A novel method to infrared thermal images vessel extraction based on fractal dimension

摘要

Breast cancer (BC) is a common cancer which is the second cause of death after lung cancer. Detection of BC in early stages results to increase the surviving chance. In the earlier stages of BC, new blood vessels are formed to feed a growing tumor. This metabolism and angiogenesis process produce heat which can be recorded by high-resolution thermal cameras which give the thermal pattern of the breast and vascular structures. The main purpose of this study is to extract vessels in order to survey vessels' irregularities. The first step is vessel extraction automatically. The suggested approach is based on the Gaussian match filter (MF) and its first derivation. This algorithm can distinguish the vessel and non-vessel edges. It is applied in breast thermography vessel extraction for the first time. The second step is investigating the vessel's structure by fractal analysis to differ the various kind of vessels. It is considerably indicated that the fractal dimension (FD) results in the complexity of the vessels. The more complex the vessel, the larger is FD. It is shown the mean of simple vessel fractal dimension is 1.2 and it is 1.4 for the others. It can be applied to study vascular changes to the early detection of breast lesions.