EUVIMEDDescoperă EUVIMED
Open access · CC BY 4.0

Recall-Oriented Adaptive Loss with Knowledge Distillation for Efficient Melanoma Detection Under Class Imbalance

Samah El Kah, My Abdelouahed Sabri, Abdellah Aarab

Machine Learning and Knowledge Extraction · 2026-11-17 · 8 (11) · 384

DOI: 10.3390/make8110384 ↗

At a glance

Journal article in Machine Learning and Knowledge Extraction, published online on 2026-11-17. Written by Samah El Kah and 2 co-authors. 45 references. The full text is freely available (open access, CC BY 4.0).

RezumatDate bibliograficeCiteRelated studies

Abstract (original language)

Melanoma is a highly aggressive skin cancer for which early and accurate detection remains essential. However, automated analysis of dermoscopic images is challenged by severe class imbalance, which can bias deep learning models toward the non-melanoma class and reduce sensitivity. This study presents a recall-oriented knowledge distillation framework that combines the Recall-Oriented Adaptive Loss (ROAL) with teacher–student learning. ROAL increases the positive-class weight whenever melanoma recall on a development subset falls below a predefined target, while a ResNet50 teacher transfers softened predictions to a MobileNetV2 student with approximately 2.2 million parameters, compared with 23.5 million for the teacher. Using a lesion-level split, three random seeds, and a fixed protocol, eleven teacher, student-only, and distillation configurations were evaluated on HAM10000 and further assessed on the more severely imbalanced ISIC 2020. At the common predefined operating threshold of 0.5, distilling a ROAL-trained teacher into a downsampled student achieved the highest melanoma recall on both datasets: 0.924 on HAM10000, with an average of seven false negatives out of 92 melanoma cases, and 0.914 on ISIC 2020, with 9.7 out of 112. In both cases, the distilled student exceeded the recall of its own teacher while using ten times fewer parameters, a gain that reflects the combination of distillation with balanced-batch training rather than distillation alone. On the second dataset, ROAL showed a larger performance gap over fixed class weighting, indicating that the proposed adaptive strategy remained effective under a substantially different and more imbalanced data distribution. Overall, the framework achieves high melanoma sensitivity at a fraction of the model size, supporting its potential deployment in resource-constrained screening settings.

Cite

APA

Kah, S. E., Sabri, M. A., & Aarab, A. (2026). Recall-Oriented Adaptive Loss with Knowledge Distillation for Efficient Melanoma Detection Under Class Imbalance. Machine Learning and Knowledge Extraction, 8(11), 384. https://doi.org/10.3390/make8110384

Vancouver

Kah SE, Sabri MA, Aarab A. Recall-Oriented Adaptive Loss with Knowledge Distillation for Efficient Melanoma Detection Under Class Imbalance. Machine Learning and Knowledge Extraction. 2026;8(11):384. doi:10.3390/make8110384

BibTeX
@article{Kah2026,
  author = {Kah, Samah El and Sabri, My Abdelouahed and Aarab, Abdellah},
  title = {Recall-Oriented Adaptive Loss with Knowledge Distillation for Efficient Melanoma Detection Under Class Imbalance},
  journal = {Machine Learning and Knowledge Extraction},
  year = {2026},
  volume = {8},
  number = {11},
  pages = {384},
  doi = {10.3390/make8110384}
}
Start for free →