用深度学习自动分析花粉，效率提升6倍，准确率超95%。
Automated Palynological Analysis System: Integrating Deep Metric Learning and $U^{2}$-Net Detection in $H\infty$ bright field microscopy
- 结合U²-Net检测与DINOv2模型，实现花粉精准识别
- 分类召回率达95.8%，处理速度比人工快6倍
- 可解释性标注助力专家验证，适合生态研究者
传统蜜源花粉分析耗时4-6小时/样本，主观性强。本文提出一种自动化高通量显微系统，融合$H\infty$鲁棒机械控制与深度学习流水线，对智利南中央地区比奥比奥区域的花粉进行精确计数、分类与形态分析。系统采用U²-Net进行显著目标检测，基于DINOv2视觉变换器骨干网络，通过深度度量学习训练完成分类，并结合梯度加权注意力生成可解释的纹理与诊断特征标注。相比人工专家分析，系统实现95.8%的分类召回率和6倍的处理速度提升。
原文摘要 · Abstract (English)
Traditional melissopalynology is a time-consuming and subjective process, often taking 4-6 hours per sample. We present an automated, high-throughput microscopy system that integrates $H\infty$ robust mechanical control with advanced deep learning pipelines for the precise counting, classification, and morphological analysis of pollen grains from Bio Bio region in south central territory in Chile. Our system employs $U^{2}$-Net for salient object detection and a DINOv2 Vision Transformer backbone trained via Deep Metric Learning for classification. By integrating Gradient-Weighted Attention, the model provides human-interpretable texture and diagnostic feature annotations. The system achieves a 95.8$\%$ classification recall and a 6x processing speedup compared to manual expert analysis.
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。