Examination on Key Technologies of Segmentation and Retrieval in Medical Image Processing.

In recent years, medical imaging technology has rapidly spread in the modern medical industry, producing a large amount of medical image information. The purpose of this paper is how to combine the key means of image processing with medical images, providing doctors with efficient and convenient means of medical image search. This paper studies medical image segmentation, extraction and representation of feature, index and related retrieval techniques. Aiming at image segmentation, a PCNN image segmentation method based on canny edge detection is proposed. Based on the traditional gray histogram feature extraction technique, an adaptive weighted improved gray histogram method is proposed, and it is proved by experiments that this method can enhance some important features of the image. It is easier to calculate the similarity and help the doctor to find the image features of interest in the complex learning image. For the image indexing technology, used the artificial neural network BP algorithm to classify the image, proposed a diffusion matching algorithm based on image segmentation. In the corresponding feedback, the rough set and the support vector set can be better combined, and introduced corresponding feedback techniques to retrieve the image. Make full use of the rough set of big data processing, the benefits of redundant data's reduction, which can reduce SVM training data, can not only improve SVM classification efficiency, but also improve retrieval efficiency. Experimental results show that the improved corresponding feedback method has certain advantages in feedback accuracy and feedback speed. [ABSTRACT FROM AUTHOR]