This paper researches on the issue of how to detect the diagnostic lesion of breast cancer from near-infrared images. Firstly, it sets up the imaging model after studying the way of how near-infrared images come into being. With this model, the image restoration is conducted to eliminate the background and achieve better quality. A new method, which utilizes the imaging features of blood vessels, is also proposed to detect blood vessels based on enhanced images with poor quality. The paper also studies P. K. Saha and Jayaram K. ' s scale-based fuzzy connectivity methods of how to segment dense tissue regions from fat within breasts. Finally, classification of breast diseases based on near-infrared images is studied, and the way to distinguish blood vessels from tumors is given, and then discussion is conducted on how to determine the property of different tumors based on the feature vectors. |