Detection of Breast Tumor in Mammograms Using Single Shot Detector Algorithm

The influence of AI over Healthcare is increasing day by day. Lot of efforts are being made to increase the efficiency of Cancer diagnosis at the early stages, where medical images play a vital role. Among the women, Breast cancer is on a steep rise with an alarming statistics of, Indian woman being diagnosed with Breast cancer, for every four minutes. Though Breast cancer is a curable disease, identifying the problem early is an important metric for the successful outcome. Earlier days, Mammograms were the only way of detecting Breast cancer. However, Mammogram is not effective for women belonging to every age, and hence yields to many false positive and false negative cases. This leads to disastrous results. To overcome this limitation, this experimental study tells, about using one of the Deep Learning algorithm called Single Shot Detector algorithm to detect the cancerous portion in the Mammograms. This experiment was done in the Real Time Data set of mammograms collected from a 1250 bed hospital, after the scientific and Ethical committee approval. With a total of 80 mammography images belonging to either category of being, either Malignant or Benign. The Single Shot Detector algorithm gave an accuracy of 74%, which was better compared with CNN (64%) and VGG (60%) deep learning algorithms.