An Efficient Re-rank and Fuzzy based Color & Edge Feature Extraction for CBIR

摘要

Recently, feature extraction methods are in require today for Content Based Image Retrieval (CBIR) and object recognition applications. In previous decade, large database of image sets has grown quickly and will continue in future. Querying and Retrieval of these images in efficient way is needed in order to access the visual content from huge database set. Content based image retrieval (CBIR) gives the explanation for competent retrieval of image from these huge image databases the new propose system attribute is called “Edge Directivity Descriptor and Colour” and integrates in a texture information and histogram colour with reranking feature. CEDD feature extraction development consists of a HSV colour twostage fuzzylinking algorithm. This descriptor is apposite for correctly retrieving images even in deformation cases such as bend, smoothing and noise. Imperative feature of the CEDD is the low computational power needed for its extraction, in association to the requirements of the most MPEG7 descriptors. The researchers are makes using WANG database which consists of 1500 images from 10 different classes. Experimental result explains that the proposed approach execute better in terms of precision compared to other existing methods.