Breast Blood Perfusion (BBP) Model and Its Application in Differentiation of Malignant and Benign Breast

摘要

Preprocessing of the TBI plays a crucial role in the computer-assisted diagnosis of breast cancer using TBIs. The raw breast thermogram often suffers from the intra-class scatter problem that makes it very difficult to differentiate a malignant breast thermogram from the benign one. Hence, in this paper, we have proposed a BBP model, which is also the primary contribution of this paper, to convert the raw gray-level breast thermogram into the blood perfusion image. The BBP model used thermal physiology of the breast and reduced the intra-class variability of the breast thermograms. The BBP model can be used as a preprocessing step for the diagnosis of the breast using breast thermogram. The experiment results have shown that the blood perfusion image of the breast is more powerful than the raw gray-level breast thermogram for the prediction of the breast abnormalities. Additionally, we have introduced a new combination of texture features for the analysis of the breast thermograms. In future, we will try to address the problem of suspicious region segmentation using BBP image, which is comparatively very hard for the gray-level breast thermal image.