The System Development Of Mdeical Image Retrieval Based On Texture Analysis

With the improvement of image-based clinical technique, large amounts of digital images are produced every day. The management of medical image database and how to use those images in Clinical Diagnose Process (CDP) becomes an especially challenge in medical field. In this situation, content-based medical image retrieval develops quickly.Extracting the characteristic of image is one of the key technologies to content-based medical image retrieval. In fact, medical images have some particular characteristic. We review adequately current methods of extracting the features of images, and then combine the imaging mechanism and special characters of medical images to research the suitable algorithms. Image texture is an important feature of image content, so the extraction methods of image texture are studied in this article, including Gray Level Co-occurrence Matrix, Wavelet Transform and Gabor Filter. All of these, we give the experimental results, and discuss the affectivity of the results by using the precision ratio and recall ratio. Then the method of medical image retrieval based on the Gabor filter and relevance feedback is proposed for the problem of medical image retrieval, which has been proved a good performance.Using the method of Gabor filter texture feature is a key research by discussing the method of medical image texture feature firstly, and then the similarity calculated between the two images is adjusted by the technology of relevance feedback to improve the performance of human-computer interaction in the image retrieval. Experiments show that this method can effectively improve the recall ratio and precision ration in the retrieval result by the relevance feedback.We have a method of texture-analysis based medical image retrieval and implement our medical image retrieval system form it. In the development of the platform, we employed object-orient techniques to make it independent to the knowledge of filed. Under the rules we have made, one can easily add additional features and matching rules. There are so many kinds of images that we cannot expect to use a single pattern to distinguish them, every kind of images has its own characteristics and each type of image methods has a regular use method. Our texture analysis-based medical image retrieval system for medical images is right developed from the platform by some such feature on it.Using the method of Gabor filter and the technology of relevance feedback will have a better retrieval precision and can also improve the flexibility of the user query.