Content based Image Retrieval Using Support Vector Machines

Vishal Goyal and Aasheesh Shukla

Lot of image database has been created due to technological development. The images are created by the equipment used in various applications like fashion and graphic design, crime prevention medicine, engineering design and architectural, etc. To handle this image, image processing tools are required. Image retrieval is used to retrieve an image in a given database. The requirements of image processing are met by developing various techniques. One of the most useful as well as important techniques developed is called CBIR. It is an image retrieving process and it is used retrieve an image from a database with huge collection of images. The image retrieval is done based on the visual contents. Visual contents are named as features (colour, shape, texture). Retrieval of image between query image features and database images features are conducted by a meta-heuristic algorithm (genetic algorithm with Iterated local search). But it produced less accuracy in retrieval. The classifier performance may be affected by the reduction of a dimensionality of a feature and it is a challenging task. In this work, dimensionality reduction is done based on the bat algorithm in order to enhance the CBIR performance. Images are enhanced using contrast stretching process. Image retrieval is done by SVM at last.

Volume 11 | 10-Special Issue

Pages: 625-632

DOI: 10.5373/JARDCS/V11SP10/20192851