Papers › NMS Strikes Back
NMS Strikes Back
Jeffrey Ouyang-Zhang, Jang Hyun Cho, Xingyi Zhou, Philipp Krähenbühl
Detection Transformer (DETR) directly transforms queries to unique objects by using one-to-one bipartite matching during training and enables end-to-end object detection. Recently, these models have surpassed traditional detectors on COCO with undeniable elegance. However, they differ from traditional detectors in multiple designs, including model architecture and training schedules, and thus the effectiveness of one-to-one matching is not fully understood. In this work, we conduct a strict comparison between the one-to-one Hungarian matching in DETRs and the one-to-many label assignments in traditional detectors with non-maximum supervision (NMS). Surprisingly, we observe one-to-many assignments with NMS consistently outperform standard one-to-one matching under the same setting, with a significant gain of up to 2.5 mAP. Our detector that trains Deformable-DETR with traditional IoU-based label assignment achieved 50.2 COCO mAP within 12 epochs (1x schedule) with ResNet50 backbone, outperforming all existing traditional or transformer-based detectors in this setting. On multiple datasets, schedules, and architectures, we consistently show bipartite matching is unnecessary for performant detection transformers. Furthermore, we attribute the success of detection transformers to their expressive transformer architecture. Code is available at https://github.com/jozhang97/DETA.
In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.
Code
Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.
Code Syntology ran Syntology
Not run by Syntology. Nothing on this page verifies that the listed code works.
Tasks
Results from the paper archive 2025-07-28
| Task | Dataset | Model | Metric | Value | Rank at snapshot | Leaderboard | Report |
|---|---|---|---|---|---|---|---|
| Object Detection | COCO test-dev | DETA (Swin-L) | AP50 | 80.4 | #16 of 225 | Archive leaderboard | report |
| Object Detection | COCO test-dev | DETA (Swin-L) | AP75 | 70.2 | #16 of 225 | Archive leaderboard | report |
| Object Detection | COCO test-dev | DETA (Swin-L) | APL | 76.9 | #16 of 225 | Archive leaderboard | report |
| Object Detection | COCO test-dev | DETA (Swin-L) | APM | 66.9 | #16 of 225 | Archive leaderboard | report |
| Object Detection | COCO test-dev | DETA (Swin-L) | APS | 46.1 | #16 of 225 | Archive leaderboard | report |
| Object Detection | COCO test-dev | DETA (Swin-L) | box mAP | 63.5 | #16 of 225 | Archive leaderboard | report |
| Object Detection | COCO-O | DETA (Swin-L) | Average mAP | 48.5 | #2 of 45 | Archive leaderboard | report |
| Object Detection | COCO-O | DETA (Swin-L) | Effective Robustness | 20.15 | #2 of 45 | Archive leaderboard | report |
Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.
Methods
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections