Recognition in Terra Incognita
Sara Beery Grant Van Horn and Pietro Perona
Abstract
It is desirable for detection and classification algorithms to generalize to unfamiliar environments, but suitable benchmarks for quantitatively studying this phenomenon are not yet available. We present a dataset designed to measure recognition generalization to novel environments. The images in our dataset are harvested from twenty camera traps deployed to monitor animal populations. Camera traps are fixed at one location, hence the background changes little across images; capture is triggered automatically, hence there is no human bias. The challenge is learning recognition in a handful of locations, and generalizing animal detection and classification to new locations where no training data is available. In our experiments state-of-the-art algorithms show excellent performance when tested at the same location where they were trained. However, we find that generalization to new locations is poor, especially for classification systems.111The dataset is available at https://beerys.github.io/CaltechCameraTraps/
中文速览
野生动物监测中常用的相机陷阱（camera trap）会在固定地点自动拍摄动物照片，研究者希望在少数已知地点训练好的识别算法能直接用于从未见过的新地点，但学界此前缺乏专门衡量这种"跨地点泛化"能力的数据集和基准。为此，作者构建了 Caltech Camera Traps（CCT）数据集，从美国西南部20个相机陷阱地点收集了近6万张带有边界框标注的图像，并设计了"同地点测试"与"新地点测试"两种评估协议，系统对比算法在已见地点与未见地点上的表现。实验结果显示，当前最先进的检测与分类模型在训练地点表现优异，但一旦迁移到新地点，性能大幅下滑，分类任务尤为严峻。这一发现揭示了现有深度学习模型本质上仍依赖背景纹理等地点特有线索而非真正理解动物外观，为推动更具泛化能力的视觉识别研究提供了清晰的量化基准。
原文 arXiv:1807.04975;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1807.04975v2