发明名称 METHOD FOR CODING IMAGE SIGNAL
摘要 PURPOSE:To reduce the coding variable of a quantizing value without deteriorating picture quality by comparing a threshold value with the absolute value of an orthogonal transformation coefficient, and when the coefficient is smaller than the threshold value, executing processing for converting the coefficient to '0'. CONSTITUTION:A forecast errors from a forecast image are divided into blocks, orthogonal transformed for each block, each of the obtained orthogonally transformed coefficients is quantized by a previously determined quantization step group Q, and then the quantized value is coded. Each quantization step for each orthogonally transformed coefficient in the group Q is multiplied by a value (a)>1/2 to find a threshold group SA, the absolute value of each orthogonally transformed coefficient is compared with the threshold value of its corresponding threshold group SA, and when the former is smaller than the latter, the coefficient value is set to be '0'. In the process for coding the forecasting error of an image formed between reference images, coding errors are not accumulated. Consequently, the coding variable of a quatized value can be reduced without deteriorating the picture quality.
申请公布号 JPH0410789(A) 申请公布日期 1992.01.14
申请号 JP19900114893 申请日期 1990.04.26
申请人 VICTOR CO OF JAPAN LTD 发明人 ANDO ICHIRO
分类号 H04N19/50;H04N19/132;H04N19/136;H04N19/176;H04N19/196;H04N19/503;H04N19/60;H04N19/61;H04N19/85 主分类号 H04N19/50
代理机构 代理人
主权项
地址
您可能感兴趣的专利