30 January 2003 Joint image compression and indexing technique using wavelet transform
Author Affiliations +
In this paper, a joint image compression and indexing technique using wavelet transform is presented. For compression, scalar quantization and classified vector quantization are applied in the wavelet domain to remove redundancies from different sub-bands according to their distinct characteristics. For indexing, two statistical feature vectors are constructed directly from compression outputs (quantized sub-band data before entropy coding), which facilitate a hierarchical (coarser to finer) indexing procedure and achieve image indexing in the compressed domain. Experimental results show that the joint technique performs with equal effectiveness as either compression or indexing standing alone, while the computational cost for decompression is greatly reduced (only entropy decoding is needed). Thus, the advantages of this joint (dual) image compression-indexing technique and its feasibility for online distributed image retrieval in the arena of exploding networked image applications are demonstrated.
© (2003) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Hai Wei, Hai Wei, David Y. Y. Yun, David Y. Y. Yun, } "Joint image compression and indexing technique using wavelet transform", Proc. SPIE 4793, Mathematics of Data/Image Coding, Compression, and Encryption V, with Applications, (30 January 2003); doi: 10.1117/12.453522; https://doi.org/10.1117/12.453522

Back to Top