Papers › Generalized Focal Loss V2: Learning Reliable Localization Quality Estimation for Dense...
Generalized Focal Loss V2: Learning Reliable Localization Quality Estimation for Dense Object Detection
25 Nov 2020CVPR 2021 1arXiv:2011.12885archive 2025-07-28
Xiang Li, Wenhai Wang, Xiaolin Hu, Jun Li, Jinhui Tang, Jian Yang
Localization Quality Estimation (LQE) is crucial and popular in the recent advancement of dense object detectors since it can provide accurate ranking scores that benefit the Non-Maximum Suppression processing and improve detection performance. As a common practice, most existing methods predict LQE scores through vanilla convolutional features shared with object classification or bounding box regression. In this paper, we explore a completely novel and different perspective to perform LQE -- based on the learned distributions of the four parameters of the bounding box. The bounding box distributions are inspired and introduced as "General Distribution" in GFLV1, which describes the uncertainty of the predicted bounding boxes well. Such a property makes the distribution statistics of a bounding box highly correlated to its real localization quality. Specifically, a bounding box distribution with a sharp peak usually corresponds to high localization quality, and vice versa. By leveraging the close correlation between distribution statistics and the real localization quality, we develop a considerably lightweight Distribution-Guided Quality Predictor (DGQP) for reliable LQE based on GFLV1, thus producing GFLV2. To our best knowledge, it is the first attempt in object detection to use a highly relevant, statistical representation to facilitate LQE. Extensive experiments demonstrate the effectiveness of our method. Notably, GFLV2 (ResNet-101) achieves 46.2 AP at 14.6 FPS, surpassing the previous state-of-the-art ATSS baseline (43.6 AP at 14.6 FPS) by absolute 2.6 AP on COCO test-dev, without sacrificing the efficiency both in training and inference. Code will be available at https://github.com/implus/GFocalV2.
In Syntology View this paper on Syntology, its page in Syntology's graph. That page lists the repositories linked to the paper, the abstract and the calls for agents.
Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.
For agents, on Syntology's MCP service (how to connect):
get_citation_path(paper_1="2011.12885", paper_2="…") with another paper's arXiv id or titleget_concepts_for_paper(arxiv_id="2011.12885")
Code
PaddlePaddle/PaddleDetection
paddleSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to PP-YOLOE: An evolved version of YOLO (arXiv:2203.16250; that paper's own run record: 22 ran (of which 0 constructed an object rather than computing a result; 19 with no instrument failure: 2 honoured, 0 violated, 17 with no contract checked; 3 where Syntology's instrument failed) · 5 unverified); nothing checks that they implement this paper's method. report
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.
A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.
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 |
GFLV2 (Res2Net-101, DCN, multiscale) |
AP50 |
70.9 |
#66 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (Res2Net-101, DCN, multiscale) |
AP75 |
59.2 |
#66 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (Res2Net-101, DCN, multiscale) |
APL |
65.6 |
#66 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (Res2Net-101, DCN, multiscale) |
APM |
56.1 |
#66 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (Res2Net-101, DCN, multiscale) |
APS |
35.7 |
#66 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (Res2Net-101, DCN, multiscale) |
box mAP |
53.3 |
#66 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (Res2Net-101, DCN) |
AP50 |
69 |
#89 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (Res2Net-101, DCN) |
AP75 |
55.3 |
#89 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (Res2Net-101, DCN) |
APL |
63.5 |
#89 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (Res2Net-101, DCN) |
APM |
54.3 |
#89 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (Res2Net-101, DCN) |
APS |
31.3 |
#89 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (Res2Net-101, DCN) |
box mAP |
50.6 |
#89 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNeXt-101, 32x4d, DCN) |
AP50 |
67.6 |
#99 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNeXt-101, 32x4d, DCN) |
AP75 |
53.5 |
#99 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNeXt-101, 32x4d, DCN) |
APL |
61.4 |
#99 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNeXt-101, 32x4d, DCN) |
APM |
52.4 |
#99 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNeXt-101, 32x4d, DCN) |
APS |
29.7 |
#99 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNeXt-101, 32x4d, DCN) |
Hardware Burden |
3G |
#99 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNeXt-101, 32x4d, DCN) |
box mAP |
49 |
#99 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNet-101-DCN) |
AP50 |
66.5 |
#107 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNet-101-DCN) |
AP75 |
52.8 |
#107 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNet-101-DCN) |
APL |
60.7 |
#107 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNet-101-DCN) |
APM |
51.9 |
#107 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNet-101-DCN) |
APS |
28.8 |
#107 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNet-101-DCN) |
Hardware Burden |
3G |
#107 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNet-101-DCN) |
box mAP |
48.3 |
#107 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNet-101) |
AP50 |
64.3 |
#128 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNet-101) |
AP75 |
50.5 |
#128 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNet-101) |
APL |
57 |
#128 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNet-101) |
APM |
49.9 |
#128 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNet-101) |
APS |
27.8 |
#128 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNet-101) |
box mAP |
46.2 |
#128 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNet-50) |
AP50 |
62.3 |
#148 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNet-50) |
AP75 |
48.5 |
#148 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNet-50) |
APL |
54.1 |
#148 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNet-50) |
APM |
47.7 |
#148 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNet-50) |
APS |
26.8 |
#148 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO test-dev |
GFLV2 (ResNet-50) |
box mAP |
44.3 |
#148 of 225 |
Archive leaderboard |
report |
| Object Detection |
COCO-O |
GFLv2
(R2-101-DCN) |
Average mAP |
25.1 |
#26 of 45 |
Archive leaderboard |
report |
| Object Detection |
COCO-O |
GFLv2
(R2-101-DCN) |
Effective Robustness |
2.6 |
#26 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