Journal
IEEE ROBOTICS AND AUTOMATION LETTERS
Volume 7, Issue 1, Pages 215-222Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/LRA.2021.3123374
Keywords
Object detection; segmentation and categorization; deep learning for visual perception
Categories
Funding
- Australian Research Council (ARC) Centre of Excellence for Robotic Vision [CE140100016]
- QUT Centre for Robotics
- CSIRO Machine Learning and Artificial Intelligence Future Science Platform
Ask authors/readers for more resources
In this paper, 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 modeled with class-specific Gaussian Mixture Models. Experimental results show that GMM-Det consistently outperforms existing uncertainty techniques for identifying and rejecting open-set detections.
Deployed 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.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available