Picture compression process, especially of the MPEG2 type

Abstract

A picture compression process, especially of the MPEG2 type, in which each picture macroblock (10) is subjected to a coding chosen, for each macroblock, from among several types of coding. In order to select the type of coding to be applied to the macroblock, the parameter MB_activ−α_Q, is determined, in which MB_activ = ∑ 4 ⁢ luma ⁢ ⁢ blocks ⁢ Block_activ and block_active = MAX ⁡ ( Max i , j = 0 i = 6 , j = 7 ⁢ ⁢ [LeftBracketingBar] x ⁡ ( i , j ) - x ⁡ ( i + 1 , j ) [RightBracketingBar] , MAX i , j = 0 i = 7 , j = 6 ⁢ ⁢ [LeftBracketingBar] x ⁡ ( i , j ) - x ⁡ ( i , j + 1 ) [RightBracketingBar] ) , wherein Q is the quantization interval, and α is a predetermined parameter, for example equal to 1. The mode of coding selected is that corresponding to a minimum value of the parameter MB_activ−α_Q.

Dependencies are inferred from the claim text.