Wavelet correlogram: A new approach for image indexing and retrieval

In this paper, a new approach for content-based image indexing and retrieval is presented. The proposed method is based on a combination of multiresolution analysis and color correlation histogram of the image. According to the new algorithm, the wavelet coefficients of the image are computed first using Daubechies3 wavelet. Then, monodimensional color correlogram of the horizontal and vertical wavelet coefficients are computed. Finally, the index vectors are constructed using these wavelet correlograms. The image retrieval results by applying the new method on a 1000 image database show a significant improvement in effectiveness and efficiency compared to the indexing and retrieval methods based on image correlogram or wavelet transform.