EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Multimodal Deep Learning with Attention-Based Fusion for Skin Cancer Diagnosis

Wiem Abdelbaki iD, Hend Alshaya iD, Inzamam Mashood Nasir iD, Sara Tehsin iD, Wided Bouchelligua iD

Bioengineering · 16.05.2026 · 13 (5) · 564

DOI: 10.3390/bioengineering13050564 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Bioengineering, online erschienen am 16.05.2026. Verfasst von Wiem Abdelbaki und 4 weiteren Autor:innen. Beteiligte Einrichtung: College of Engineering and Technology, American University of the Middle East, Egaila 54200, Kuwait. Die Arbeit nennt 40 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Imam Mohammad ibn Saud Islamic University.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The diagnosis of skin cancer remains a growing challenge because of its high variability as a result of the varying imaging conditions in clinical settings. This paper proposes a multimodal deep learning framework to address these challenges by combining the auxiliary clinical information with dermoscopic image features. This proposed architecture uses an attention-based feature encoder with a structured multimodal fusion approach to utilize the integrated feature representation across all channels. Evaluations of the proposed architecture were conducted across a range of benchmark datasets, including ISIC 2019, ISIC 2020, and HAM10000, using a unified experimental approach. This proposed model achieved accuracies of 90.5%, 88.7%, and 91.8% and AUCs of 95.8%, 94.6%, and 96.3%, respectively, on the selected datasets. For the baseline models, ResNet50 and EfficientNet-B4, our approach increased the AUC by 6.5% and the F1 score by 8.0%. Furthermore, across various datasets, the model achieved an AUC of 90.9%, proposing strong generalization. From the ablation analysis results, the attention and multimodal fusion mechanisms showed a 4.1% decrease in AUC when key components were removed, confirming their effectiveness. With 34.7 million parameters and an average of 19.3 Ms., the model has adequate intensity to deploy in a real clinical setting without affecting its performance. Additionally, the improvements to the model were statistically significant across all evaluation metrics (p = 0.01). The proposed multimodal framework demonstrates strong performance and robustness across multiple benchmark datasets.

Schlagwörter

Stichwörter der Autor:innen

attention mechanismsclinical data fusioncross-dataset generalizationdermoscopic image analysismultimodal deep learningskin cancer diagnosis

Autor:innen und Institutionen

  • College of Engineering and Technology, American University of the Middle East, Egaila 54200, Kuwait.
  • Applied College, Imam Mohammad Ibn Saud Islamic University (IMSIU), Riyadh 11432, Saudi Arabia.
  • Human-Environment-Technology (HET) Systems Centre, Mykolas Romeris University, 08303 Vilnius, Lithuania.
  • Faculty of Informatics, Kaunas University of Technology, 51368 Kaunas, Lithuania.

ORCID-Profile: Wiem Abdelbaki, Hend Alshaya, Inzamam Mashood Nasir, Sara Tehsin, Wided Bouchelligua

Förderung

  • Imam Mohammad ibn Saud Islamic University

Zitieren

APA

Abdelbaki, W., Alshaya, H., Nasir, I. M., Tehsin, S., & Bouchelligua, W. (2026). Multimodal Deep Learning with Attention-Based Fusion for Skin Cancer Diagnosis. Bioengineering, 13(5), 564. https://doi.org/10.3390/bioengineering13050564

Vancouver

Abdelbaki W, Alshaya H, Nasir IM, Tehsin S, Bouchelligua W. Multimodal Deep Learning with Attention-Based Fusion for Skin Cancer Diagnosis. Bioengineering. 2026;13(5):564. doi:10.3390/bioengineering13050564

BibTeX
@article{Abdelbaki2026,
  author = {Abdelbaki, Wiem and Alshaya, Hend and Nasir, Inzamam Mashood and Tehsin, Sara and Bouchelligua, Wided},
  title = {Multimodal Deep Learning with Attention-Based Fusion for Skin Cancer Diagnosis},
  journal = {Bioengineering},
  year = {2026},
  volume = {13},
  number = {5},
  pages = {564},
  doi = {10.3390/bioengineering13050564}
}
Kostenlos starten →