In shape feature extraction, the extracted feature vector is magnitude of five types edges (horizontal, vertical, 45, 135 degree and isotropic). Also Hough Transform is used to extract the edge features. Using different types of masks, the feature vector is obtained from the original image. Similarity measurement is performed by Euclidean Distance measure. | ISSN:2249-5789 Alok Pandey et al , International Journal of Computer Science & Communication Networks,Vol 3(6),358-367 An Integrated Approach for Image Retrieval Based on Amelioration of Color Mean and Edge Detection using Novel Masks 1 1 Vasagan, PG Part Time Student, ECE Department, Anna University, Chennai 2 Assistant Professor, ECE Department, Kamaraj College Of Engineering and Technology, Virudhunagar, Tamilnadu, India sabarivasagan@ Abstract The growth of computer vision technology in all over the world has led to necessity of the efficient retrieval system. In the proposed system, amalgamation of two features such as color and shape has been implemented which are extracted by amelioration of color mean and edge detection using novel masks respectively. In color feature extraction, row mean, column mean are extracted. These feature vectors are applied on to the original, original with even and original with odd image. In shape feature extraction, the extracted feature vector is magnitude of five types edges (horizontal, vertical, 45, 135 degree and isotropic). Also Hough Transform is used to extract the edge features. Using different types of masks, the feature vector is obtained from the original image. Similarity measurement is performed by Euclidean Distance measure. The performance of the system is analyzed by precision and recall. This system provides the efficient retrieval of images with high precision rate. 1. Introduction An image retrieval system is a computer system for browsing, searching and retrieving images from a large database of digital images. There are two approaches to image retrieval. In the first conventional approach, the visually extracted attributes are manually entered into the database. This requires a considerable level of effort, and fixes a priori types of queries that can be addressed to the database. In the second approach, where the investigation is the .
