Query implementation technique for large image databases

An image indexing technique using wavelet decomposition and clustering approach, which can be employed for retrieval of images from an image database, is presented. An algorithm incorporating image indexing on the clusters of regions has been developed. This method can be employed for region-based querying of image. The querying method described in this paper has applications in different domains, including graphic design, multimedia, geology, satellite imaging, medical imaging, defence, etc. Some experimental results obtained for different feature sets using Daubechies wavelet transform and Haar wavelet transform have been presented.