Modulation Based Combination of High Level Features Generated from SCCF & Contourlet Transforms for CBIR Applications

In this research, for content based image retrieval, an outstanding feature combination based on modulation methods is presented. Including generated improved contourlet transform with spectral correlation coefficient functions features are proposed. These specifications are re-composed as frequency statistic specification, proposing spatial signals dominion. The produced features are Norm-1 values for transformation of two dimensional signals. Since the contourlet transform are more efficient compared to the other transforms, the modified modulation based recombination was used. Finally, modified features was employed for all database images. This paper utilize high pass modulation estimation, for efficient image retrieval high level features. The representations achieved from large-scale database of 10,000 images using the proposed features were applied as textural specification. Recombination of Textural features based on modulated transformation provides robust feature matrix for database image retrieval. This system illustrate higher retrieval percentage up to 90% accuracy in mentioned database. In all classes, proposed method has more than 90% accuracy. The extracted results certify the precision and efficiency of the proposed system. The results were compared with different systems and were discovered to be notable.