用分块推理和视觉聚类先验，零样本实现多物种植物识别
Tile-Based ViT Inference with Visual-Cluster Priors for Zero-Shot Multi-Species Plant Identification
- 分块推理：4×4切片匹配模型感受野，逐块预测
- 零样本提升：用聚类与地理位置过滤构建先验，F1达0.348
- 无需训练：仅靠聚类先验重加权，适合快速部署
我们介绍DS@GT在PlantCLEF 2025多物种植物识别挑战赛中的第二名解决方案。该流程结合（i）微调的Vision Transformer ViTD2PC24All进行像素级推理，（ii）4×4分块策略，使图像块大小与网络518×518的感受野对齐，（iii）通过PaCMAP + K-Means视觉聚类与地理信息筛选进行领域先验适配。分块预测通过多数投票聚合，并使用聚类特定的贝叶斯先验重新加权，在私有排行榜上取得0.348的宏平均F1，且无需额外训练。所有代码、配置文件及可复现脚本已公开于https://github.com/dsgt-arc/plantclef-2025。
原文摘要 · Abstract (English)
We describe DS@GT's second-place solution to the PlantCLEF 2025 challenge on multi-species plant identification in vegetation quadrat images. Our pipeline combines (i) a fine-tuned Vision Transformer ViTD2PC24All for patch-level inference, (ii) a 4x4 tiling strategy that aligns patch size with the network's 518x518 receptive field, and (iii) domain-prior adaptation through PaCMAP + K-Means visual clustering and geolocation filtering. Tile predictions are aggregated by majority vote and re-weighted with cluster-specific Bayesian priors, yielding a macro-averaged F1 of 0.348 (private leaderboard) while requiring no additional training. All code, configuration files, and reproducibility scripts are publicly available at https://github.com/dsgt-arc/plantclef-2025.
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。