CVPIC Colour/Shape Histograms for Compressed Domain Image Retrieval

Compressed domain image retrieval allows image indexing to be performed directly on the compressed data without the need of decoding. This approach hence provides a significant gain in terms of speed and also eliminates the need to store feature indices. In this paper we introduce a compressed domain image retrieval technique based on the Colour Visual Pattern Image Coding (CVPIC) compression algorithm. CVPIC represents an image coding technique where the compressed form is directly meaningful. Data that is readily available includes information on colour and edge (shape) descriptors of image subblocks. It is this information that is utilised by calculating a combined colour and shape histogram. Experimental results on the UCID dataset show this novel approach to be both efficient and effective, outperforming methods such as colour histograms, colour coherence vectors, and colour correlograms.

• Type: Conference
• Year: 2004
• Where: DAGM
• Authors: Gerald Schaefer
• Keywords: Compressed Domain Image, DAGM 2004, Domain Image Retrieval, Retrieval Allows Image