Metraipolis › archive › dc_a0ebf81219bf
Robust Lightweight Deep Learning Models for Oral Cancer Screening — arXiv:2608.21583
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.21583
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-18T07:38:45.229Z
- SHA-256 of the archived text
477ab744acc0a966f7b07ccaa646c2dc5d095f85a29a894da131c789a14d93a8- Stable URL of this page
https://metraipolis.com/documento/dc_a0ebf81219bf
How to cite
Robust Lightweight Deep Learning Models for Oral Cancer Screening — arXiv:2608.21583 — arXiv (en), https://arxiv.org/abs/2608.21583 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a0ebf81219bf (Metraipolis archive dc_a0ebf81219bf, sha256 477ab744acc0a966f7b07ccaa646c2dc5d095f85a29a894da131c789a14d93a8)
BibTeX
@misc{dca0ebf81219bf,
title = {Robust Lightweight Deep Learning Models for Oral Cancer Screening — arXiv:2608.21583},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.21583},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a0ebf81219bf; sha256 477ab744acc0a966f7b07ccaa646c2dc5d095f85a29a894da131c789a14d93a8},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.21583. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Oral cancer is a leading cause of mortality in low-to-middle-income countries, where a shortage of specialists delays diagnosis. While point-of-care screening via smartphones offers a scalable solution, developing robust AI for resource-constrained settings poses significant challenges, including class imbalance in training data, variable data quality, and computational constraints on edge devices. In this paper, we present the optimisation of lightweight deep learning models for smartphone-based oral cancer screening. Using a diverse, multi-centre retrospective dataset of approximately 30,000 images acquired over a decade, we systematically evaluate state-of-the-art convolutional, transformer, and hybrid architectures. Through rigorous pipeline ablation, we demonstrate that directly optimising hybrid architectures for the edge strictly outperforms computationally heavy paradigms, such as large models or knowledge distillation. Furthermore, interpretability analysis and simulated noise-stress tests revealed that the system anchors on clinical features and remains robust to unstructured sensor noise, despite vulnerabilities to impulse bit errors. In the held-out test set, our optimised MobileViTv2 models achieved an average sensitivity of 83.2 $\pm$ 1.5% and an average specificity of 86.0 $\pm$ 0.8%, with the best model exhibiting 87.4% sensitivity, 86.5% specificity, and a critical negative predictive value of 97.2% with reference to specialist labels. These results confirm that with targeted architectural selection and streamlined optimisation, interpretable and robust lightweight AI models exhibit high potential for edge deployment to enable automated triage in primary care settings. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-25 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2608.21583 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-18T07:37:52.650Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-25?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).