Quantitative Analysis of Vascular Structures Using Image Processing

Vascularization, the growth of new blood vessels form the existing vessels, implies many pathological processes and needs to be reasonably quantified. However, most vascular analysis is done manually. This is a tedious and laborious work without consistence. In this paper, we will demonstrate the feasibility of automatic quantification of vascular structures by image processing. To quantify the formation of blood vessels, the following parameters are measured automatically from the processed images: total length of tubes, total number of loops, total tube area, total confluent areas, the number of confluent area and the number of nodal structures. In addition, the obtained number of loops and tube length are compared with the values measured manually. The experimental results show that the highest Hit Rate is 90.3% and the highest False Alarm Rate is 6.675%.