Wavelet transform on pixel distribution of rows & column of BMP image for CBIR
Soham Pinge, R. C. Patil · 2009
Retrieving image from large & varied collections using image content(such as color, shape, texture) as a key is challenging & important problem. This paper describes a novel & effective approach to Content Based Image Retrieval (CBIR) that represent each image in database by a vector of feature values called "Wavelet Transform on Pixel Distribution of Row & Column of BMP for CBIR". Here we propose a simple and effective approach that can be easily implemented in a programming language. In this technique standard deviation & mean of wavelet coefficients of color distribution of row & column used as a feature vector of image. We use Harr wavelet for this purpose. We obtain compact feature vector of size 24 that create image signature in term of both texture & color. We use simple Euclidean distance to compute the similarity measures of images for Content Based Image Retrieval application. This technique gives acceptable results in a simple and fast way.