ORTHOGONAL TRANSFORMATION ENCODING METHOD FOR PICTURE

摘要

PURPOSE:To realize the duplicate block orthogonal transformation encoding method for a picture taking local property relating to the movement of a picture into account so as to utilize the basic characteristic of the duplicate block orthogonal transformation representing an excellent encoding efficiency more than that of DCT and the hybrid encoding efficiency being the combination of the duplicate block orthogonal transformation and inter-frame prediction is further improved. CONSTITUTION:An input signal series of an input picture data is given to a duplicate block processing section 2, in which the data is reconstituted into an input block whose length is 2N and in which adjacent blocks are overlapped by N/2 picture element each. Then the data is converted into N sets of transformation coefficients by a duplicate block orthogonal transformation section 3, a block discrimination section 4 compares a transformation coefficient block (in-frame block) with an inter-frame coefficient difference block (inter-frame block) with respect to a transformation coefficient block to a same coordinate as a preceding frame stored in the coefficient memory 5, a block with less sum of in-block power is selected, the block is quantized by a quantization section 6 and an entropy coding section 7 applies variable length coding to the block and an output is obtained therefrom.