Applications of Deep Learning (DL) Techniques in Detecting Breast Cancer and Malignant Cells

According to the most recent statistics, the most common type of cancer globally is the breast carcinoma and kills close to 900,000 people annually. Early and accurate diagnosis of the illness can increase the likelihood of successful treatment and lower the mortality rate. In fact, an early diagnosis can help stop it from spreading and prevent the premature victims from getting it. Researchers who study cancer have a number of difficulties when attempting to differentiate between benign and malignant tumors as well as attempting to make judgments about benign and metastatic breast carcinoma. Examine the effectiveness of automated deep learning algorithms at identifying malignant cells in women’s breasts and cancer stage. This paper suggests applying deep learning algorithms to whole-slide pathology images in order to possibly increase diagnostic efficacy and accuracy. The convolutional neural networks (CNN), sparse auto encoders (SAE), and stacked sparse auto encoders are illustrations of techniques of deep learning were used in this research work. There are numerous public mammographic databases available. The methods discussed in this paper are put to the test using the mini-MIAS mammographic database. The stacked sparse auto encoder performs better, this method has to be tested in a clinical setting before being used. It has higher accuracy and precision as compared to CNN and SAE. Better diagnostic performance was achieved by several deep learning methods. Deep learning algorithms are used to better reliably identify tiny tumors while detecting breast cancer via a mammogram.