Scalable compression based on tree structured vector quantization of perceptually weighted block, lapped, and wavelet transforms

This paper presents an algorithm for scalable compression using tree structured vector quantization (TSVQ) of perceptually weighted block, lapped, or wavelet transforms. The algorithm produces an embedded bit-stream to support decoders with various spatial and temporal resolutions. Bandwidth scalability with a dynamic range from a few kbps to several Mbps is provided. The algorithm further supports decoders with varying alphabet size, computation, memory, latency and power requirements. The embedded bit-stream produced is prioritized with bits arranged in order of visual importance. The algorithm also allows easy joint-source channel coding on heterogenous networks. The subjective quality of compressed images improves significantly by the use of perceptual distortion measures.