Leveraging the Machine Learning and Deep Learning Techniques to Detect Breast Cancer Earlier

Worldwide the Breast cancer is a major reason of cancer-related mortalities among women. This research aims to compare Support Vector Machines (SVM), Convolutional Neural Networks (CNN), and Recurrent Neural Networks (RNN) as potential models for breast cancer detection using full mammogram images from the CBIS DDSM dataset. The primary objective is to identify early-stage breast cancer and classify images as Calcification or Mass by extracting relevant features from mammograms. Given the image processing nature of this study, CNN offers a distinct advantage, as it can autonomously learn intricate features from mammogram images through its multiple convolutional layers. In contrast, SVM employs the kernel function to classify images by extracting features, while RNN excels in capturing temporal dependencies in sequential data, making it valuable for cancer detection. The evaluation of these models relies on the confusion matrix, providing insights into correct and incorrect classifications of Cancerous and Non-Cancerous images. The essential metrics, like accuracy, precision, recall, and F1-score, are utilized to get the performances of each model. The study's findings reveal that CNN achieves the high accuracy rate. This comparative analysis sheds light on the strengths and limitations of both Machine Learning (SVM) and Deep Learning models (CNN and RNN) in the context of early breast cancer detection. The valuable information derived from this research holds significant potential for advancements in the medical field and could lead to early cancer detection, ultimately saving lives.