A Dual Transfer Learning Based Model for Mammogram Images Enhancement and Classification
Vandana Saini, Meenu Khurana, Rama Krishna Challa · BENTHAM SCIENCE PUBLISHERS eBooks · 2025
Accurate and timely detection of breast cancer is very important to save a patient’s life. Therefore, designing an accurate computer-aided diagnosis (CAD) for mammogram cancer detection is quite important for providing an interpretation to radiologists. In this paper, a CAD-based model has been proposed based on double transfer learning. The CAD system is trained to detect various abnormalities or cancers from the input mammogram images. In this study, the MIAS mammogram dataset is used to evaluate the proposed work. The original images in the dataset are also enhanced in this paper using a pre-trained VGG-16 network. The pre-processing of images has shown a better peak signal-to-noise (PSNR) value. The proposed network has shown a promising PSNR of more than 70 and classification accuracy of more than 99% with lesser system training complexity.