用于口腔癌筛查的鲁棒轻量级深度学习模型
Robust Lightweight Deep Learning Models for Oral Cancer Screening
浏览论文内容
中文总结 AI 辅助
本研究针对中低收入国家口腔癌筛查的需求，优化了轻量级AI模型，发现直接优化混合架构优于大型模型等方案，其MobileViTv2模型在边缘设备上实现高灵敏度与特异度，可用于基层自动分诊。
中文摘要 AI 辅助
口腔癌是中低收入国家的主要致死原因之一，这些地区专科医生短缺导致诊断延迟。虽然通过智能手机进行即时检验（point-of-care）筛查提供了一种可扩展的解决方案，但在资源受限环境下开发鲁棒的人工智能（AI）面临重大挑战，包括训练数据的类别不平衡、数据质量不稳定以及边缘设备的计算约束。在本文中，我们提出针对智能手机口腔癌筛查的轻量级深度学习模型优化方案。我们使用一个涵盖十年采集的约30000张图像的多中心回顾性多样化数据集，系统评估了最先进的卷积、Transformer及混合架构。通过严格的流水线消融实验，我们证明针对边缘设备直接优化混合架构的性能严格优于计算密集型范式，如大型模型或知识蒸馏。此外，可解释性分析和模拟噪声压力测试显示，该系统以临床特征为锚点，对非结构化传感器噪声保持鲁棒性，尽管存在脉冲位错误的脆弱性。在保留的测试集中，我们优化的MobileViTv2模型实现了平均灵敏度83.2±1.5%，平均特异度86.0±0.8%，最优模型表现出87.4%的灵敏度、86.5%的特异度，以及参考专科医生标签的关键阴性预测值97.2%。这些结果证实，通过针对性的架构选择和精简优化，可解释且鲁棒的轻量级AI模型在边缘设备部署中具有巨大潜力，可实现基层医疗环境中的自动分诊。
英文摘要
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.
发表机构
- Indian Institute of Science（印度科学学院）
- KLES’ Institute of Dental Sciences（KLES牙科科学研究所）
- Biocon Foundation（百康基金会）
机构由 AI 辅助整理，请以论文原文为准。
