To enhance compressibility of no-loss compression in a camera signal processing.

A camera signal processing part 200 converts an RGB signal inputted from an imaging device, in which pixels corresponding to three primary colors are Bayer arranged into an YCbCr signal, and outputs it to a camera signal compressing part 100. An optimal adaptive function for making correlation between referable pixels around an object pixel a maximum is calculated, and a color image signal is bisected based on the optical adaptive function by an optimal adaptive no-loss separation converting means 110 of the camera signal compressing part 100. DPCM encoding is performed by calculating a predicted value, so that predicted errors of a prescribed object pixel for prediction corresponding to a signal of low-band components among signals with separated bands becomes a minimum by a DPCM encoding means 120. A sub- band conversion encoding signal and a DPCM encoding signal are made into entropy codes by an entropy encoding and compressing means 130, and a compressed image signal is generated.