CBIR using Simplified Wavelet-Based Color Histogram

Abstract

Content Based Image Retrieval System (CBIR) is the technique which enhances the advance image graphics technology by taking image content to seek images from bulky image databases. The methodology adopted in the concept has been simulated using Matlab 7.5 and Matlab 11a and tested on the WANG dataset. The search mechanism used here is based on resemblance rather than the perfect matching. We have followed the image retrieval technique, as described in the Simplified Wavelet-Based Color Histogram (SWBCH). Each randomly selected query images of each category returns the similar images from database record, and the calculated precision values. This paper shows retrieval system’s average precision is (0.654). The effectiveness of simplified wavelet based color histogram method is calculated by taking different query images from the dataset. The time taken to do the whole processing of the 1000 image database takes approximately 22 – 26 seconds.