Image coding with polynomial transforms

Compression schemes for digital audio, video and images are necessary to accommodate the increasing digital data traffic. Desirable features of such compression schemes are good perceptual quality, flexibility, higher compression ratios and fast decoding time. This paper focuses on low bit rate compression of digital images. We propose an incomplete orthogonal polynomial transform that can be implemented by a multiplication-free architecture. A subband coder is employed to quantize the transform domain coefficients so that the spatial correlation between the coefficients of the neighboring blocks can be exploited. The proposed approach improves the perceptual quality and the rate-distortion performance.