JPH11355575A - Image processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processor capable of reproducing an image without making conspicuous block distortion or gradation deficiency even when the quantity of information is reduced. SOLUTION: This image processor has an edge control means 1 for detecting whether the edge component of an image exists in the block of input image data or not and controlling the quantity of information to be allocated to the said block, frequency converting means 2 for conversion to the coefficient signals of plural frequency bands, quantization means 2 for quantizing each of coefficient signals of these frequency bands, inverse converting means 4 for recovering the image data of inputted style and gradation processing means 5 for performing gradation processing. Then, the allocation of the number of quantized bits is changed corresponding to the coefficient signals of converted frequency bands, and any one of gradation processing among more than two kinds of gradation processings in the gradation processing means is selected and performed based on the allocation of the number of quantized bits to the inversely converted image data.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus such as a black-and-white digital electrophotographic copying machine, a color digital electrophotographic copying machine, a laser printer, a facsimile, and the like. The present invention relates to an image processing device that converts a signal into a coefficient signal and processes the coefficient signal.

[0002]

2. Description of the Related Art In conventional image processing, image data sent from a scanner or the like is quantized by an AD converter, converted into a digital real space image signal, filtered, scaled, and processed.
Image processing such as gamma conversion processing and gradation processing was performed.
However, in the image processing using the real space image signal, since various processes are performed with the image signal having redundancy, the amount of the image signal to be processed becomes enormous, and a large amount of memory is required for storing the image signal. In addition, much time was required for image processing. In order to solve the problem, the input image data is divided into a plurality of blocks, each block is subjected to a transform coding process such as a wavelet transform, and converted into a plurality of frequency band coefficient signals. Providing a small number of bits for quantization reduces the amount of processing that must be performed on the image, or reduces the amount of memory required by performing image processing on only part of the transformed coefficient signal. In this case, the processing time was shortened. However, if the amount of information given for each block is reduced in order to increase the amount of information compression, block distortion becomes conspicuous. This is particularly conspicuous in an edge portion. As in the technique disclosed in Japanese Patent Laid-Open No. 2-222386, a contour component is detected from image data before orthogonal transformation, and the information amount given for each block is obtained based on the result. Some have decided on the assignment.

[0003]

As described above, the block distortion is surely reduced by allocating the information amount at the edge portion, but the block distortion is reduced in the portion where the information amount is less allocated. There is a problem that it stands out. Therefore, the problem to be solved by the present invention is to reproduce a high-definition image in which the amount of information is compressed, so that a block having an outline portion of an image such as a character or a line drawing has a large amount of information so that block distortion is not conspicuous. The amount of image data is reduced by allocating an amount and allocating a small amount of information in a block having a gentle change in density, and further, an area gradation processing unit for a block to which a small amount of information is allocated, and an area for a block to which a large amount of information is allocated. Make it equal or larger than the gradation processing unit, make it larger than the reference block size used for encoding, or make the repetition cycle of the area gradation processing different from the repetition cycle of the reference block used for encoding. This makes it possible to reproduce high-quality images without noticeable block distortion and insufficient gradation even if the amount of information is reduced. And to provide a processing apparatus.

[0004]

