Reproduce “A Lightweight Attention-Based CNN with Saliency-Map Interpretability for Skin Cancer Detection” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Sanjana Islam Kasfia, Md. Ismiel Hossen Abir, AKM Bahalul Haque, Abir Mahmud Shahariar
The authors present CNN-Attention, a dermoscopic image classifier that sorts lesions as benign or malignant and produces maps of influential image regions. Its channel attention is inspired by Squeeze-and-Excitation. On a 660-image test set, it matches MobileNet’s rounded accuracy of 0.87 with 1.95 million parameters and 7.45 MB; the comparison does not establish clinical validity or faster inference.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.