Long-tail Detection with Effective Class-Margins
Long-tail Detection with Effective Class-Margins
复制标题
DOI：
10.1007/978-3-031-20074-8_40
复制
发表时间：
2023-01
期刊：
影响因子：
--
通讯作者：
Jang Hyun Cho;Philipp Krähenbühl
中科院分区：
文献类型：
--
作者：
Jang Hyun Cho;Philipp Krähenbühl
大规模目标检测和实例分割面临着严重的数据不平衡问题。对象类的粒度越细，它们在我们的数据集中出现的频率就越低。然而，在测试时，我们希望检测器对所有类都表现良好，而不仅仅是最常见的类。在本文中，我们提供了一个理论上的理解的长拖尾检测问题。我们展示了如何在一个未知的测试集上常用的平均精度评估指标是由一个长尾对象检测训练集上的基于边缘的二进制分类错误约束的。我们用一个新的代理目标有效类边际损失（ECM）来优化基于边际的二进制分类错误。ECM损失是简单的，理论上动机良好，并优于其他启发式同行LVIS v1基准在广泛的架构和检测器。代码可在https://github.com/janghyuncho/ECM-Loss上获得。
Large-scale object detection and instance segmentation face a severe data imbalance. The finer-grained object classes become, the less frequent they appear in our datasets. However, at test-time, we expect a detector that performs well for all classes and not just the most frequent ones. In this paper, we provide a theoretical understanding of the long-trail detection problem. We show how the commonly used mean average precision evaluation metric on an unknown test set is bound by a margin-based binary classification error on a long-tailed object detection training set. We optimize margin-based binary classification error with a novel surrogate objective calledEffective Class-Margin Loss(ECM). The ECM loss is simple, theoretically well-motivated, and outperforms other heuristic counterparts on LVIS v1 benchmark over a wide range of architecture and detectors. Code is available at https://github.com/janghyuncho/ECM-Loss.