Deep Learning-Based Early Detection of Breast Cancer: Improving Accuracy and Efficiency in Diagnosis for Enhanced Patient Outcomes

In Todays era breast cancer is a main reason of deaths among women across the world. If detected at early stage it increases the survival rates of the patients. In the last few years, deep learning techniques have been used in medical imaging in the early detection of disease. This research focuses on the development of a deep learning-based mechanism for the early detection of breast cancer by ultrasound scan images. The proposed technique in this paper presents the usage of CNNs to extract relevant features from ultrasound images. A large dataset of annotated ultrasound scans is used for training the deep learning model. To measure the effect of the developed system, extensive experiments and validations are conducted using independent test datasets. In our work, CNN models have been trained to identify the three cases of breast tissues such as normal, benign, and malignant, and were optimized using Recursive Feature Elimination Optimizer and divide it into 80% for the training set and 20% for the testing set. VGG-19, ResNet-101 and ResNeXt architectures optimized by Recursive Feature Elimination optimizer are used for training and validation of the breast cancer dataset and get the accuracy of 98.5% for the optimized VGG-19 model.