Jehad Q. Odeh Alnihoud, Department of Computer Science, Al-al Bayt University, Mafraq, Jordan.
Abstract— The need for automatic feature extraction and comparison has become one of the most important research topics in image based retrieval. That due to the rapid increases in images databases and the difficulties of indexing images based on textual description. CBIR (content based image retrieval) utilizes automatic feature extraction based on color, texture, and shape using image analysis and processing techniques. In this paper a novel two-leveled CBIR system is proposed. In the first level, spatial domain is considered with color features extraction and comparison. While, in the second level a hybrid feature based approach is deployed. In which edge detection and enhancement, morphological operator (Dilate), and 2D-DWT (wavelet transform) are used. WANG database of 1000 images form 10 categories is used to test the proposed system. Moreover, extensive comparisons with the most related systems are conducted and the result are better than the other compared systems.
Keywords-CBIR; Edge Detection; Color Moments; Morphological Dilate and 2D-DWT.
Citation: Jehad Q. Odeh Alnihoud, “Image Retrieval Model Based on Color and Shape Features Utilizing 2D Wavelet Transform”, The World of Computer Science and Information Technology Journal (WSCIT). 2017 Volume 7, Issue 4, pp. 20.25.