IMAGE PROCESSING UNIT AND IMAGE PROCESSING METHOD

摘要

PROBLEM TO BE SOLVED: To provide a means where dispersion errors of arithmetic accuracy of inverse discrete cosine transform IDCT in an MPEG decoder are not accumulated even when the arithmetic accuracy of the IDCT is decreased. SOLUTION: A device that decodes given image compression information is provided with a variable-length decoder(VLD) that decodes variable-length codes of the image compression information, an inverse quantizer (Q<-1> ) that applies inverse quantization to the decoding result, an inverse motion compensation device (MCDCT<-1> ) that performs inverse motion compensation in a DCT domain, and an inverse discrete cosine transform device (DCT<-1> ) that applies inverse discrete cosine transform to the result of inverse motion compensation. The variable-length coder, the inverse quantization device, the inverse motion compensation device and the inverse discrete cosine transform device conduct decoding of the image compression information in this order.