Enhanced Automated Breast Cancer Diagnostics System Using Deep Transfer Learning Techniques on Histopathological Images

As breast cancer continues to be the leading cause of death among women around the world, there is an urgent need for diagnostic tools that are efficient, accurate, and automated. The purpose of this research is to develop an Automated Breast Cancer Diagnostics System that makes use of deep transfer learning techniques on histopathology pictures. The system makes use of deep transfer learning techniques like ResNet, EfficientNet, and VGG-19 Net that integrate Global Average Pooling (GAP) layers in order to reduce the amount of time complexity and processing overhead without affecting the accuracy of the results. Traditional fully connected layers are replaced by GAP layers, significantly reducing the number of trainable parameters while preserving powerful feature extraction capabilities. Evaluation of the proposed system is carried out using the Breast Cancer Histopathological Image (BACH) dataset. This dataset comprises high-resolution microscopic images classified into benign, malignant, and normal tissue types. The results of the experiments show that the system obtains an accuracy of 96.7% and an F1-score of 96.3, which is higher than the baseline models. When compared to standard fully connected architectures, the inclusion of GAP layers results in a reduction in the computational cost, which in turn leads to training periods that are 35% faster.