The Application Investigation Of Image Retrieval Technology Of Medicinal Image Database

Life science is the most important domain of science research in twenty-one century. Biomedicine which is corporate with information engineering is main research means. It is called digital and modern medical treatment by applying information and computer technology to medical treatment. For adapting for the trend of digital medical treatment, we plan to build the medical image processing system which can put the actions of collecting, showing, depositing and transmitting the image together. A lot of medical pictures of patients' physiology, pathology and dissecting are included in the system. So it is an important and objective tool for doctor carrying on clinical diagnosis, condition following, surgical operation plan, study, and identify diagnosis. However, the huge and miscellaneous medical pictures need efficient medical picture research and manage methods providing maximum information. So I study the problem of how efficient storing, management and retrieving the huge medical image data and put content-based image retrieval, database and multimedia technologies together for the study.From software structure and system function, this thesis gives three layers medical image database model and the devices of medical image database from analysis of medical image characters and its functions and tasks. By the study of technology quality, job process ,image extractive quality and similarity of QBIC and comparing with new retrieve technology based on color, figure, texture quality and space relationship ,I give a retrieve model joining key words with image content based on practice. The model extracts the scope of image retrieving first, and then inquires exactly by content-based image retrieval. In the inquiring of content-based image retrieval, it gives the retrieving method of based on HSI color model additive vertical diagram and putting wavelet modulus maxima and moment invariant feature together, and analyses the good quality of additive vertical diagram. What is more, it improves on separation of color similarity of local additive vertical diagram and mends moment invariant feature, then it gives the reality of relative arithmetic and result analysis.In the system, I give the diagnosing report model based on dribs and drabs language. So doctors can write immediate diagnosing report. The system can manage the information database and typical illness case by knowledge. At the same time, user can inquire by methods of condition, content-based image retrieval and so on. The system offers many statistical methods and medical information network inquiring function for easing medical study and intercommunication.