Huawei Technologies Co., Limited

Tsinghua University

Xinjian Meng

Yun He

Jianhua Zheng

Jianwen Chen

Ping Yang

Wei Yu

Xiaozhen Zheng

Xiaozhong Xu

Shujuan Shi

Gang Zhu

A method and apparatus is provided for encoding/decoding an image in a process of transmission. A plurality of frequency bands are divided into different frequency zones; a scaling value is assigned for each frequency zone; a scaling matrix is established with coefficients comprising the scaling values for the different frequency zones; wherein a number of the frequency zones is smaller than a number of coefficients in the matrix, and each frequency zone corresponds to one scaling value in the coefficient matrix such that every coefficient in said each frequency zone has the same scaling value that is assigned to said frequency bands; and corresponding quantization/dequantization is performed in process of encoding/decoding respectively using the scaling matrix.

Quantization method and apparatus in encoding/decoding

An apparatus at an encoding end and an apparatus at decoding end are provided for implementing quantization in encoding and decoding an image in a process of transmission. The apparatus at an encoding end includes a memory storage and a processor coupled to the memory that executes the instructions to perform quantization using a coefficient matrix on an image, the coefficient matrix having a plurality of coefficients which being divided into a plurality of frequency bands, wherein a number of the frequency bands is smaller than a number of coefficients in the coefficient matrix; a scaling value is assigned that every coefficient in said each frequency band has the same scaling value; generate a parameter set having multiple band parameters thereof and write the parameter set into a bitstream to be transmitted. An apparatus at a decoding end performs decoding process thereof.

A method and apparatus is provided for implementing quantization in encoding/decoding. Firstly, the quantization on the picture transformed coefficient matrix is performed using a quantization parameter model, that is, the coefficient matrix is modeled as a parameter model represented by only several parameters, in which the parameter model includes two sorts of parameters, frequency band distribution parameters and frequency band weighting parameters. Thus, a quantization matrix conforming to the human vision system can be obtained by controlling the model parameters and the subjective quality of the encoded picture can be easily controlled. Secondly, the method does not need to store or transfer the quantization matrix in the bitstream which help to improve the encoding efficiency. Finally, the quantization on picture transformed coefficients using the method is well adapted to the content information of the picture to be encoded. Therefore, the present invention can improve the subjective quality of the encoded picture.

A method of obtaining a disparity vector and its encoding/decoding in the multi-view coding process is disclosed. The present invention includes essentially: determining a disparity vector between two views during the multi-view image coding, and computing a disparity vector between the other two views according to the disparity vector between the two views and the known relative location information between each of the views. Thereafter, the disparity vector is used for multi-view encoding/decoding. The present invention further makes use of the correlation between the disparity vector and depth information of the spatial vector on one hand, and on the other hand makes use of the direct relationship between the disparity vector and the location of each of the cameras. It is experimentally proved that the disparity vector between several views can be accurately computed during the multi-view coding, thereby improving the performance of multi-view coding.

Method and device for multi-view encoding/decoding using disparity vector

A method and apparatus for performing arithmetic encoding/decoding in an arithmetic encoder is disclosed. It comprises: firstly, when performing binary arithmetic encoding/decoding in the arithmetic encoder, an interval update process for arithmetic encoding/decoding to be performed in the original domain is mapped to the logarithm domain, and the interval update computation is implemented with direct add/subtract operations; during arithmetic encoding/decoding in the logarithm domain, data conversions between the original domain and the logarithm domain are performed by approximations; and after encoding/decoding a symbol, its probability is updated when necessary, the update being performed with add/subtract operations in the logarithm domain. It can be seen that with the arithmetic encoding/decoding solutions of the disclosure, the arithmetic encoding/decoding process and the probability estimation involved therein are mapped to the logarithm domain. Data conversions between the original domain and the logarithm domain are performed by means of approximation equations. Accordingly, complicated computations and table looking-up are avoided effectively. Therefore, the disclosure provides a simple, efficient and suitable solution for performing arithmetic encoding/decoding.

Method and apparatus for realizing arithmetic coding/decoding

A binary coding and decoding method and apparatus for MVD (Motion Vector Difference) absolute values, includes: dividing symbols for MVD absolute values to be coded into a plurality of subsets according to probability distribution properties; assigning a binary codeword for each subset; assigning a binary codeword for each symbol in each subset; concatenating and outputting the binary codeword for each subset and the binary codeword for each symbol in the subset, as the binary coding result. At the decoding side, decoding operations are performed by using a principle corresponding to that at the coding side. In this manner, the invention makes full use of the source properties on one hand, and effectively prevents the codeword length from being too long on the other hand. It has been experimentally testified that a better compression effect may be achieved during arithmetic coding process and thus the image coding performance may be improved effectively.

Growing community of inventors

Shenzhen, China

Xinjian Meng