In order to solve the above problems, according to the first aspect of the present invention, input image data is divided into a plurality of blocks, and each block is converted into coefficient signals in a plurality of frequency bands. An image processing apparatus for detecting the presence or absence of an edge component of an image in a block of input image data and controlling the amount of information allocated to the block; Frequency converting means for converting into coefficient signals of a plurality of frequency bands, quantizing means for quantizing each coefficient signal of the frequency band converted by the frequency converting means, and a coefficient signal quantized by the quantizing means. Inverse conversion means for returning the image data of the input form to the processed image data, and image data of the input form returned by the inverse conversion means And a gradation processing means for performing gradation processing, wherein the edge control means determines the number of quantization bits of the quantization means with respect to the converted coefficient signals of a plurality of frequency bands depending on the presence or absence of an edge component. In addition to changing the assignment, any one of two or more tone processes in the tone processing unit is performed based on the assignment of the number of quantization bits to the image data in the input form returned by the inverse transform unit. One of them is selected for processing. According to a second aspect of the present invention, in the image processing apparatus according to the first aspect, the frequency transforming means is configured to transform the signal into coefficient signals of a plurality of frequency bands by wavelet transform. According to the third aspect of the present invention, in the first aspect or the second aspect,
In the image processing apparatus described above, the area gradation processing unit for a block with a small information amount is equal to or larger than the area gradation processing unit for a block with a large allocation,
Alternatively, the configuration is made larger. According to a fourth aspect of the present invention, in the image processing apparatus of the first or second aspect, the area conversion processing unit for the block to which the information amount is small is converted to the coefficient signal of the frequency band by the frequency conversion unit. It is configured to be larger than the block size to be referred for conversion. According to a fifth aspect of the present invention, in the image processing apparatus of the first, second, third or fourth aspect, the repetition period of the area gradation processing is determined by the frequency conversion means. The configuration is such that the repetition period and the phase of the block referred to for converting into a signal are different.

[0005]

According to the first aspect of the present invention, the presence or absence of an edge component for each block is detected by the edge control means, and the converted coefficient signals of a plurality of frequency bands are used by the quantization means. By changing the allocation of the number of quantization bits,
Since any one of two or more gradation processes in the gradation processing means is selected and processed on the image data after the inverse transformation based on the allocation of the quantization bit number, the quantization is performed. Area gradation processing unit for a block to which a small amount of information is allocated by the converting means can be equal to or larger than the area gradation processing unit for a block to which a large amount of information is allocated, thereby reducing the amount of information. Also, it is possible to perform image reproduction in which block distortion and insufficient gradation are inconspicuous. According to a second aspect of the present invention, in addition to the image processing apparatus according to the first aspect, the frequency transforming means is capable of transforming into coefficient signals of a plurality of frequency bands by wavelet transform. Therefore, even if the amount of information is reduced, block distortion can be suppressed. According to a third aspect of the present invention, in addition to the invention of the image processing apparatus configured as in the first or second aspect, the area gradation processing unit for a block with a small amount of information is assigned to a block with a large allocation. Is larger than or equal to the area gradation processing unit for the block, so that the block to which the information amount is allocated is subjected to the gradation processing larger in the area gradation processing unit, The distortion is reduced. Further, the invention of claim 4 is based on claim 1
Alternatively, in addition to the invention of the image processing apparatus configured as in claim 2, the area gradation processing unit for a block to which the information amount is small is set to be larger than the reference block size used for encoding. Gradation processing taking into account the influence of adjacent blocks is performed, so that boundary steps between blocks become less noticeable. Claim 5
According to the invention, in addition to the invention of the image processing apparatus configured as in claim 1, 2, 3, or 4, the repetition period of the area gradation processing is used for a reference block used for encoding. Since the repetition period and the phase are made different, gradation processing taking into account the influence of an adjacent block is performed, and the boundary step between blocks becomes less noticeable.

[0006]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, embodiments of the present invention will be described in detail with reference to the accompanying drawings. FIG. 1 is a block diagram showing a main part of a black-and-white digital electrophotographic copying machine (hereinafter referred to as a copying machine) which is a typical image processing apparatus. A wavelet transform is taken as an example of a means for inputting image data and converting it into coefficient signals in a plurality of frequency bands. The copier shown in FIG. 1 reads an image from a document and converts it into digital image data. The scanner 8 takes in image data output from the scanner 8 and detects the presence or absence of an edge component of the image for each block. An edge control unit (edge control unit) 1 for controlling the amount of information to be allocated, a wavelet transform unit (frequency conversion unit) 2 as a frequency conversion unit for inputting digital image data and converting it into coefficient signals in a plurality of frequency bands, A quantizing unit (quantizing means) 3 for quantizing the coefficient signal converted into a plurality of frequency bands; an image storing unit 6 for storing the quantized coefficient signal; and a coefficient signal stored in the image storing unit 6 And a wavelet inverse transform unit (inverse transform unit) 4 as an inverse transform unit for returning the image data to a real space image signal having the same form as the input image data. Processing unit (gradation processing means) 5 that performs gradation processing such as image processing and error diffusion, a printer 9 that outputs image data to paper, a scanner 8, a wavelet transform unit 2, a quantization unit 3, an image storage Unit 6, wavelet inverse transform unit 4,
The image processing apparatus includes a gradation controller 5, an edge controller 1, and a system controller 7 for providing operation parameters of the printer 9.

