Finite state vector quantization with an interleaved image source designed for reconstruction of lost packets
CONSTANTINE N. MANIKOPOULOS, Haijiang Sun · International Conference on Acoustics, Speech, and Signal Processing · 2003
A novel interleaved-source finite-state vector quantizer (ISFSVQ) has been devised. It is a vector quantizer with memory, quite robust to packet loss. ISFSVQ assures access to a large codebook for quantizing purposes, thus achieving high image quality, while utilizing the low bit rate of a small codebook for the purpose of specifying the block label. Its main advantage, compared to other finite state algorithms, is that it requires a lookup table of small size and modest complexity. ISFSVQ results in high image quality upon image reconstruction at coding bit rates of 0.25-0.35 bits/pixel. An overall reduction in bit rate of about 0.3 bits/pixel has been achieved, compared to direct VQ, at no significant cost in image quality for many types of applications.>