Research And Implementation Of Content Based Image Retrieval Technology

A large number of images come forth because of the prevalence of multimedia technology and the implementation of Internet technology. Traditional text keyword-based retrieval approach can't adapt to the demand of image data retrieval, so content-based image retrieval (CBIR) technology becomes the current research focus. In this paper, according to the hot point of the current research of CBIR, based on analyzing and studying the key techniques of CBIR, we mainly research the image retrieval algorithms based on image color spatial feature and region feature. The main content of this paper are summarized as follows:1. Some key techniques and algorithms of CBIR, such as the low-level feature descriptions including color, texture, shape, the similarity measure between images, the indexing methods, the relevance feedback and so on, are deeply analyzed and discussed .2. We Propose a color-spatial feature retrieval method based on color clustering, color clustering is introduced into color-spatial feature based retrieval method in order to decrease color quantization error, increase retrieval accuracy, and a new spatial color descriptor is proposed. This descriptor involves an adaptive binning color edge histogram and an adaptive binning color smooth histogram , increases color description accuracy. Corresponding similarity measurement for the descriptor is also proposed. Experimental results show that the method is not only more accurate than normal color-spatial feature based retrieval methods but also reduces the storage space required for the image features.3. We research and implement a region-based image retrieval method . First, the image is segmented by a segmentation algorithm using color and spatial information, and the region feature such as color, texture, shape is extracted. An unbalanced region matching method based on two level description is used to measure image similarity to reduce the influence of inaccurate segmentation. Experimental results show that the method is more accurate than normal region-based retrieval methods .4. We design a CBIR system using Visual C++ 6.0 and Access2003 and implement some retrieval algorithms. The system can not only be used for algorithm evaluation and performance comparison but also combine different algorithm to reach better results.