In the scanner 8, an image is read from a document by a CCD, photoelectrically converted, and the image converted into an electric signal by an A / D converter is converted into digital data.
Further, the digitalized image data is subjected to shading correction for correcting variations in light sensitivity of individual pixels of the CCD, and then the image data is output to the wavelet transform unit 2. Further, the digital image signal is subjected to wavelet transform by the wavelet transform unit 2, and is converted into coefficient signals of a plurality of frequency bands. That is, the wavelet transform unit 2 calculates the following equations (1-1) to (1-4)
Are converted using the low-pass filters s (x) and s (y) and the high-pass filters h (x) and h (y) in the main scanning direction (X direction) and the sub-scanning direction (Y direction) as basic wavelet functions. Do. _{s (x) = (x n} + x n + 1) / 2 ···· (1-1) s (y) = (y n + y n + 1) / 2 ···· (1-2) h ( x) = x _n −x _{n + 1} ... (1-3) h (y) = y _n −y _{n + 1} ... (1-4) FIG. 2 shows a configuration of a wavelet transform unit 2 that performs a transform. Real space image signal _dij
Is decomposed into low-frequency component and high-frequency component coefficient signals in the main scanning direction by a low-pass filter s (x) 21 and a high-pass filter h (x) 22, respectively.
The signal is down-sampled by 1/2 by 3, 24 to obtain coefficient signals w1, w2. Further, each coefficient signal w1,
w2 is decomposed into low-frequency component and high-frequency component coefficient signals in the sub-scanning direction by low-pass filters s (y) 25 and 27 and high-pass filters h (y) 26 and 28, and then down-samplers 29, 30, 31, 1/2 by 32
And the coefficient signals w3, w4, w
5. Obtain w6. FIG. 4 shows the configuration of the wavelet transform unit 2 that performs wavelet transform on 4 × 4 pixels. The real space image signal _dij is a low-pass filter s (x)
After being decomposed into low frequency component and high frequency component coefficient signals in the main scanning direction by a high-pass filter 41 and a high-pass filter h (x) 42, the down-samplers 43 and 44 down-sample the signal by ２. obtain. Furthermore, low-pass filters s (y) 45 and 47 and high-pass filters h (y) 4 are applied to the respective coefficient signals w1 and w2.
The signal is decomposed into low frequency component and high frequency component coefficient signals in the sub-scanning direction by 6 and 48, and then down-samplers 49 and 5
Down-sampling is performed by 1/2 with 0, 51, and 52 to obtain coefficient signals w3, w4, w5, and w6.

Next, a low-frequency component in the main scanning direction and a high-frequency component in the sub-scanning direction are respectively applied to the coefficient signal w3 by a low-pass filter s (x) 53 in the main scanning direction and a high-pass filter h (y) 54 in the sub-scanning direction. After being decomposed into the coefficient signals of the components, the down-samplers 55 and 56 down-sample the signal by 1/2, and the obtained coefficient signals w7 and w8 are further subjected to low-pass filters s (y) 58 and 60 and a high-pass filter h. (Y) The signal is decomposed into low-frequency component and high-frequency component coefficient signals in the sub-scanning direction by 59 and 61, and then down-sampled by 1/2 by down-samplers 62, 63, 64 and 65 to obtain coefficient signals w11, w12 and w.
13, w14 is obtained. Similarly, although not shown, the low-pass filters s (x) and s (y) and the high-pass filter h (x) in the main scanning direction and the sub-scanning direction are respectively applied to the coefficient signals w4, w5, and w6. h (y) filtering,
The downsampling process is repeated in the same process as described above, and the entire signal is decomposed into 16 coefficient signals w11 to w26, including the process of the coefficient signal w3.

