Publications

ReSet: A Residual Set-Transformer approach to tackle the ugly-duckling sign in melanoma detection

Published in 2024 IEEE International Conference on Image Processing (ICIP), 2024

We introduce ReSeT (Residual Set-Transformer), a novel approach to address the Ugly-Duckling Sign in melanoma detection. By comparing skin lesions within patients, our model effectively increases accuracy while remaining simple and lightweight, making it generalizable for other applications.

Recommended citation: J. Collenne, R. Iguernaissi, S. Dubuisson, D. Merad (2024). "ReSet: A Residual Set-Transformer approach to tackle the ugly-duckling sign in melanoma detection." International Conference on Image Processing. ICIP 2024. Abu Dhabi, UAE.

Automated melanoma detection. An algorithm inspired from human intelligence characterizing disordered pattern of melanocytic lesions improving a convolutional neural network.

Published in Journal of the American Academy of Dermatology, 2024

An approach focused on the measure of disorder within skin lesions using machine learning models.

Recommended citation: J. Monnier, A. C. Foahom Gouabou, M. Serdi, J. Collenne, R. Iguernaissi, M.-A. Richard, C. Gaudy-Marqueste, J.-L. Damoiseaux, J.-J. Grob, D. Merad. (2024). "Automated melanoma detection. An algorithm inspired from human intelligence characterizing disordered pattern of melanocytic lesions improving a convolutional neural network." Journal of the American Academy of Dermatology. https://doi.org/10.1016/j.jaad.2024.02.063

Fusion between an Algorithm Based on the Characterization of Melanocytic Lesions Asymmetry with an Ensemble of Convolutional Neural Networks for Melanoma Detection

Published in Journal of Investigative Dermatology, 2024

We develop an approach focusing on the asymmetry within skin lesions to detect melanomas. An asymmetry model predicts the asymmetry rate of a given lesion, which, when combined with convolutional neural networks, improves the final results and enhances interpretability for dermatologists.

Recommended citation: J. Collenne, J. Monnier, R. Iguernaissi, M. Nawaf, M.-A. Richard, J.-J. Grob, C. Gaudy-Marqueste, S. Dubuisson, D. Merad. (2024). "Fusion between an Algorithm Based on the Characterization of Melanocytic Lesions Asymmetry with an Ensemble of Convolutional Neural Networks for Melanoma Detection." Journal of Investigative Dermatology. https://doi.org/10.1016/j.jid.2023.09.289

Enhancing Anomaly Detection in Melanoma Diagnosis Through Self-Supervised Training and Lesion Comparison

Published in Machine Learning in Medical Imaging, 2023

A new framework for entirely unsupervised anomaly detection in the field of skin lesion analysis. Our approach leverage self-supervised CNNs and an unsupervised anomaly detection algorithm to detect melanomas without any annotation.

Recommended citation: J. Collenne, R. Iguernaissi, S. Dubuisson, D. Merad (2024). "Enhancing Anomaly Detection in Melanoma Diagnosis Through Self-Supervised Training and Lesion Comparison." Machine Learning in Medical Imaging. MLMI 2023. Vancouver, BC, Canada. https://link.springer.com/chapter/10.1007/978-3-031-45676-3_16

Computer Aided Diagnosis of Melanoma Using Deep Neural Networks and Game Theory: Application on Dermoscopic Images of Skin Lesions

Published in International Journal of Molecular Sciences, 2022

A game theory approach to analyse dermatological images of skin lesions.

Recommended citation: A. C. Foahom Gouabou, J. Collenne, J. Monnier, R. Iguernaissi, J.-L. Damoiseaux, A. Moudafi, D. Merad."Computer Aided Diagnosis of Melanoma Using Deep Neural Networks and Game Theory: Application on Dermoscopic Images of Skin Lesions." International Journal of Molecular Sciences. 2022; 23(22) https://www.mdpi.com/1937126