Abstract:
Breast disease occurs in the divided area with the nipple as the center. Accurate positioning of the nipple has important clinical value for the diagnosis of breast disease. In this paper, an algorithm for the nipple location automatic detection is proposed based on automated 3D breast ultrasound (ABUS) coronal images. Firstly, a mask is applied to find the region of interest (ROI). Then, the image pre-processing operations are used to improve the accuracy and efficiency of the target area identification. Finally, central coordinate and radius of the nipple are obtained by using the circle Hough transform and the exclusion of misjudgment circles. It is shown in the results that the proposed algorithm can effectively detect the location of the nipple, and the detection accuracy can achieve 94.7%.