Miller, D.Sunderhauf, N.Milford, M.Dayoub, F.2022-03-032022-03-032022IEEE Robotics and Automation Letters, 2022; 7(1):215-2222377-37662377-3766https://hdl.handle.net/2440/134512Deployed into an open world, object detectors are prone to open-set errors, false positive detections of object classes not present in the training dataset.We propose GMM-Det, a real-time method for extracting epistemic uncertainty from object detectors to identify and reject open-set errors. GMM-Det trains the detector to produce a structured logit space that is modelled with class-specific Gaussian Mixture Models. At test time, open-set errors are identified by their low log-probability under all Gaussian Mixture Models. We test two common detector architectures, Faster R-CNN and RetinaNet, across three varied datasets spanning robotics and computer vision. Our results show that GMM-Det consistently outperforms existing uncertainty techniques for identifying and rejecting open-set detections, especially at the low-error-rate operating point required for safety-critical applications. GMM-Det maintains object detection performance, and introduces only minimal computational overhead. We also introduce a methodology for converting existing object detection datasets into specific open-set datasets to evaluate open-set performance in object detection.en© 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.detection; segmentation and categorization; deep learning for visual perceptionJournal article10.1109/lra.2021.31233742022-03-03599050Dayoub, F. [0000-0002-4234-7374]