Lifted wavelet-based image dataset compression with column random access for image-based virtual environment navigation

摘要

A lifted wavelet-based compression scheme with effective image-column random access for coding image datasets is proposed. This image dataset compression scheme is based on the approach of a lifted wavelet transform combined with embedded block entropy coding. This approach is adapted to compressing the image datasets used for image-based virtual environment navigation and facilitating a column random access mechanism. A specific data structure is designed to achieve efficient compression and accommodate column random access. The code stream syntax is constructed to efficiently assemble bit streams corresponding to the designed data structure. The two-level index tables in the code stream are used to assist random access. This image dataset compression scheme features a proper compromise between the coding efficiency and the random access flexibility, and thus ensures fast interactive image rendering. It reduces the decoding delay, simplifies the implementation of a random access mechanism and easily facilitates real-time image decoding. Experimental results of the software implementation of this image dataset compression scheme show its superior coding performance over the standard JPEG and JPEG 2000 schemes.