 Image processing device 11 CPU
20 recording media

Claims (14)

In an image processing apparatus that performs orthogonal transformation on image data to generate data composed of a plurality of spatial frequency components, and performs image processing based on the generated data,
Detection means for detecting image characteristics based on the image data from the data composed of the plurality of spatial frequency components, and processing corresponding to the detection result of the detection means is performed on data belonging to a predetermined spatial frequency region among the data An image processing apparatus comprising: means; and means for performing inverse orthogonal transform on the data obtained by performing the processing.   A determination unit that determines a magnitude relationship between the magnitude of the spatial frequency component and a preset threshold value, and the detection unit detects a feature of the image based on a determination result of the determination unit; The image processing apparatus according to claim 1, wherein:   The detection includes: a calculation unit that performs a predetermined calculation using a DC component and an AC component among the spatial frequency components; and a determination unit that determines a magnitude relationship between a value obtained by the calculation and a preset threshold value. The image processing apparatus according to claim 1, wherein the unit is configured to detect a feature of the image based on a determination result of the determination unit.   The image processing apparatus according to claim 2, wherein the detection unit is configured to detect whether the image is an edge image based on the determination result.   The image processing apparatus according to claim 1, wherein the processing is edge enhancement processing or smoothing processing.   When the detection means detects that the image is an edge image, an edge enhancement process is performed. When the detection means detects that the image is a non-edge image, a smoothing process is performed. 6. The image processing apparatus according to claim 5, wherein the image processing apparatus is configured as described above.   7. The apparatus according to claim 1, further comprising a region determination unit that determines whether the image data belongs to a character region, a halftone dot region, or a photo region. Image processing apparatus.   When the detection means detects that the image is an edge image, and the area determination means determines that the image data is data belonging to a character area or a photo area, an edge enhancement process is performed. Yes, when the detection means detects that the image is a non-edge image and the area determination means determines that the image data belongs to a halftone area or a photographic area, smoothing processing is performed. The image processing apparatus according to claim 7, wherein the image processing apparatus is applied.   When the edge enhancement process is performed, the data belonging to the spatial frequency range is multiplied by a real number larger than 1, and when the smoothing process is performed, the data is divided by a real number larger than 1. The image processing apparatus according to claim 6.   An image processing apparatus comprising: the image processing apparatus according to claim 1; and means for forming an image on a sheet based on image data processed by the image processing apparatus. Forming equipment. In an image processing method of performing orthogonal transformation on image data to generate data composed of a plurality of spatial frequency components and performing image processing based on the generated data,
A first step of detecting image characteristics based on the image data from data consisting of the plurality of spatial frequency components, and a second step of performing processing according to the detection result on data belonging to a predetermined spatial frequency region among the data. And a third step of performing inverse orthogonal transform on the data obtained by performing the processing.   12. The step according to claim 11, wherein the first step is a step of determining a magnitude relationship between the magnitude of the spatial frequency component and a preset threshold value, and detecting the feature of the image based on the determined result. The image processing method as described.   The first step performs a predetermined calculation using a DC component and an AC component among the spatial frequency components, determines a magnitude relationship between a value obtained by the calculation and a preset threshold value, and results of the determination The image processing method according to claim 11, wherein the image processing method is a step of detecting a feature of the image based on the image. In a computer program for causing a computer to perform orthogonal transformation on image data to generate data composed of a plurality of spatial frequency components, and to perform image processing based on the generated data,
Causing the computer to detect image characteristics based on the image data from the data composed of the plurality of spatial frequency components; and causing the computer to perform processing according to the detection result of the step in a predetermined spatial frequency region of the data. A computer program comprising the steps of: applying to the data to which the computer belongs; and causing the computer to perform inverse orthogonal transform on the data obtained by performing the processing.

Images