JP4469643B2 - Image processing circuit - Google Patents

Description

  The present invention relates to an image processing circuit for compressing image data for each predetermined block by an MPEG (Moving Picture Experts Group) method.

  The MPEG system is known as a moving image compression technique. In the MPEG method, image data of each frame (screen) constituting a moving image is compressed for each predetermined block using a picture type such as an I picture, a P picture, or a B picture. For example, an image of one frame is divided into a plurality of blocks each consisting of 8 pixels × 8 pixels = 64 pixels, and compression processing is performed in units of each block.

  Here, the I picture is compressed only with the image data of the current frame without referring to the image data of other frames. The P picture is compressed using image data of the current frame and image data of another frame preceding in time. The B picture is compressed using the image data of the current frame and the image data of other frames that precede and follow in time.

  In the case of a P picture, for example, the image data of the current frame is compared with the image data of the previous frame to generate a motion vector (motion vector). Also, difference data between the image data of the current frame and the image data of the previous frame that has been motion-compensated based on the motion vector is generated. Then, a P picture is obtained by performing compression processing based on the motion vector and the difference data.

  As described above, by performing the compression process based on the motion vector and the difference data, it is possible to obtain a relatively high-quality image even after the compression process. However, in the conventional image processing circuit that performs image data compression processing for each predetermined area in the MPEG method, if the difference data value is small or the bit number allocation is small, the difference data is not used in the block. In many cases, compression processing is performed using only vectors.

  Images with little color change meet the above conditions. However, if the above processing is performed on an image that gradually changes its color, such as the sky or walls, the color gradation boundary line is noticeable and compressed. There has been a problem that the image quality of the processed image deteriorates.

  One conventional method for dealing with this problem is to set a small coefficient value of a quantization table used at the time of quantization, for example, and allocate a large number of bits. However, with this, a high compression rate cannot be realized.

  An object of the present invention is to provide an image processing circuit that can solve the problems based on the conventional technique and improve the image quality of an image after compression without increasing the amount of data for an image with little color change. There is to do.

In order to achieve the above object, the present invention provides an image processing circuit for compressing image data for each predetermined block in the MPEG system,
If the value of the DC coefficients of the transformed data obtained by the image data included in the block of each of the current frame by the DCT transform is greater than the reference value set in advance, the spatial change in color image data A color change detection circuit that determines that there are few, a difference generation circuit that generates difference data between image data of the current frame and image data of another frame, and image data of the current frame and image data of another frame A motion detection circuit for detecting a motion vector between
When the color change detection circuit determines that the image has little color change, a compression process is performed based only on the difference data generated by the difference generation circuit without detecting a motion vector by the motion detection circuit. The present invention provides an image processing circuit characterized in that it is performed.

  According to the present invention, in the case of an image with little color change, by performing compression processing based only on difference data without detecting a motion vector, the amount of data after compression processing is increased without increasing the amount of data. Image quality can be improved.

  Hereinafter, an image processing circuit of the present invention will be described in detail based on preferred embodiments shown in the accompanying drawings.

  FIG. 1 is a block diagram of an embodiment showing an internal configuration of an image processing circuit of the present invention. An image processing circuit 10 shown in the figure compresses image data for each predetermined block by the MPEG method, and includes a color change detection circuit 12, a frame memory 14, a difference generation circuit 16, a DCT conversion circuit 18, and a quantization circuit. 20, an encoding circuit 22, a motion detection circuit 24, a motion compensation circuit 26, an encoding circuit 28, a multiplexing circuit 30, and the like.

  In this embodiment, in order to simplify the description, an image of one frame is divided into a plurality of blocks of 8 pixels × 8 pixels = 64 pixels, and image data for 64 pixels is compressed for each block. A case will be described as an example.

  In the image processing circuit 10, a color change detection circuit 12 detects a spatial color change in the image data of the current frame for each block. The color change detection circuit 12 is sequentially inputted with image data of the current frame in units of blocks. Then, the color change detection circuit 12 detects whether each block image is an image with little color change, and outputs a detection signal thereof.

