内容发布更新时间 : 2024/9/20 16:45:37星期一 下面是文章的全部内容请认真阅读。

Image coding device for coding image signal to reduce the amount of the information in the image

BACKGROUND OF THE INVENTION 1. Field of the Invention

The present invention relates to an image coding device which is capable of reducing the amount of information contained in an image by coding an input image signal, in particular, a moving image. 2. Description of the Related Art

In recent days, with the prevail of service integrated digital networks (referred to as a ISDNs), as new communication service, image communication service has been more and more requested.

Concretely, a TV phone or a TV conference system has requested such service.

In general, in the case of transmitting image information as in a TV phone or a TV conference system, since the image consumes a massive amount of information, in light of a transmission speed of a line and the cost in the transmission, it is necessary to compress and code the amount of information of the transmitted image for reducing the amount of information in transmission.

As a coding method for compressing the information, various methods have been considered. Some of them are going to be realized in the field of a TV phone or a TV conference system. Further, a new communication network, representatively a broadband integrated

service digital network (referred to as a B-SIDN), is now actively studied for providing higher service. As a high-efficiency coding method for an image being now on the way of implementation, there has been known a motion compensated inter-frame prediction orthogonal transformation method. This method takes the steps of dividing an image into plural small blocks and performing an orthogonal

transformation such as a discrete cosine transformation with respect

to various predictive error signals in each small block, quantizing the resulting transforming coefficient for coding, and transmitting the coded data.

FIG. 9 shows a coding device arranged on the foregoing prior art. FIG. 10 is a block diagram showing an arrangement of the foregoing conventional decoding device.

An input image signal is inter-frame predicted as compensating for the motion at each image frame by means of a motion inter-frame compensation predicting unit 40. To detect the motion in the motion compensation, the image signal of a one-previous frame stored in a frame memory 41 and the input image signal are divided into blocks. Which portion of the image signal of a one-previous frame is moved to the block is detected. The motion inter-frame compensation predicting unit 40 operates to read an image frame coded of a one-previous frame stored in the frame memory 41 as a predictive value. Then, a

difference between the input image frame and the predictive value read from the frame memory 41 is derived in a differential operating unit 42 for performing an inter-frame coding operation.

In succession, the predictive error signal output from the differential operating unit 42 is transmitted to an orthogonal transforming unit 43 in which an orthogonal transformation is

performed. The orthogonal transforming unit 43 serves to output an orthogonally transformed coefficient. The coefficient is quantized in the quantizing unit 44 at the quantization level sent from a coding control unit 45. The quantized output from the quantizing unit 44 is output to the external as a coded result. The quantized output from the quantizing unit 44 is also sent to a reverse-quantizing unit 46. The reverse-quantizing unit 46 performs a reverse process to the quantizing unit 44 so that the unit 46 may output an orthogonally transformed coefficient and a reverse orthogonally transformed coefficient. In the reverse orthogonally transforming unit 47, the reverse orthogonally transforming operation is carried out by using the coefficients. In these serial operating process, to efficiently perform an operation like motion detection in compensating for the motion, the image is divided to blocks each having a predetermined size. The output of the reverse orthogonal transforming unit 47 is added to the predictive value read out of the frame memory 41 in the adding unit 48. Then, the added value is stored in the frame memory 41 and is used for inter-frame prediction in the next input image frame. The input image signal is coded according to such a loop (coding loop).

In the decoding device, like the process in the coding loop of the coding device, the input coded signal is sent to a reverse-quantizing unit 49 in which the reverse operation to the quantizing unit 44 is performed. The orthogonally transformed coefficient is output to the reverse orthogonal transforming unit 49.

In an adding unit 51, the output of the reverse orthogonal

transforming unit 50 is added to a predictive value read out of a frame memory 53 based on a predictive motion value from an inter-frame predicting unit 52. The added result is output as a signal and is stored in the frame memory 53.

The foregoing conventional image coding device allows a certain degree of error for reducing the amount of information when

quantizing the orthogonally transformed result. The allowed error

makes the quality of the image degraded. This has been a disadvantage. The quantization is controlled by the amount of generated codes and the transmission capacitance. In each divided block, the quantizing condition is changed. When the coded image is viewed as an overall image, a block distortion in which a block being processed is striking or coarseness variable in each block of the screen takes place so that the quality of the image may be disadvantageously degraded.

Further, whether or not the content of the image on the block to be coded may be an important portion to transmission or storage, the coding operation is evenly controlled for any of the blocks. Hence, the image quality on an important portion cannot be held so that the overall image may be disadvantageously degraded. Further, by reverse-quantizing and reverse orthogonally transforming all the outputs of the quantizing unit for locally decoding, the locally decoded result is stored in the frame memory and is used for inter-frame predictive coding. If lost of information takes place in a line such as a

broadband ISDN, the inter-frame predictive coding does not properly function, thereby remarkably lowering the quantity of the image. SUMMARY OF THE INVENTION

It is therefore an object of the invention to provide an image area detector which can recognize an important portion in the image to be coded.

It is another object of the invention to provide an image coding

device which uses an image area detector for recognizing an important portion in the image to be coded and enables to properly code the