FIG. 5A shows a basic wavelet function for extracting low-frequency components, which corresponds to the low-pass filters s (x) and s (y) in FIG. 3 or FIG. FIG.
(B) represents a basic wavelet function for extracting high-frequency components, and is a high-pass filter h (x) shown in FIG. 3 or FIG.
And h (y). The image data d _ij shown in FIG. 3 or FIG. 4 is converted into coefficient signals w3 to w6 or w11 to w2 in a plurality of frequency bands by wavelet transform.
6 and output from the wavelet transform unit 2. Further, in the quantizing unit 3, the edge control unit 1 quantizes the quantizer 1 to the quantizer 4 or the quantizer 1 to
The quantization is performed by the quantizer 16 with different quantization bit numbers. Here, the number of allocated bits in each quantizer is determined based on the result of the edge control unit 1 detecting the presence or absence of an edge. Here, in the edge detection of the edge control unit 1, the presence or absence of an edge component included in the image data is detected by pattern matching or by a differential filter. This technique is well known and will not be described here.

[0010] Next, in the quantizing section 3, the number of quantized bits differed by the quantizer 1 to quantizer 4 in FIG. 3 and the quantizer 1 to quantizer 16 in FIG. At this time, the edge control unit 1 detects the presence or absence of an edge component for each block, and in a pixel or an area where no edge component exists, a small number of bits for a high frequency component and a low frequency component Can be assigned a large number of bits. Conversely, a larger number of bits is assigned and quantized so that a sufficient number of bits can be allocated to a high-frequency component in a region where an edge component exists. That is, a larger amount of information is assigned to a block having an edge component, and a smaller amount of information is assigned to a block having no edge component. The image accumulating unit 6 accumulates the quantized coefficient signals in a memory and changes the reading order of the memory to perform image processing such as editing processing and rotation processing. Next, the output coefficient signal from the image storage unit 6 is input to the inverse wavelet transform unit 4 and converted from the coefficient signal into real space image data in the form of the original input image data.
In the inverse wavelet transform unit 4, the wavelet transform unit 2
, Ie, equations (1-1) to (1-
The conversion exactly the reverse of the conversion shown in 4) is performed. The following equations (2-1) to (2-4) show inverse transformation equations for equations (1-1) to (1-4) for obtaining a real space image. _{x n = s (x) +} h (x) / 2 ···· (2-1) x n + 1 = s (x) -h (x) / 2 ···· (2-2) y n = s (y) + h (y) / 2 (2-3) y _{n + 1} = s (y) -h (y) / 2 (2-4) In addition, FIG. 2 shows a configuration example of the tone processing unit 5. The real space image data from the inverse wavelet transform unit 4 is input to a first gradation processing 5b and a second gradation processing 5c, and gradation processing is performed according to each processing procedure. The selector 5a receives the output results of the first gradation processing 5b and the second gradation processing 5c, and selects one of the outputs based on the information amount allocation by the edge control unit 1 depending on the presence or absence of an edge. It has become. Here, the area gradation processing unit which is the area processing unit for performing the gradation processing is, for example, the first gradation processing 5b is a simple multi-value processing in the unit of one pixel, and the second gradation processing 5c is N
The area gradation processing unit of the second gradation processing 5c is set to be equal to or larger than the area gradation processing unit of the first gradation processing 5b such as multi-value dither processing in × N (N is a natural number) pixel units. Set.

