VBIRS-Visual based Image Retrieval System for Generic Web Image Database

In this paper, we discussed Visual Based Image Retrieval System to retrieve set of relevant images for the given input image from the large generic image database. We proposed HSV color space model and Haar transform to extract color and texture features. The images are transformed into set of features. These features are used as inputs in Self Organizing Maps (SOM) to train the network for generate the code word. The advantage of SOM is able to preserve topology structure. The cosine similarity measure is used to retrieve similar images with new representation. The experimental results are evaluated over a collection of 10,000 general purpose images to demonstrate the effectiveness of the proposed system.