The Research Of Image Classification And Retrieval Based On Rough Set

It brings a mass of digital images with the rapid development of the computer software and hardware, Internet and the multimedia technologies. So, it has become a research hotspot how to effectively manage and retrieve large scale image database, the content-based image retrieval (Content-based Image Retrieval, CBIR) technology came into being. CBIR is comprehensive technology which combines information retrieval, image processing, computer vision, artificial intelligence. At present, the investigation of CBIR technologies is in the stage of based on low-level features. Retrieval has not high validity.This paper discusses several key technologies in CBIR, the main work are summarized as follows: 1. Introduced the background, the present stage and the future of CBIR, and some typical domestic and foreign CBIR systems; Research on the key technologies of CBIR: feature extraction, including the extraction and expression of image color, texture and shape; Usual similarity measure model, such as geometric model, related calculation model; Usual image retrieval ways, query by external pictorial example, query by internal pictorial example, query by sketch, integrated query; The system performance evaluation, simply referred to the security of the system; 2. Expatiated on Rough Set Theory and its application area in image processing, such as image enhancement, image segmentation; 3. Acquired decision attribution set by color histogram, main area histogram and average area histogram, forecast image classification using confident degree and supported degree, construct the rough sets of objects and backgrounds in images for similarity measures between image objects. Finally, it's show the feasibility of this method by doing image classification and image retrieval experiments.do simulation experiments, achieving image classification and image retrieval.