Content-based Image Retrieval Technology Research And Implementation Of The Algorithm

With the rapid development of computer technology, high-speed Internet and multimedia technology, people can get more and more image information. How to find out desired images from large-scale image databases rapidly and effectively is becoming an active hot point in the area of retrieval research. Content-based image retrieval (CBIR) is a set of techniques to solve the problem, which retrieves relevant images from image databases based on their features. It analyzes the image according to the visual features derived from its content, then establishes a variety of image feature vector databases and builds the query module to find out desired images. In recent years, CBIR is a very active research direction and has been applied to many fields. This dissertation bases on the study of related research and technology materials home and abroad, focuses on the low-level visual features of images, which are color, texture and shape. Then, a new image retrieval algorithm based on sub-block image features is presented. The image is partitioned into several sub-blocks, which reflect local information of the image, and the combination of all sub-blocks’ features describes the whole region shape. The algorithm projects the sub-block feature to a hexadecimal character, then the image feature becomes a hexadecimal string. By computing strings’ similar degree, we get the images’ similarity. At the end of the dissertation, an image database retrieval prototype system is developed to validate the research results of the algorithm, which has a good human-computer interface. After a series of experiments, some expected results have been obtained. By analyzing the results, new ideas are put forward to improve the performance of the system.