<![CDATA[Skin cancer is one of the most prevalent cancers worldwide, making early and accurate diagnosis essential for improving patient outcomes and reducing mortality. However, automated skin lesion classification remains challenging due to high inter-class similarity, class imbalance, and variations in lesion appearance. This study proposes an attention-enhanced hybrid deep learning for skin cancer diagnosis with hierarchical feature fusion. The proposed framework integrates channel and spatial attention with hierarchical feature fusion to enhance discriminative feature learning and improve classification robustness. Experiments were conducted on the PAD-UFES-20 dataset using image preprocessing and data augmentation. The proposed model achieved approximately 98% training accuracy, 95% validation accuracy, and a validation loss below 0.5, outperforming DRMv2Net, DenseNet201, ResNet101, and MobileNetV2 while demonstrating faster convergence and stronger generalization. These findings demonstrate the potential of hybrid attention and hierarchical feature fusion to improve the reliability and robustness of artificial intelligence-based diagnostic systems in dermatology, supporting more effective clinical decision-making and early skin cancer screening.]]>
Copyrights © 2026
