期刊名称：International Journal of Soft Computing & Engineering
电子版ISSN：2231-2307
出版年度：2011
卷号：1
期号：5
页码：122-131
出版社：International Journal of Soft Computing & Engineering
摘要：In this paper, we investigated image retrieval based on image content, Content Based Image Retrieval (CBIR) and proposed a framework to characterize the image content and similarity between the images. Our paper discusses the CBIR problem and the solution. Due to the enormous increase in image database sizes, the need for the development of CBIR systems arose. Firstly, this paper outlines feature extraction methods for color and texture. The extracted features for each image in the database used as the basis for similarity between the images. We built user interface based on Java, in which user can easily select query image and view top ten retrieved images based on decreasing order. We extended our approach to sub image retrieval also. Our results report that HSV based color features and contrast based texture features outperform than RGB based features and results reported in the paper are convincing
