arXiv 2512.090692026-02-25cs.CVcs.AIcs.LG

KD-OCT：用于临床级视网膜OCT分类的高效知识蒸馏

KD-OCT: Efficient Knowledge Distillation for Clinical-Grade Retinal OCT Classification

Artificial Intelligence Department University of Isfahan（人工智能系伊斯法罕大学）；Labbafinejad Hospital Shahid Beheshti University of Medical Science（拉巴法尼贾医院谢赫·贝赫什提医学科学大学）；Department of Mechanical Engineering Isfahan University of Technology（机械工程系伊斯法罕技术大学）

Erfan Nourbakhsh, Nasrin Sanjari, Ali Nourbakhsh

备注：: 7 pages, 5 figures (Accepted at ICSPIS 2025)

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

KD-OCT通过知识蒸馏将高性能ConvNeXtV2-Large模型压缩为轻量EfficientNet-B2模型，实现高效且准确的视网膜OCT分类，适用于临床实时部署。

摘要：

年龄相关性黄斑变性（AMD）和脉络膜新生血管（CNV）相关疾病是全球导致视力丧失的主要原因，光学相干断层扫描（OCT）是早期检测和管理的基石。然而，部署最先进的深度学习模型如ConvNeXtV2-Large在临床环境中受到其计算需求的阻碍。因此，开发能够在保持高诊断性能的同时实现实时部署的高效模型是有必要的。在本研究中，提出了一种名为KD-OCT的新知识蒸馏框架，将增强的高级增强、随机权重平均和聚焦损失的高性能ConvNeXtV2-Large教师模型压缩成轻量级的EfficientNet-B2学生模型，用于分类正常、黄斑和CNV病例。KD-OCT采用实时蒸馏，结合损失平衡软教师知识转移和硬真实监督。所提出方法的有效性在Noor Eye Hospital（NEH）数据集上通过患者层面交叉验证评估。实验结果表明，KD-OCT在效率-准确性平衡方面优于同等多尺度或特征融合OCT分类器，实现了接近教师模型的性能，同时在模型大小和推理时间上实现了显著减少。尽管进行了压缩，学生模型仍超过了大多数现有框架，便于AMD筛查的边缘部署。代码可在https://github.com/erfan-nourbakhsh/KD-OCT上获得。

英文摘要：

Age-related macular degeneration (AMD) and choroidal neovascularization (CNV)-related conditions are leading causes of vision loss worldwide, with optical coherence tomography (OCT) serving as a cornerstone for early detection and management. However, deploying state-of-the-art deep learning models like ConvNeXtV2-Large in clinical settings is hindered by their computational demands. Therefore, it is desirable to develop efficient models that maintain high diagnostic performance while enabling real-time deployment. In this study, a novel knowledge distillation framework, termed KD-OCT, is proposed to compress a high-performance ConvNeXtV2-Large teacher model, enhanced with advanced augmentations, stochastic weight averaging, and focal loss, into a lightweight EfficientNet-B2 student for classifying normal, drusen, and CNV cases. KD-OCT employs real-time distillation with a combined loss balancing soft teacher knowledge transfer and hard ground-truth supervision. The effectiveness of the proposed method is evaluated on the Noor Eye Hospital (NEH) dataset using patient-level cross-validation. Experimental results demonstrate that KD-OCT outperforms comparable multi-scale or feature-fusion OCT classifiers in efficiency-accuracy balance, achieving near-teacher performance with substantial reductions in model size and inference time. Despite the compression, the student model exceeds most existing frameworks, facilitating edge deployment for AMD screening. Code is available at https://github.com/erfan-nourbakhsh/KD-OCT.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。