Interpolation approaches to vector quantization for image compression

Vector quantization (VQ) applies effectively to very low bit-rate image transmission. The high compression ratios are often balanced by limited reconstruction quality due to blockiness effects due to coarse quantization. Interpolating different codevectors can overcome such drawback by enhancing the generalization ability and the adaptiveness of the coding system at run time. Experimental results on real testbeds show the method's advantages as compared with related techniques.