CBIR by Using Features of Shape and Color

Abstract

Geometrical Feature is a key issue in content based image retrieval (CBIR). In the prior work, various surface highlights have been proposed in writing, including literature, including statistic ethos and spectral methods. But in many cases most of them are not precisely captured. The most critical texture feature in an image called edge information. As of late, a portion of the authors on multi-scale analysis, particularly the curve-let research about, gave great chance to remove more accurate texture features for image recovery. Curve-let has indicated promising execution, anyway it was initially proposed for image de-noising. In this paper, another image include in view of curve-let transform has been proposed. We apply discrete curve-let transform on surface image and transformed images; we process the low order statistics. Images are then represented using the extracted texture features. We discuss design, implementation, and performance analysis of Tamara’s new statistical feature based image retrieval system. One of our major contributions is to propose a new scalable image retrieval scheme using shape and color based features, which is shown to be scalable to high dimensional of image data.