JPH0514735A - Image processing device - Google Patents

Abstract

(57) [Summary] [Object] To provide an image processing device having a function of reducing block distortion occurring in an image obtained by reproducing a compressed image. An image processing apparatus comprising: a reproduction unit 1 for reproducing image data compressed for each pixel block composed of a plurality of pixels; and a correction unit 8 for correcting block distortion between pixel blocks. The means 8 is a discrimination means 9 for discriminating the boundary area of the pixel block, and the discrimination means 9
The filter means 11 acts on the boundary area between the adjacent pixel blocks in the boundary area determined by.

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 having reproducing means for reproducing image data compressed for each pixel block consisting of a plurality of pixels, and correcting means for correcting block distortion between pixel blocks. With respect to the above, for example, a device having a compression / expansion function for improving data transmission and storage efficiency for image data of video conference / videophone, color still image, color moving image, etc., AVCC (Audio Visual
The present invention relates to an image processing device used in a device that contributes to computer / communication) fusion.

[0002]

2. Description of the Related Art In a conventional image processing apparatus, the correction means uses an average value filter that uses an average value of eight pixels adjacent to each pixel as new data of the pixel, or an average value filter of eight pixels adjacent to each pixel. There has been a configuration in which a median filter having a median value as new data of the pixel is configured to act on all pixels. In this type of image processing apparatus, international standardization has been attempted as an image compression method for improving data transmission and storage efficiency for an original image, and an orthogonal row transform coding method using DCT (discrete cosine transform) is adopted. It The orthogonal transform coding method is a method in which an image is divided into blocks of several pixels to several tens of pixels, orthogonal transform is applied to each block, and the transform coefficient is encoded and transmitted. On the receiving side, orthogonal inverse transform is performed. This is a method of reproducing an image.

[0003]

However, the above-mentioned prior art has the following drawbacks. Not limited to DCT, if high-efficiency encoding is performed to reduce the average number of bits per pixel, or if the compression rate is increased, the DCT outputs in the block are all linearly summed when the pixels are inversely converted and reproduced, so that the reproduced image quality is improved. However, there is a drawback in that distortion occurs between adjacent blocks at this time, resulting in discontinuous and unsightly images. There is little obstruction of seams between blocks D
LT (discrete Legendre transform) has less block distortion than DCT, but cannot be used because there is no high-speed algorithm. Therefore, all the pixels were averaged by using an average value filter or the like for the reproduced image, but the resolution deteriorates, so that the image is blurred as a whole, such as blurred characters or meandering lines. There was a drawback.
An object of the present invention is to eliminate the above-mentioned conventional drawbacks.

[0004]

To achieve this object, the image processing apparatus according to the present invention is characterized in that the correction means determines the determination means for determining the boundary area of the pixel block and the determination means. The filter means that acts on the boundary area between the adjacent pixel blocks in the selected boundary area. Preferably the filter means is a moving average filter. It is preferable that the correction means operates when the compression rate of the original image is equal to or larger than a set value.

[0005]

With respect to the image reproduced by the reproducing means and having the distortion between the adjacent blocks, the judging means judges the distortion occurrence place, that is, the boundary area of the pixel block, and the filter means detects the difference between the adjacent pixel blocks. Pixels corresponding to the boundary area are smoothed to avoid blurring of the image inside the block, reduce distortion between blocks, and ensure image continuity. By using a moving average filter as the filter means, adjacent blocks are converted into smooth continuous images.
The correction means, the compression ratio that the effect of block distortion is conspicuous,
For example, by setting 1/100 as a set value and activating only when the compression rate is equal to or higher than the set value, filter processing is performed even for an image in which block distortion is not noticeable in a reproduced image for a small compression rate. Avoid unnecessary blurring of the image due to applying

[0006]

Therefore, according to the present invention, even when the high-efficiency coding is performed to reduce the average number of bits per pixel or the compression rate is increased, discontinuity caused by block distortion occurring in a reproduced image is prevented. It has become possible to provide an image processing device that can reduce unsightly.