Next, in the selector 5a, based on the edge detection result for each block by the edge control unit 1 described above, for example, for a pixel or block having no edge component,
Since a small number of bits are assigned by the quantization unit 3, the result of performing the second gradation processing 5c, which is a larger area gradation processing unit, on the decoded image data is selected, and conversely. Since a large number of bits are assigned by the quantization unit 3 to the block in which the edge component exists, the first gradation processing 5b which is a small area gradation processing unit for the image data after the inverse transform is performed. To select and output the result. Further, in the second gradation processing 5c, the reference block size used for the wavelet transform is 4 × 4.
Assuming pixels, the matrix size of the dither processing is set to be larger than the reference block size used for the wavelet transform, for example, 6 × 6 pixels. In FIG.
The dither matrix size of the second gradation processing 5c is switched by a dither matrix size selection signal (not shown) from the system controller 7. Here, a switching method in which the dither matrix is configured from a threshold table in a memory and the matrix size is switched by switching addresses is generally known. For example, if the converted output of the inverse wavelet transform unit 4 in FIG. 2 is a 4 × 4 pixel block, and each pixel is an 8-bit signal, as an example of a dither matrix used in the second gradation processing 5c, When the size is equal to the reference block size used for coding, the dither matrix size of 4 × 4 pixels in FIG. 6A is used, and when the size is larger than the reference block size used for coding, 6 × in FIG.
A dither matrix size of 6 pixels or the like is used.

FIG. 7 shows the relationship between a two-dimensional array of reference block sizes and a two-dimensional array of dither processing matrices for the reference block size. Solid block 1
The size of 1 indicates the reference block size used for the wavelet transform, and the size of the dashed block 12 indicates the block size of the dither matrix used in the second gradation processing 5c. If the area gradation processing unit for the block to which the information amount is small is larger than the reference block size used for the wavelet transform and gradation processing is performed, gradation processing taking into account the influence of adjacent blocks is performed. As a result, image processing in which distortion generated for each block is alleviated can be performed. FIG. 8 shows a state in which the phase of a reference block used for encoding is different from the phase of a block of a dither matrix used in the second gradation processing 5c. A solid block 13 is a reference block used for wavelet transform, and a broken block 14 is a second gradation processing 5.
The block of the dither matrix used in c is shown, and both blocks are shown by the same 4.times.4 pixels, but there is no problem as long as the size is equal or larger. Here, the operation is performed such that the phase of the block of the dither matrix is shifted by two pixels in the main scanning direction and the sub-scanning direction with respect to the reference block used for the wavelet transform. If gradation processing is performed with different phases in this way, gradation processing incorporating the influence of adjacent blocks is performed, and image processing in which distortion generated for each block is reduced is performed. Can be. Further, the output from the gradation processing unit 5 is input to the printer 9 to perform paper output. Also,
FIG. 2 shows two gradation processes held by the gradation processing unit 5, a first gradation process 5b and a second gradation process 5c. May be configured to select any one of the gradation processes.

[0013]

As described above, according to the first aspect of the present invention,
The area gradation processing unit for a block with a small amount of information allocation can be equal to or larger than the area gradation processing unit for a block with a large information amount allocation. It is possible to provide an image processing apparatus capable of reproducing an image with inconspicuous gradation and insufficient gradation. According to the invention of claim 2, in addition to the invention of the image processing apparatus of claim 1, it is possible to convert the signal into coefficient signals of a plurality of frequency bands by wavelet transform. It is possible to provide an image processing apparatus capable of reproducing an image with inconspicuous gradation and insufficient gradation. According to the invention of claim 3, in addition to the invention of the image processing apparatus of claim 1 or 2, gradation processing of a larger area gradation processing unit is performed on a block to which a small amount of information is allocated. As a result, the block distortion has been alleviated, so that it is possible to provide an image processing apparatus capable of reproducing an image in which the amount of information is reduced and block distortion and insufficient gradation are not noticeable. According to the invention of claim 4, in addition to the invention of the image processing apparatus of claim 1 or claim 2, the area gradation processing unit for the block to which the information amount is small is
Since the size of the reference block used for encoding is larger than that of the reference block and the influence of adjacent blocks is taken into account to perform gradation processing in which the boundary step between the blocks is inconspicuous, even if the information amount is reduced, the block size is reduced. It has become possible to provide an image processing apparatus capable of reproducing an image with no noticeable distortion or insufficient gradation. According to a fifth aspect of the present invention, in addition to the image processing apparatus of the first, second, third or fourth aspect, the repetition period of the area gradation processing is used for a reference block used for encoding. By making the repetition cycle and phase different, it is possible to apply the influence of adjacent blocks and perform gradation processing with inconspicuous boundary steps between blocks. It has become possible to provide an image processing apparatus capable of inconspicuous image reproduction.

