Improvising Breast Cancer detection using CNN, VGG and SSD Algorithms
S. Ruban, Mohammed Moosa Jabeer, Ram Shenoy Basti · 2022 Third International Conference on Intelligent Computing Instrumentation and Control Technologies (ICICICT) · 2022
Healthcare is one area where artificial intelligence technology is having a rising impact. In the early stages of cancer, medical imaging is critical, to increase the effectiveness of cancer diagnosis. With an alarming rate of one Indian woman being diagnosed with breast cancer every four minutes, breast cancer is on the increase among women. Despite the fact that breast cancer is a curable disease, early identification is vital to a favourable prognosis. Until recently, mammograms were the only way to identify breast cancer. Mammography, however, fails to detect breast cancer in women of all ages, leading to a high rate of false positive and false negative instances. This has dire repercussions. This experimental study discusses about using several Deep Learning algorithms to identify the cancerous area in mammograms. Three deep learning algorithms—Convolutional Neural Network, Visual Geometry Group, and Single Shot Detector—were applied in this research. This experiment, which had already been approved by the scientific and ethical authorities, was run on a Real Time Data collection of mammograms acquired from a 1250 bed hospital. There are 90 patients, mammography images in all, which are classified as benign or malignant. The accuracy of the Single Shot Detector method was 74%, which was greater than that of the other deep learning algorithms CNN (64%) and VGG (60%). Findings suggest that these deep learning algorithms can tremendously help the radiologists to improve their findings with respect to accuracy and precision.