Design And Implementation Of Many Characteristics-Based Image Retrieval System

The rapid development of high-speed Internet and multimedia technology and the continuous improvement of mass storage devices and processor performance have greatly promoted the popularization and development of the graphic image system. Image database is being widely used in more and more fields like digital library, medical image databases, satellite image database, trademarks and architecture database, and so on. However, how to find the necessary image information in a large number of images is now a pressing problem in the image application.CBIR analyzes the image according to visual features of images such as color, texture, shape and spatial position relation and conducts the image retrieval by establishing a variety of image feature vector databases and building a query module. Based on a wide consultation of related research and technology materials home and abroad, studies and analyzes the color feature, texture feature and shape feature, offered based on comprehensive features of retrieval methods. The experiment verifies that the texture and color, shape of the search method overcome by using a single item of localized to retrieve precision and retrieval universality.So this paper is to take many characteristics based image retrieval system which I have designed in China Youth University For Political Sciences as an example, discussing to carry out design and implementation of many characteristics based image retrieval system according to color space, color features, texture features, shape features, relevance feedback, etc.The article described the following parts:1. First Demand analysis of the image retrieving service based on collecting relevant information from students and teachers, found the main problems existing in current applications, such as indicated the need for manual labor on every picture, marked with a degree of subjectivity key words, lack of Content of those images. According to the content of the systems exist in the main question, to retrieve the contents of the image is based on solutions and define the design target.2. According to business requirements, images retrieval system needs to ensure retrieve recall ratio and retrieve precision ratio. Based on study and discussion through comparative analysis, choose the right algorithm or key technology.3. System analysis and module design of the many characteristics based on image retrieval with high cohesion and low coupling principle. Design the database structure in images retrieval system, and implement on the SQL Server database.4. Detailed design and implemented on each module the image retrieval system. The paper obtained the image color feature in the human visual characteristics consistent with HSV color space, and extracted texture feature and shape feature based on combining the grey level grows matrix and HU-based of seven invariant moments algorithm. By combining relevance feedback on the learning sample set, obtained the user object and image features between queries object model, and then learn the model to the new round of search. Finally, using Visual C++, built a small query system, realized the algorithm mentioned in this article, complete the basic functions of image retrieval.5. According to retrieve recall ratio and retrieve precision ratio, the image retrieve system performance standards, the paper validated the image retrieval system had greatly promoted based on many characteristics than based on a single character, and achieved the desired effect, the use of the tests for data collection system testing. In addition, the simple navigating interface and good interaction could lead the whole system to a better user experience.Finally, the paper summarizes the research on CBIR and proposes the future research in this area, based on analyzed the difficulties in the realization process.