Research On The Algorithm Of Artificial Neural Network For Segmentation Of Blood Cells Image

With the development of computer technology, computers are widely used in medical study field. Among all these applications, the computer assisted surgery and computer assisted diagnosis are getting more and more attention. Especially the processing of medical microscopic images, the work is very tired with low efficiency in the traditional method. So the members of the examination often have subjective observation errors and have been had higher request to the technique level. It can only record the images by microscopic photograph and not process the image in necessity, and it can't save and recur the image quickly. Even it can't deliver the image long-distance by the network, so it isn't fit for the new request of the information age development. Now we can show the color digital medical microscopic images by the modern light-electricity conversion, hardware of the computer, data and image processing software etc. To reduce work intension and assist diagnosis for doctor, the blood cells are recognized by the specific request and analyzed on the quantitative colors and morphology in the computer. The most important part of this pager explored and proposed BP neural network method on the digital image processing. Based on the method, the color cells auto-counting system is developed by VC++.Chapter two introduces the basic concepts and expounds several familiar models about artificial neural network in detail. Then makes overall discussion and gives the improving arithmetic on the BP neural network. Chapter three narrates the basic knowledge of the digital image processing technology, and sets forth the color image processing methods in particular, then analyzes the application of color and morphology in image segmentation. The effects of threshold segmentation are compared in different color space. And the method is brought out from two aspects of color and morphology. Chapter four expounds the choosing rule about target characteristic extract and the extract method, and gives the basic content on statistic classifying theory and several statistic segmentation methods based on target area distribution. Chapter five gives the working principle and system structures of the color cells auto-counting system based on artificial neural network, then expounds the course of realization with software and introduces the operation of color cells auto-counting system.This thesis has three innovations:1. Exposed the result that the threshold segmentation in HIS color space is better than the threshold segmentation in RGB color space.2. Giving the images segmentation methods using two aspects synthesize both color and morphology, it can reduce the processing steps of the irrelevant data cost and the process is speed up on the color target segmentation.3. Realizing in color target segmentation and morphology extraction by using BP neural network model, improving on the accurate degree of image segmentation and made a easy condition for image recognition.