Color Spaces for Transform-based Image Retrieval

The discrete image transforms are used for energy compaction primarily and so used in image data compression. The level of energy in the image depends on level of colors used. In this paper we use two discrete image transforms namely Discrete Hadamard Transform (DHT) and Discrete Wavelet Transform (DWT). These transforms are applied on two different color models namely HSV and YCbCr separately in a given large standard database with 1000 images formed from 10 different classes taken from the Corel collection. The proposed features are effective and useful for image indexing and retrieval.