Support vector machine image segmentation algorithm applied to angiogenesis quantification

Angiogenesis is an interactive process of creating a network for oxygen and nutrients supply among tumor, endothelial and stromal cells, and this phenomenon is necessary for tumor growth. The angiogenic is usually estimated by counting the number of blood vessels in particular areas. One of the most popular experiment model to study the angiogenesis phenomenon is developing chick embryo and its chorioallantoic membrane (CAM). We present a new image segmentation method using support vector machine to segment the preprocessed image. Then we fill and extract the skeleton of the image segmented, apply image automatic counting software that gives an unbiased quantification of the length and branching points of angiogenic CAM images´ micro-vessels. Experimental results demonstrate that the proposed image segment algorithm based on support vector machine is effective which is able to provide reliable quantitative analysis data of CAM model. Moreover, the analysis speed, measurement accuracy and data repeatability have an advantage over manual expert assessment.