Research On Medical Image Processing And Information Retrieval Technology

In the recent years, the medical imaging technology is widely used. Along with rapid development of medicine equipment, amount of medical image information have be produced. It becomes a problem that how to find the target image efficiently and rapidly in large-scale image database. There are lots of images, graphics, videos, audios and cartoon besides that simple text information on Internet, therefore content-based multimedia retrieval techniques emerge. The content-based image retrieval techniques have been applicated in the fields of medicine, teaching and scientific research very importantly. How will the medical image retrieval and image combined for physicians to provide convenient and accurate image search tool for the diagnosis and the provision of complementary proposals, a the target of study in this thesis is using image retrieval technologies in medical image retrieval and giving some more help to doctor to retrieve medical images and give diagnosis exactly.According to the key technology of image processing technology and image retrieval techniques, including the medical image segmentation, feature extraction, image matching technology and related feedback, the thesis gives a lot of studies. In the aspect of image segmentation, a Canny PCNN edge detection based on image segmentation method is put forward. In the aspect of image feature extraction technology, two aspects of work mainly are finished. In the traditional graylevel histogram based on feature extraction technology, a method of adaptive weighted improvement graylevel histogram are proposed and proves that this method can make an important part of the image characteristics to strengthen and make the similarity calculation easier. In this thesis, by using the rough sets and support vector machine theory, an improved algorithm on the rough sets and SVM related feedback for image retrieval is put forward. Because of the advantages of eliminate redundant information, we can use the rough sets theory to process the large amount of data and reduce the training data of SVM. This method can improve the ability and efficiency of the classical SVM. Experimental result shows that the improved method can improve the speed and accuracy of related feedback obviously.Based on the above key technologies, a content-based image retrieval system is designed and implemented. The result shows that the system has good effects and efficiency.