Breast tumour segmentation using thresholding and canny edge detector / Fatin Rasyidah Rosli ... [et al.]

Mammogram acts as a screening tool is used to acquire images of the breast in order to detect early signs of breast cancer. However, the limitation of the mammogram images is it turn out to be too dark or too bright which endangers the loss of useful information. Numerous techniques have been introduced to improve the mammograms including quantitative evaluation. Unlike existing research that required additional hardware to be implemented in the segmentation process on the mammogram, this paper proposes an automated approach to segment breast tumours using image processing. The segmentation process is performed on the mammogram images using thresholding and canny edge detection algorithms. Thirty-three images are collected and tested. Qualitative evaluations showed that the proposed system outperformed segmented breast tumour at an acceptance rate of 52.09 percent, whereas quantitative evaluation using Area Overlap, False Positive Rate and False Negative Rate produced an acceptance rate 52.09 percent, 33.34 percent and 14.57 percent respectively. The findings could improve the quality of mammography images and help radiologists and doctors to detect breast tumours more accurate in a shorter period of time.