Medical Image Retrieval Based On Shape

With the improvement of image-base clinical technique, large amounts ofdigital images are produced everyday. The management of medical imagedatabase and how to use those images in Clinical diagnose Process (CP)becomes a especially challenge in medical field.The most important thing to improve the efficiency of traditional PACS isto introduce the CBIR (Content based Image Retrieval) technique to themanage process of medical images. Traditional PACS use a handy (orsemi-automatic) method to manage images in database. It not only runs witha very slow speed, but also reduces its clinical benefits in CDP. CBIRprovides the ability to access images by their content, which meets therequirement of PACS. So, it is obvious that we should provide thequery-by-content ability to PACS. But it is a pity that, although the need tomigrate CBIR in medical images treatment has been proposed many times, itstill in its junior stage?￡As the ICOM standard has been made for a long time, the need forremote diagnose became more and more popular, all of this challenge theaccess method of images in a rough manner. The ICOM standard also givessome convenience to CBIR, because the structure of ICOM image containsnot only original image data, but many other information for use in thediagnose process.The motive of this article is exploring the usage of CBIR technique inmanagement of medical images, and develops a CBIR system base onmedical images that can improve the access ability of medical images. Maythose work will contribute to the implement process of CBIR in medical.Image thresholding division is the most commonly used technique ,at thesame time it is the most ordinary image division technique as well.It isparticularly applied to image which the object and background occupydifferent grey-scale level range .it can not only compress galactic data, butalso can predigest the steps of analysis and management So under manyconditions ,it is the indispensable image manage progress before imageanalysis , character distilling and pattern identification . The aim of imagethresholding is compartmentalize pels concourse according to the grey-scalelevel ,every hypo-concourse form a region that corresponding to the realscenery ,each region have accordant attribute .The division method can beachieved by select one thresholding or multi-thresholding from grey-scalelevel .Liver segmentation and PCBIR identification are discussed. Imagesegmentation is one of the key techniques of medical CBIR systems. Thisdissertation presents watershed technique for liver segmentation base onmathematical morphology and deformable models. In this algorithm, firstly,the edge of the original image is followed and candidate corners in theoriginal image are detected. Then the counterpoints of the candidate cornersin the result of edge follow are found, and boundary corners whosecounterpoints have been found are queue in the order of their counterpoints inedge follow.The shape is one of most important feature for characterising an object.However, most shapes that are expressed with primitive uniform featureshave difficulty reflecting their logical and structural properties. In this paper,we propose a structural analysis scheme for the shape feature structured bylogical properties, as well as a similar retrieval method. A shape isrepresented as a set of curve segments with a specific pattern. As afundamental unit, a curve segment has adaptive features based on the logicalproperty of its pattern. The relationship information of curve segments isexpressed as a structural feature. We also use it as a feature for ?°coarse-fine?±matching because our shape features have global characteristics as astructural feature and local characteristics as an adaptive feature of shape.Our experiments show that structural adaptive features through logicalanalysis result in effectively classifying shapes according to their cognitivecharacteristics. Various experiments show that our approach reducescomputational complexity and retrieval cost.When CBIR is applied to specific field, the most important and mostdifficult problem is how to use the knowledge in the field. There are so manykinds of images that we can not expect to use a single pattern to distinguishthem. Every kind of images has its own characteristics, this is especially truewhen it was used in diagnose process. The rules we use images in diagnoseprocess is what field knowledge means. How to employ these fieldknowledge in diagnose process is the key we introduce CBIR to medical field.This is the idea of semantic based image retrieval, is the heading direction ofcontent based image retrieval, and it is the future work of this article.