On the SIIM-ISIC melanoma dataset, the model achieves 94.5% accuracy and an F1 score of 0.94, outperforming traditional image-only models like ResNet-50 and EfficientNet.

The study is set to be published in Information Fusion (Volume 124), with online availability mid-year and formal print publication on December 1, 2025, reflecting international collaboration across Korea, the UK, and Canada.

Researchers frame the work as a foundation for real-world tools, emphasizing transparency and practical deployment in healthcare settings.

The project is led by Prof. Gwangill Jeon of the Department of Embedded Systems Engineering at Incheon National University.

The author list spans affiliations in the UK, Korea, and Canada, including Misbah Ahmad, Imran Ahmed, Abdellah Chehrid, and Gwangill Jeon.

Feature importance analysis identifies lesion size, patient age, and anatomical site as key contributors, enhancing transparency and potential clinical trust in AI decisions.

The report is accessible without a paywall, and the study appears under DOI 10.1016/j.inffus.2025.103304.

Officials suggest practical applications in smartphone-based diagnostics, telemedicine, and AI-assisted dermatology workflows to improve early detection and reduce misdiagnosis.

The work is framed as a step toward personalized diagnosis and preventive medicine by converging imaging data with basic patient information to enable earlier melanoma detection.

Affiliations include Incheon National University (Korea), University of the West of England, Anglia Ruskin University, and the Royal Military College of Canada.

A multimodal deep learning system from Incheon National University and international collaborators fuses dermoscopic images with basic patient metadata to improve melanoma detection beyond image-only approaches.

Led by Prof. Gwangill Jeon and colleagues Misbah Ahmad, Imran Ahmed, and Abdellah Chehrid, the project demonstrates near-perfect accuracy metrics in its reported evaluation and discusses implications for clinical decision-making.