AI Based Framework for Breast Cancer Detection for Histopathological Images

Abstract

Breast cancer continues to be one of the primary causes of death among women globally, and timely, precise diagnosis is essential for enhancing survival rates. This research introduces a deep learning approach that utilizes the Inception-ResNet-V2 model to classify histopathological breast tissue images. The publicly available BreakHis dataset, consisting of benign and malignant samples at multiple magnification levels, was used to evaluate the proposed method. Images were preprocessed through resizing and normalization before being fed into the model. For binary classification, the proposed model achieved an accuracy of 83.15%, with precision, recall, and F1-score of 0.9487, 0.7912, and 0.8628, respectively. In the multi-class scenario involving eight histological subtypes, the model achieved a testing accuracy of 94%, with macro and weighted F1-scores of 0.94, indicating consistent performance across all classes. The results demonstrate that the proposed approach effectively captures both low- and high-level features from histopathological images, offering a reliable tool for supporting breast cancer diagnosis. The study highlights the potential of deep learning models, particularly Inception-ResNet-V2, in enhancing diagnostic precision and reducing the burden on medical professionals.

Article Content

Keywords

Article Info