  Note that the color change detection method is not limited in any way. For example, the color change can be detected by comparing a predetermined number of high-order bits of image data having a predetermined bit length included in each block. For example, when only the upper 4 bits of 8-bit image data for 64 pixels included in a block are compared and the difference is smaller than a preset reference value, the image has little color change. It can be judged.

  It is also possible to detect a color change based on the DC coefficient value of converted data obtained by DCT conversion of image data included in each block. The DCT transform and the DC coefficient will be described later. For example, when the image data for 64 pixels included in the block is DCT converted and the value of the DC coefficient is larger than a preset reference value, it can be determined that the image has little color change.

  Subsequently, the frame memory 14 holds image data of a temporally preceding frame. Although a part of the illustration is omitted in FIG. 1, in the case of the present embodiment, the motion compensation circuit is converted into data obtained by inverse quantization of the quantized data output from the quantization circuit 20 and further inverse DCT conversion. The image data obtained by adding the motion compensated image data output from H.26 is held in the frame memory 14 as the image data of the previous frame.

  It should be noted that the number of frames of image data that the frame memory 14 holds in time can be appropriately determined as necessary.

  The difference generation circuit 16 generates, for each block, difference data between the image data of the current frame and the image data of other frames according to the detection signal output from the color change detection circuit 12. The difference generation circuit 16 includes the image data of the current frame, the image data of the previous frame output from the frame memory 14, and the motion-compensated previous frame output from the motion compensation circuit 26. Image data is input in units of blocks.

  When the color change detection circuit 12 determines that the image has little color change, the difference generation circuit 16 determines the image data of the current frame and the previous frame output from the frame memory 14 according to the detection signal. Difference data with image data is generated. On the other hand, if it is determined that the image has a large color change, difference data between the image data of the current frame and the image data of the previous frame that has been subjected to motion compensation output from the motion compensation circuit 26 is generated. To do.

A DCT (Discrete Cosine Transform) conversion circuit 18 orthogonally transforms the difference data output from the difference generation circuit 16 and outputs the converted data. That is, the difference data is converted into data of each frequency component. Of the converted data (r _XY ) included in the block of 8 pixels (X = 0 to 7) × 8 pixels (Y = 0 to 7), r ₀₀ is called a DC coefficient, and r _{01 to} r ₇₇ are AC coefficients. Called. The DC coefficient r ₀₀ indicates an average value (DC component) of 64 pixels.

  The quantization circuit 20 quantizes the conversion data output from the DCT conversion circuit 18 using a quantization table, and outputs the quantized data. That is, the converted data is divided by the coefficient of the quantization table and converted into an integer to the nearest numerical value.

  The encoding circuit 22 performs variable length encoding on the quantized data output from the quantizing circuit 20 using, for example, a Huffman encoding method, and outputs the encoded data.

  For each block, the motion detection circuit 24 determines between the image data of the current frame and the image data of the previous frame output from the frame memory 14 according to the detection signal output from the color change detection circuit 12. The motion vector is detected. The motion detection circuit 24 receives image data of the current frame and image data of the previous frame output from the frame memory 14 in units of blocks.

  When the color change detection circuit 12 determines that the image has many color changes, the motion detection circuit 24 determines the image data of the current frame and the previous frame output from the frame memory 14 according to the detection signal. A motion vector between image data is detected. On the other hand, when it is determined that the image has little color change, the motion vector is not detected.

  The motion compensation circuit 26 performs motion compensation on the image data of the previous frame output from the frame memory 14 on the basis of the motion vector output from the motion detection circuit 24 for each block. The image data of the previous frame is output.

  The encoding circuit 28 performs variable length encoding on the motion vector output from the motion detection circuit 24 using, for example, a Huffman encoding method, and outputs encoded data.

  The multiplexing circuit 30 multiplexes the encoded data output from the encoding circuit 22 and the encoded data output from the encoding circuit 28, and outputs the result as compressed data.

