Research On Techniques Of Segmentation-Based Medical Image Retrieval

Currently,doctors in the study of the diagnosis of disease and treatment are inseparable from the reference and use of the information of the medical images.The update of imaging techniques of medical Images and the widely use of the equipments make the number of the medical images grow at an alarming rate all the time, and the size of the database of medical images also expand rapidly.Currently,how to use of the information of medical images stored in the images database effectively is one of the problems the medical field is facing.To get the information of medical images effectively can not be separated from the image retrieval.The traditional Text-based Image Retrieval on the description of the information of images was confined to some limited text information. The text information includes the name of the image,some other artificial notes and so on.But people's understanding of the image is subjective,The description of the image with the limited text information is far away from the people's understanding of the image. Therefore,the result of Text-based Image Retrieval was unable to meet the requirements of most of the users very well.To solve this problem,Content-based Image Retrieval has been proposed. Content-based Image Retrieval is based on various features of the image to describe the content of the image.Compared with the limited text information,the description of the image with various features of the image is almost as same as the people's understanding of the image.First,the article make a describption of research background,current situation and research significance of Content-based Image Retrieval and a few representative retrieval system at home and abroad briefly.It shows the purpose and significance of the research of Content-based Image Retrieval.This article also describes the characteristics of medical images.These characteristics of the medical images have caused many great difficulties to application of Content-based Image Retrieval in the medical field.This problem shows the need of the research of the application of Content-based Image Retrieval in the medical field.Solving these difficulties and promoting the application of Content-based Image Retrieval in the medical field are the purposes of this article.Content-based Image Retrieval is the combination of the technologies in many areas. There are many relevant technologies.In this article, Several of these relevant techniques are briefly introduced,including Image Segmentation,Feature Extraction,Similarity Measurement,Database Indexing,Image Matching,Retrieval Result Evaluation Criteria, Relevance Feedback,etc.Feature Extraction is to describe the content of the image. And Feature Extraction is one of the key technologies of Content-based Image Retrieval.Because only with the accurate description of contents of the image the retrieval system can make a clear understanding of the requirements of the query of the users and retrieve in accordance with the requirements of users.Similarity Measurement technique is related to measuring the sample image and the medical images in the database.The proposed High-dimensional Indexing technique is an improved database indexing technique with good application.How to develop an objective and fair Retrieval Result Evaluation Criteria is also discussed in this article.The use of Relevance Feedback technique allows users to evaluate the results of the search. Evaluation results will be returned to the computer. It allows the system to learn the results of the evaluation of the search results to make the search results better.To be the basis of image processing,image segmentation is able to combined with Feature Extraction to describe the image content accurately.In this article, we make a detailed description of the concept and classification of image segmentation and list some more suitable techniques for the segmentation of medical images.One of the innovations of this article are to propose an improved watershed algorithm.And applying it combined with the image retrieval achieves good search results.Finally,the article makes a full summary of the work,and points out the directions for future research.