[Brief description of the drawings]

FIG. 1 is a block diagram of a main part of an image processing apparatus showing an example of an embodiment of the present invention.

FIG. 2 is a block diagram of a main part of a gradation processing unit in the image processing apparatus according to the embodiment of the present invention.

FIG. 3 is an explanatory diagram illustrating a configuration of a wavelet transform unit that performs wavelet transform on 2 × 2 pixels.

FIG. 4 is an explanatory diagram illustrating a configuration of a wavelet transform unit that performs wavelet transform on 4 × 4 pixels.

FIG. 5A is an explanatory diagram illustrating a basic wavelet function for extracting low frequency components. (B) is an explanatory view for explaining a basic wavelet function for extracting a high-frequency component.

FIG. 6A is an explanatory diagram illustrating an example of a 4 × 4 pixel dither matrix used in the second gradation processing.
FIG. 3B is an explanatory diagram illustrating an example of a 6 × 6 pixel dither matrix used in the second gradation processing.

FIG. 7 is an explanatory diagram illustrating an example in which a unit of area gradation processing is larger than a reference block size used for wavelet transform and gradation processing is performed.

FIG. 8 is an explanatory diagram illustrating a state in which the phase of a reference block used for wavelet transform and the phase of a block of a dither matrix used in second gradation processing are arranged to be different.

[Explanation of symbols]

1 edge control unit (edge control means), 2 wavelet transformation unit (frequency transformation means), 3 quantization unit (quantization means), 4 wavelet inverse transformation unit (inverse transformation means),
5 gradation processing section (gradation processing means), 5a selector, 5 gradation processing section
b First gradation processing, 5c Second gradation processing, 6 Image storage unit, 7 System controller, 8 Scanner, 9 Printer

 ──────────────────────────────────────────────────続 き Continuing from the front page (72) Inventor Hiroshi Ishii 1-3-6 Nakamagome, Ota-ku, Tokyo Inside Ricoh Company (72) Inventor Takahiro Yanagita 1-3-6 Nakamagome, Ota-ku, Tokyo Share Inside Ricoh Company (72) Inventor Yukiko Yamazaki 1-3-6 Nakamagome, Ota-ku, Tokyo Stock Company Ricoh Company

Claims (5)

[Claims] 1. An image processing apparatus which divides input image data into a plurality of blocks, converts each block into coefficient signals of a plurality of frequency bands, and processes the coefficient signals. Edge control means for detecting the presence or absence of a component and controlling the amount of information allocated to the block; frequency conversion means for converting each block of the input image data into coefficient signals in a plurality of frequency bands; and the frequency conversion means Quantizing means for quantizing each coefficient signal of the frequency band converted by the transforming means, and inverse transforming means for converting the coefficient signal quantized by the quantizing means into image data of a form inputted to the processed image data; and Gradation processing means for performing gradation processing on the image data in the input form returned by the inverse conversion means, wherein the edge control means Depending on the presence or absence of the di component, while changing the assignment of the number of quantization bits of the quantization means to the converted coefficient signals of the plurality of frequency bands, the image data of the input form returned by the inverse transformation means An image processing apparatus for selecting one of two or more gradation processes in the gradation processing means based on the assignment of the number of quantization bits. 2. The image processing apparatus according to claim 1, wherein
An image processing apparatus characterized in that the frequency converting means converts the signal into coefficient signals of a plurality of frequency bands by wavelet transform. 3. The image processing apparatus according to claim 1, wherein an area gradation processing unit for a block having a small amount of information is equal to or larger than an area gradation processing unit for a block having a large amount of information. An image processing apparatus comprising: 4. The image processing apparatus according to claim 1, wherein the area conversion processing unit for the block to which the information amount is small is converted by the frequency conversion unit into a frequency band coefficient signal. An image processing apparatus characterized in that the size of the block is larger than the size of the block. 5. The image processing apparatus according to claim 1, wherein said frequency conversion means converts a repetition period of the area gradation processing into a coefficient signal in a frequency band. An image processing apparatus characterized in that a repetition period and a phase of a block referred to in (1) are made different.