[0007]

EXAMPLES Examples will be described below. As shown in Figure 1,
Image compression in an image processing device that handles color still images /
The reproducing means 1 stores one frame of raster data as input image data, divides and outputs it in blocks of 8 × 8 pixels, and outputs the reproduced image data in reverse, and an image buffer 2 and each pixel block thereof. An orthogonal transformer 3 for performing DCT (discrete cosine transform) / inverse DCT (inverse discrete cosine transform) on the input, and a two-dimensional coefficient after the transform output is quantized according to a quantization table that presets the transform ratio and determines the compression rate. Is performed in a one-dimensional zigzag scan, and the inverse operation is performed by the quantizer 4, and the one-dimensional quantized data is Huffman coded, and the inverse operation is performed by the encoder / decoder 5. It is composed of a compressed data buffer 6 for storing data and a controller 7 for controlling the operation thereof. The image data reproduced by the image compression / reproduction means 1 and output from the image buffer 2 is output via a correction means 8 for correcting block distortion between pixel blocks. Correction means 8
Is a discrimination unit 9 for discriminating the address of 28 pixels which is the boundary region of the pixel block based on the address information from the orthogonal transformer 3, and between the adjacent pixel blocks in the boundary region discriminated by the discrimination unit 9. Based on the filter means 10 acting on the boundary area and the control signal from the controller 7, the filter means 10 is operated when the compression rate of the original image is equal to or more than a set value, and is output as it is when the compression rate is less than the set value. It is composed of an operation control means 11 for switching the operation / non-operation of the filter means 10.
As shown in FIG. 2, the filter unit 10 includes two adjacent pixels of a boundary pixel of an arbitrary pixel block BL and boundary pixels of each block of four pixel blocks BL adjacent to the pixel block BL. The moving average processing is performed for.

Another embodiment of the present invention will be described below. The filter means 10 used in the previous embodiment is composed of a line buffer of at least two lines in the raster scanning direction, and a digital arithmetic circuit that performs an average value process on specific two pixels adjacent between the line buffers or within the line buffer. Although it can be configured, it is not particularly specified and can be configured by a known digital filter technique. In the above embodiment, the moving average process is performed on two adjacent pixels of the boundary pixels of an arbitrary pixel block BL and the boundary pixels of each block of four pixel blocks BL adjacent to the pixel block BL. However, as shown in FIG. 3, moving average processing may be performed on four adjacent pixels, or as shown in FIG. Alternatively, the moving average process may be performed by including pixels inside. In the previous example,
Although the moving average filter has been described as the filter means 10, the present invention is not limited to this, and any smoothing filter such as a median filter can be used. In the above embodiment, an example in which DCT (discrete cosine transform) is used as the orthogonal transformer 3 has been described. However, the orthogonal transformer 3 is not limited to this, and any discrete Fourier transform, Hadamard transform, or the like may be used. It can be configured with a converter. In the above embodiment, compression / compression is performed on a pixel block of 8 × 8 pixels.
Although the example of performing the decompression process has been described, the present invention is not limited to this, and the pixel block can be configured with an arbitrary number of pixels, such as 16 × 16 pixels, as long as the correlation between pixels can be seen. In the above embodiment, the operation of the filter means is switched and controlled by the operation control means, but the operation switching reference may be performed by a command from the controller based on the compression ratio determined by the quantization table. Further, it may be configured to always operate. The image processing device may be a device that handles moving images, or a device that handles still images such as videophones.

It should be noted that reference numerals are given in the claims for convenience of comparison with the drawings, but the present invention is not limited to the configurations of the accompanying drawings by the entry.

[Brief description of drawings]

FIG. 1 is a configuration diagram of a main part of an image processing apparatus.

FIG. 2 is a simulated view of boundary pixels of a pixel block to be smoothed.

FIG. 3 is a simulated view of boundary pixels of a pixel block to be smoothed according to another embodiment.

FIG. 4 is a simulated diagram of boundary pixels of a pixel block to be smoothed according to another embodiment.

[Explanation of symbols]

1 reproduction means 8 Correction means 9 discrimination means 11 Filter means

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Koji Horikiri             100-1 Sakado, Takatsu-ku, Kawasaki City, Kanagawa Prefecture             Gawa Science Park (KSP) Building B 11F             1113 Kubota Electronic Technology Center Co., Ltd.

Claims (3)

[Claims] 1. An image processing apparatus comprising a reproducing means (1) for reproducing image data compressed for each pixel block composed of a plurality of pixels, and a correcting means (8) for correcting block distortion between the pixel blocks. The correction means (8) determines the boundary area between the adjacent pixel blocks among the boundary areas judged by the judging means (9) for judging the boundary area of the pixel block and the judging means (9). An image processing apparatus comprising a working filter means (11). 2. The image processing apparatus according to claim 1, wherein the filter means (11) is a moving average filter. 3. The image processing apparatus according to claim 1, wherein the correction means (8) operates when the compression rate of the original image is equal to or more than a set value.