Research And Application Of Content Based On Image Retrival

With the continuous development of IT the landscape of visualinformation become more and more huge. At the same time, the rapidgrowth of images resourse increased enormously. How to find the requiedimage rapidly is the centre issue in the field. There are two types ofdifferent image retrieval techniques namely text based image retrieval andcontent based image retrieval techniques. Text-Based image retrieval usestraditional database techniques to manage images. Content-based imageretrieval (CBIR) uses the visual features of an image such as color, shape,texture, and spatial layout to represent and index the image. The aim ofthis paper is to review the current development of the technology incontent-based image retrieval (CBIR), a technique for retrieving imageson the basis of automatically-derived features such as colour and shape,and to propose a system of content based image retrival for shoppingonline.This paper proposed a method for image retrival, which extract thefeatures from color and shap, and this paper also found a efficient index structure for query. The paper improve the method of dominant colorextraction, which named by adaptive dominant color extraction. Themethod can reduce the dimension of the color characteristics. Accordingto the different layout of colors in each image,we can receive thedifferent number of the dominant colors. There is a deep research on theshape feature extraction methods and ideas, and select the area shapefeature extraction, which is the block Hu moments feature extraction.This approach resolves the problem of Hu moments in segementation.Finally, the system for product image retrival is developped, and someproduct images can be queried by this system.