Deep Learning CNNs for Breast Cancer Classification and Detection ”Enhancing Diagnostic Accuracy in Medical Practice

This article explores the use of deep learning (DL), specifically Convolutional Neural Networks (CNNs), for the early detection of breast cancer (BC). Traditional methods of breast cancer detection, such as mammography, clinical breast examination and breast self-examination, have limitations and can result in false positives and false negatives. Moreover, these traditional methods are limited in their availability in low-resource settings. Deep learning has shown promising results in medical image analysis and this study investigates the capabilities of CNNs for breast cancer detection. The article provides an overview of machine learning (ML), including supervised, unsupervised and reinforcement learning and describes the advantages, challenges and limitations of deep learning in medical image analysis. A bag-of-tricks approach is taken to analyze the effects on performance and Competency using diverse deep learning techniques such as different architectures (VGG19, ResNet50, ConceptionV3, DenseNet121, MobileNetV2), class weights, Input sizes, learning transfer quantities, types of mammograms. Good resolution ratios have been achieved, depending on the techniques mentioned earlier, but the idea that we touched on in this article to improve the accuracy and efficiency in cancer detection, is that we focused on the nature of image insertion. Typically, RGB or gray-scale format is used for image processing, but in this study, we will rely on the YCbCr color space for breast cancer images to determine its effectiveness and improve the image quality and to improve classification accuracy and this with several other techniques and improvements and this is what we will discuss in the results section of the last article.