AI RESEARCH
Skin Lesion Classification Based on ResNet-50 Enhanced With Adaptive Spatial Feature Fusion
arXiv CS.CV
•
ArXi:2510.03876v2 Announce Type: replace Skin cancer classification is challenging due to high inter-class similarity, intra-class variability, and artifacts in dermoscopic images. To address these issues, we propose an improved ResNet-50 with Adaptive Spatial Feature Fusion (ASFF), which adaptively integrates multi-scale semantic and surface features to refine representations and reduce overfitting. The ResNet-50 model is enhanced with an adaptive feature fusion mechanism to achieve effective multi-scale feature extraction and improve overall performance.