The recently proposed open-world object and open-set detection achieve a breakthrough in finding never-seen-before objects and distinguishing them from class-known ones. However, their studies on knowledge transfer from known classes to unknown ones need to be deeper, leading to the scanty capability for detecting unknowns hidden in the background. In this paper, we propose the unknown sniffer (UnSniffer) to find both unknown and known objects. Firstly, the generalized object confidence (GOC) score is introduced, which only uses class-known samples for supervision and avoids improper suppression of unknowns in the background. Significantly, such confidence score learned from class-known objects can be generalized to unknown ones. Additionally, we propose a negative energy suppression loss to further limit the non-object samples in the background. Next, the best box of each unknown is hard to obtain during inference due to lacking their semantic information in training. To solve this issue, we introduce a graph-based determination scheme to replace hand-designed non-maximum suppression (NMS) post-processing. Finally, we present the Unknown Object Detection Benchmark, the first publicly benchmark that encompasses precision evaluation for unknown object detection to our knowledge. Experiments show that our method is far better than the existing state-of-the-art methods. Code is available at: https://github.com/Went-Liang/UnSniffer.

翻译：近期提出的开放世界目标检测与开放集检测方法在发现从未见过的目标并将其与已知类别区分方面取得了突破性进展。然而，这些方法在从已知类别到未知类别的知识迁移研究上尚需深入，导致对隐藏在背景中的未知目标检测能力不足。本文提出未知目标嗅探器（UnSniffer），可同时检测未知与已知目标。首先，引入广义目标置信度（GOC）评分，该评分仅使用已知类别样本进行监督，避免对背景中未知目标的不当抑制。值得注意的是，这种从已知目标学习到的置信度评分可泛化至未知目标。此外，我们提出负能量抑制损失函数，以进一步限制背景中的非目标样本。其次，由于训练过程中缺乏未知目标的语义信息，推断时难以获取每个未知目标的最佳边界框。为解决该问题，我们引入基于图的判定方案，替代传统的手工设计的非极大值抑制（NMS）后处理。最后，提出未知目标检测基准（Unknown Object Detection Benchmark），据我们所知，这是首个公开包含未知目标检测精度评估的基准。实验表明，我们的方法显著优于现有最先进方法。代码开源地址：https://github.com/Went-Liang/UnSniffer。