A hybrid GAN-based deep learning framework for thermogram-based breast cancer detection.

Breast cancer remains one of the most prevalent and life-threatening diseases among women worldwide, necessitating early and accurate detection methods. Traditional diagnostic approaches often face limitations in sensitivity and specificity, highlighting the need for advanced computational techniques. This study proposes BCDGAN, a novel deep learning framework integrating a Generative Adversarial Network (GAN) and a Hybrid Deep Learning (HDL) model for breast cancer detection using thermogram imagery. The aim is to enhance diagnostic accuracy by synthesizing critical regions of interest (ROIs) and leveraging deep feature extraction for improved classification performance. The proposed GAN-HDL-BCD method first employs a hybrid deep learning model to extract features from thermogram images, followed by a GAN-based approach to generate synthetic ROIs, augmenting the dataset and improving model generalization. Experimental validation using the DMR-IR benchmark dataset demonstrates that the proposed framework achieves an accuracy of 98.56%, outperforming conventional deep-learning models. These findings suggest that BCDGAN can be effectively integrated into clinical decision-support applications for thermogram-based breast cancer screening, offering a promising tool for enhancing early detection and patient outcomes.