CNN-Based Deep Learning Model for Detecting Regions of Interest in Breast Cancer Images

Early identification of breast cancer is high in priority to accelerate the survival rate of women with the improved outcome of treatment. This paper explores the development of a deep learning framework to classify from the detected images of breast cancer. Previous algorithms involve multiple sub-functions, leading to increased computational time with higher complexity. This research introduces a modified deep learning structure to reduce the computational complexity of cancer image detection. The proposed framework operates with a higher number of unseen layers and epochs. The algorithm of CNN is employed for feature extraction from cancer images, and the standard IBM cancer image dataset is used to test the system. Since the dataset is fully labelled, training is executed on the cancer images. Classification is performed using a support vector classifier, achieving a result of 99.78% accuracy, which surpasses that of other conventional methods