用3000张农业图像自监督预训练，比换架构还能提升4.57%病害分类准确率。
Domain-Specific Self-Supervised Pre-training for Agricultural Disease Classification: A Hierarchical Vision Transformer Study
- 用3000张未标注农业图像做SimCLR自监督预训练
- 在3个数据集上最高达96.3%准确率，较基线提升4.57%
- 适合想提升农业图像分类的开发者和农学研究者
我们研究了领域特定自监督预训练对农业病害分类的影响，采用分层视觉变换器。关键发现：仅用3000张未标注农业图像进行SimCLR预训练，可带来+4.57%的准确率提升，超过分层架构设计带来的+3.70%收益。更重要的是，该自监督收益与架构无关：应用于Swin-Base和ViT-Base分别获得+4.08%和+4.20%提升，表明应优先收集领域数据而非纠结模型结构。基于Swin风格的分层视觉变换器（HVT），在三个数据集上表现优异：棉叶病害（7类，90.24%）、PlantVillage（38类，96.3%）、PlantDoc（27类，87.1%）。在参数量相近条件下，HVT-Base（78M）达到88.91%，优于Swin-Base（88M）的87.23%，提升1.68%。为评估部署可靠性，报告校准分析：HVT的ECE为3.56%（温度缩放后降至1.52%）。代码开源：https://github.com/w2sg-arnav/HierarchicalViT
原文摘要 · Abstract (English)
We investigate the impact of domain-specific self-supervised pre-training on agricultural disease classification using hierarchical vision transformers. Our key finding is that SimCLR pre-training on just 3,000 unlabeled agricultural images provides a +4.57% accuracy improvement--exceeding the +3.70% gain from hierarchical architecture design. Critically, we show this SSL benefit is architecture-agnostic: applying the same pre-training to Swin-Base yields +4.08%, to ViT-Base +4.20%, confirming practitioners should prioritize domain data collection over architectural choices. Using HierarchicalViT (HVT), a Swin-style hierarchical transformer, we evaluate on three datasets: Cotton Leaf Disease (7 classes, 90.24%), PlantVillage (38 classes, 96.3%), and PlantDoc (27 classes, 87.1%). At matched parameter counts, HVT-Base (78M) achieves 88.91% vs. Swin-Base (88M) at 87.23%, a +1.68% improvement. For deployment reliability, we report calibration analysis showing HVT achieves 3.56% ECE (1.52% after temperature scaling). Code: https://github.com/w2sg-arnav/HierarchicalViT
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。