  The specific circuit configurations of the color change detection circuit 12, the difference generation circuit 16, and the motion detection circuit 24 are not limited at all, and various configurations that perform the same function can be used. Also, the frame memory 14, the DCT conversion circuit 18, the quantization circuit 20, the encoding circuits 22, 28, the motion compensation circuit 26, and the multiplexing circuit 28 are conventionally known, and those having various known configurations are used. Is possible.

  Next, the operation of the image processing circuit 10 will be described by taking as an example a case where a P picture is obtained by performing compression processing for each block.

  In the image processing circuit 10, when image data of the current frame is input in units of each block, the color change detection circuit 12 detects the color change of the block image. That is, it is detected whether the image of the block is an image with little color change. From the color change detection circuit 12, detection signals of the respective blocks are output and input to the difference generation circuit 16 and the motion detection circuit 24.

  Here, when the color change detection circuit 12 determines that the image of the block is an image with little color change, the motion detection circuit 24 does not detect the motion vector according to the detection signal.

  Further, according to the detection signal, the difference generation circuit 16 generates difference data between the image data of the current frame and the image data of the previous frame output from the frame memory 14. The difference data is DCT converted by the DCT conversion circuit 18, quantized by the quantization circuit 20, variable-length encoded by the encoding circuit 22, and then output from the multiplexing circuit 30 as compressed data.

  On the other hand, when the color change detection circuit 12 determines that the image of the block is an image with many color changes, the motion detection circuit 24 outputs the image data of the current frame and the frame memory 14 according to the detection signal. A motion vector between the image data of the immediately preceding frame is generated. The motion vector is variable-length encoded by the encoding circuit 28.

  Subsequently, the motion compensation circuit 26 performs motion compensation on the image data of the previous frame output from the frame memory 14 based on the motion vector output from the motion detection circuit 24, and the motion compensated 1 The image data of the previous frame is output.

  Subsequently, the difference generation circuit 16 generates difference data between the image data of the current frame and the image data of the previous frame subjected to motion compensation output from the motion compensation circuit 26. The difference data is DCT converted by the DCT conversion circuit 18, quantized by the quantization circuit 20, and variable-length encoded by the encoding circuit 22.

  Thereafter, the encoded data output from the encoding circuit 22 and the encoded data output from the encoding circuit 28 are multiplexed by the multiplexing circuit 30 and output as compressed data.

  As described above, in the case of an image with little color change, by performing compression processing based only on the difference data without detecting a motion vector, the image quality of the image after compression processing can be improved without increasing the data amount. Can be improved.

  The block is not limited to 8 pixels × 8 pixels, and an image of one frame is divided into a plurality of blocks of m pixels × n pixels (m and n are integers of 1 or more), and each block is compressed as a unit. Processing can be performed. In addition, the compression processing may be performed with reference to image data of a plurality of frames up to two or more before, not just the previous frame. Further, although the case of the P picture has been described, the same applies to the case of the B picture.

The present invention is basically as described above.
The image processing circuit of the present invention has been described in detail above. However, the present invention is not limited to the above-described embodiment, and various modifications and changes may be made without departing from the spirit of the present invention. is there.

It is a block diagram of one embodiment showing an internal configuration of an image processing circuit of the present invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Image processing circuit 12 Color change detection circuit 14 Frame memory 16 Difference generation circuit 18 DCT conversion circuit 20 Quantization circuit 22 Encoding circuit 24 Motion detection circuit 26 Motion compensation circuit 28 Encoding circuit 30 Multiplex circuit

Claims (1)

An image processing circuit for compressing image data for each predetermined block in the MPEG system,
If the value of the DC coefficients of the transformed data obtained by the image data included in the block of each of the current frame by the DCT transform is greater than the reference value set in advance, the spatial change in color image data A color change detection circuit that determines that there are few, a difference generation circuit that generates difference data between image data of the current frame and image data of another frame, and image data of the current frame and image data of another frame A motion detection circuit for detecting a motion vector between
When the color change detection circuit determines that the image has little color change, a compression process is performed based only on the difference data generated by the difference generation circuit without detecting a motion vector by the motion detection circuit. An image processing circuit characterized by performing the processing.

Images