An Automated and Smart Breast Cancer Detection and Classification Framework using DenseNet based Deep Learning Approach

The breast cancer detection is very critical in the field of medicine because it is among the most common types of cancer in the world and a leading cause of death among women. Diagnosis then becomes difficult due to the complicated nature of the mammogram images, noise presence, variation of characteristics in tumors, and difficulty in the distinction between benign and malignant growths. The aforementioned reasons normally bring about false positives and false negatives, hence another reason to research more effective ways of detecting and diagnosing tumors. To overcome the challenges, this paper proposed a deep-learning-based approach to detect and classify breast cancer in a new style. This article exploits the possibility offered by DenseNet-Dense Convolutional Network, powerful architecture for a dense network owing to the densely connected way, which efficiently propagates the feature and gradients' flow for deep networks, offering an overall best-performing model and thereby better prediction in mammogram classification. Hence, the model is designed for the classification between benign and malignant tumors, hence proving a strong solution to the challenges posed by the traditional methods.