Research On CAM Angiogenesis Image Processing Algorithm

Angiogenesis is a process of growing new blood vessels from existing blood vessels. Angiogenesis not only plays an important role in the normal physiological changes, such as growth and wound healing, but also has a close relationship with the mechanism of cancer. Therefore, angiogenesis is important for the treatment and prevention of cancer. We usually evaluate level of vascular growth by calculating length, area and branch number of arteries in certain area.As the experimental material is easy to obtain, experimental period is short, simple operation and results are easy to observe. Therefore, CAM is often used as the experimental carrier in the study of angiogenesis. In view of the low contrast of the target and the background area, the local blood vessel breaks, and the difference of the gray value of the blood vessels in the CAM vascular image, through the Otsu threshold segmentation algorithm combined with OPTA thinning algorithm to obtain the initial seed points, then using the iterative threshold tracking algorithm to segment image and extract vessel information. Comparing with the traditional single threshold segmentation method and edge detection segmentation method, the iterative threshold tracking algorithm can be used to segment CAM blood vessels in the image of small vessels, and detected the vascular information more accurate and completely. Skeleton information extraction directly affects the quantitative of the follow assessment. For CAM blood vessel image after the division has a width varies, branch points, etc., the vascular direction is determined by the Hessian matrix and vascular skeleton point information is obtained by the discrete Gauss kernel function convolution CAM vascular image.By contrast, the skeleton extraction algorithm in paper is superior to the traditional thinning algorithm, and has fine connectivity and axis.In order to see the effect of each processing module more directly, the GUI system is designed. Compared with the experimental results, we can see that the gap between the result of this work and that by manually labeled by experts is small enough, and can be used repeatedly.