Combined Robust Algorithm for Image Retrieval for Minority Inheritance

Authors

  • Chigurupati Navyasree, Dr. Dhulipalla Venkata Rao,

Abstract

There is huge requirement for image retrieval technique which can retrieve similar image from available database. This algorithm comes under computer vision task. In this proposed method we used combination of 2 different techniques to get robust results and improvement in accuracy. The retrieval of people images that comes under minority is an important task for government and private organization as it is a difficult task if we do it manually. In this combination we used color image histogram and with it we used edge orientation histogram. The first step of the project is divide image into sub-blocks. The final feature vector obtained is combination of both features provided by the 2 techniques. The feature vector obtained for query image is compared with available database feature vectors with the help of Canberra Distance. Proposed work is designed for minority image database which is obtained from internet source. Proposed work focus on some important features of an image such as color, shape, texture and image orientation. Proposed design is compared with many states of art techniques to show the improvement in it.

Published

2020-11-01

Issue

Section

Articles