Breast cancer detection and recognition by using deep learning techniques - ultrasound images classification in to benign, malignant and normal tissue by the resnet model

Breast cancer has become a growing concern nowadays for global countries. In 2020, there were 2.3 million women diagnosed with breast cancer and 685,000 deaths globally. It is mostly associated with women. In addition, it is observed that doctors use various techniques such as X-ray based observation - Mammography, MRI, ultrasound, biopsy to diagnose this deadly disease. From the technical point of view, different technologies emerged such as Machine Learning and Deep Learning, subfields of AI for easy and early detection of breast cancer. Many kinds of research were conducted on machine learning algorithms such as KVM, SVM, KNN classifier etc to predict breast cancer. In our research, we used an approach called CNN (convolutional neural networks) in deep learning. It extracts the best features of the images with accuracy far greater than the machine learning models. Our research is based on How accurate a CNN model can be in detecting breast cancer, the model is RESNET9 in PyTorch and also predicting what type of breast cancer it is, like Benign or malignant type of breast cancer. We also considered various performance evaluation metrics such as F1 scores, Precision, Recall for accurate classification of data.