A Novel Method For CBIR Using ACO-SVM With DTCWT And Color Features

Content Based Image Retrieval (CBIR) has been the major thing for looking associated content material on the net in the form of pictures, knowledge and so on. A retrieval approach which organization color and texture feature is future in this paper. In this research, implement a new technique for CBIR by ant colony optimization (ACO) and Support vector machine (SVM) with dual tree complex wavelet transform (DTCWT) and color features. In this work, has been used color or texture features for CBIR. In the increase of feature extraction, firstly extract texture feature using DTCWT for resolving the problematic of redundant CWT. After that, extract color feature using dominant color descriptor (DCD) on RGB and HSV color space for improving computation and efficiency. The experimental dataset contain 1000 images including horses, elephants, food, African people, etc. The equal dimension is calculated using weighted Euclidean distance (WED). For improving effectiveness of the system, used ACO for the greatest result and classify data using SVM. The performance analysis is based on precision and F-measure. (c) 2017 Published by Elsevier Ltd.