Research On The Aided Diagnosis System Of Tumour Cells

In order to change the current situation of the tumour cell diagnosis, an aided diagnosis of tumour cells based on Automated Vision Inspection theory is developed in this thesis. With a view to designing of the capture system, preprocessing of cell images, extraction of karyon, filling of karyon, and tracking of karyon contour, a great deal of study is devoted and satisfying results are achieved. The main research work of the full thesis is as follows: In order to realize automatic diagnosis, this thesis has realized the change of the microscope worktable, using automatic way to move worktable instead of manual. According to the characteristic of the manual worktable, modified transmission device, the worktable can be moved along X ,Y direction on the level on the control of computer. Tumour cells are at cell smear random, the count and the position of capturing have greater influence on the result of diagnosis. This thesis analysis the probability of capture tumour cells, and the cross-interval method of capture is betaken to capture images. The method distributes the seemly position of captured images, and enhances the probability of gathering tumour cells. Studying plenty of color cell images and comparing the characteristics of the morphology and colourometry between the normal cells and the cancer cells, establish a method to extract the karyon with the comprehensive threshold value that combined with karyon's morphology and colourometry characteristic. An extracting tumour karyon method is betaken to treat tumour cell block as a whole tumour karyon. For expressing tumour cells clearly, according to the characteristics of pels's 4-domain, the edge of the connectedness area is calculated, and the suspicious tumour cells are stamped clearly on the former image. A new sequencing thinking is putted forward. Differing from the traditional system that shows the results by sorted single cancer cells, this system sequences the images captured automatically with sorting the whole color images by their karyon's general parameter, and show the result with thumbnail images. This method can help the doctors hold the patient's condition on the whole, and enhance the correct rate of diagnosing. The experiment indicates, this device can realize the aided diagnosis of the tumour cell, can lighten the doctors working pressure, and can realize automatic measuring instead of manual, promote the popularization of general survey of tumour.