METHOD AND DEVICE FOR MULTI-VIEW-POINT IMAGE CODING AND DECODING

摘要

PROBLEM TO BE SOLVED: To encode a multi-view-point image at a high efficiency and to decode an optional block of an optional multi-view-point image in real time in the case of coding/decoding the multi-view-point image. SOLUTION: A multi-view-point image input means 101 gives images resulting from photographing a three-dimensional object from a plurality of points respectively to a multi-view-point image coding means 102, the means 102 divides the multi-view-point image into m×n pixels to design a code book 103 and vector quantization is applied to each multi-view-point image to obtain index data 104, a multiplexer means 105 multiplexes the designed code book 103 and the calculated index data 104 to generate coded data 106. A demultiplexer means 107 demultiplexes the transmitted coded data 106 into a code book 108 and index data 109, a multi-view-point image decoding means 110 references a representative vector in a code book 108 corresponding to the index data 109 to decode an optional block of the multi-view-point image from a required view point.