Papers › RetinaMask: Learning to predict masks improves state-of-the-art single-shot detection for free
RetinaMask: Learning to predict masks improves state-of-the-art single-shot detection for free
Cheng-Yang Fu, Mykhailo Shvets, Alexander C. Berg
Recently two-stage detectors have surged ahead of single-shot detectors in the accuracy-vs-speed trade-off. Nevertheless single-shot detectors are immensely popular in embedded vision applications. This paper brings single-shot detectors up to the same level as current two-stage techniques. We do this by improving training for the state-of-the-art single-shot detector, RetinaNet, in three ways: integrating instance mask prediction for the first time, making the loss function adaptive and more stable, and including additional hard examples in training. We call the resulting augmented network RetinaMask. The detection component of RetinaMask has the same computational cost as the original RetinaNet, but is more accurate. COCO test-dev results are up to 41.4 mAP for RetinaMask-101 vs 39.1mAP for RetinaNet-101, while the runtime is the same during evaluation. Adding Group Normalization increases the performance of RetinaMask-101 to 41.7 mAP. Code is at:https://github.com/chengyangfu/retinamask
In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.
Code
53 repositories listed; official and paper-mentioned ones first.
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 minival | RetinaMask (ResNet-101-FPN) | AP50 | 60.2 | #168 of 220 | Archive leaderboard | report |
| Object Detection | COCO minival | RetinaMask (ResNet-101-FPN) | AP75 | 44.1 | #168 of 220 | Archive leaderboard | report |
| Object Detection | COCO minival | RetinaMask (ResNet-101-FPN) | box AP | 41.1 | #168 of 220 | Archive leaderboard | report |
| Object Detection | COCO test-dev | RetinaMask (ResNeXt-101-FPN-GN) | AP50 | 62.5 | #172 of 225 | Archive leaderboard | report |
| Object Detection | COCO test-dev | RetinaMask (ResNeXt-101-FPN-GN) | AP75 | 46.0 | #172 of 225 | Archive leaderboard | report |
| Object Detection | COCO test-dev | RetinaMask (ResNeXt-101-FPN-GN) | APL | 53.8 | #172 of 225 | Archive leaderboard | report |
| Object Detection | COCO test-dev | RetinaMask (ResNeXt-101-FPN-GN) | APM | 45.6 | #172 of 225 | Archive leaderboard | report |
| Object Detection | COCO test-dev | RetinaMask (ResNeXt-101-FPN-GN) | APS | 24.8 | #172 of 225 | Archive leaderboard | report |
| Object Detection | COCO test-dev | RetinaMask (ResNeXt-101-FPN-GN) | Hardware Burden | 12G | #172 of 225 | Archive leaderboard | report |
| Object Detection | COCO test-dev | RetinaMask (ResNeXt-101-FPN-GN) | box mAP | 42.6 | #172 of 225 | Archive leaderboard | report |
| Object Detection | COCO test-dev | RetinaMask (ResNet-50-FPN) | AP50 | 58.6 | #204 of 225 | Archive leaderboard | report |
| Object Detection | COCO test-dev | RetinaMask (ResNet-50-FPN) | AP75 | 42.3 | #204 of 225 | Archive leaderboard | report |
| Object Detection | COCO test-dev | RetinaMask (ResNet-50-FPN) | APL | 51.0 | #204 of 225 | Archive leaderboard | report |
| Object Detection | COCO test-dev | RetinaMask (ResNet-50-FPN) | APM | 42.0 | #204 of 225 | Archive leaderboard | report |
| Object Detection | COCO test-dev | RetinaMask (ResNet-50-FPN) | APS | 21.9 | #204 of 225 | Archive leaderboard | report |
| Object Detection | COCO test-dev | RetinaMask (ResNet-50-FPN) | Hardware Burden | 9G | #204 of 225 | Archive leaderboard | report |
| Object Detection | COCO test-dev | RetinaMask (ResNet-50-FPN) | box mAP | 39.4 | #204 of 225 | 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
Introduced by this paper: RetinaMask, Self-Adjusting Smooth L1 Loss
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