| 1 | Focal-Stable-DINO (Focal-Huge, no TTA) | 71.4 | Yes | A Strong and Reproducible Object Detector with O... | 2023-04-25 | Code |
| 2 | EVA | 70.8 | Yes | EVA: Exploring the Limits of Masked Visual Repre... | 2022-11-14 | Code |
| 3 | UNINEXT-H | 66.7 | Yes | Universal Instance Perception as Object Discover... | 2023-03-12 | Code |
| 4 | QueryInst (single scale) | 61.7 | No | Instances as Queries | 2021-05-05 | Code |
| 5 | SOLQ (Swin-L, single scale) | 61.3 | No | SOLQ: Segmenting Objects by Learning Queries | 2021-06-04 | Code |
| 6 | YOLOv4-P7 CSP-P7 (single-scale, 16 fps) | 60.7 | No | Scaled-YOLOv4: Scaling Cross Stage Partial Network | 2020-11-16 | Code |
| 7 | YOLOR-D6 (1280, single-scale, 31 fps) | 60.6 | No | You Only Learn One Representation: Unified Netwo... | 2021-05-10 | Code |
| 8 | EfficientDet-D7x (single-scale) | 59 | No | EfficientDet: Scalable and Efficient Object Dete... | 2019-11-20 | Code |
| 9 | UniverseNet-20.08d (Res2Net-101, DCN, multi-scale) | 58.9 | No | USB: Universal-Scale Object Detection Benchmark | 2021-03-25 | Code |
| 10 | YOLOR-P6 (1280, single-scale, 72 fps) | 57.4 | No | You Only Learn One Representation: Unified Netwo... | 2021-05-10 | Code |
| 11 | ResNeSt-200 (multi-scale) | 57.07 | No | ResNeSt: Split-Attention Networks | 2020-04-19 | Code |
| 12 | GCNet (ResNeXt-101 + DCN + cascade + GC r4) | 56.1 | No | Global Context Networks | 2020-12-24 | Code |
| 13 | ELSA-S (Cascade Mask RCNN) | 56 | No | ELSA: Enhanced Local Self-Attention for Vision T... | 2021-12-23 | Code |
| 14 | FocalNet-T (LRF, Cascade Mask R-CNN) | 56 | No | Focal Modulation Networks | 2022-03-22 | Code |
| 15 | DINO-5scale (24 epoch) | 56 | No | DINO: DETR with Improved DeNoising Anchor Boxes ... | 2022-03-07 | Code |
| 16 | DINO-5scale (36 epoch) | 55.8 | No | DINO: DETR with Improved DeNoising Anchor Boxes ... | 2022-03-07 | Code |
| 17 | FocalNet-T (SRF, Cascade Mask R-CNN) | 55.8 | No | Focal Modulation Networks | 2022-03-22 | Code |
| 18 | ResNeSt-200-DCN (single-scale) | 55.4 | No | ResNeSt: Split-Attention Networks | 2020-04-19 | Code |
| 19 | UniverseNet-20.08d (Res2Net-101, DCN, single-scale) | 55.4 | No | USB: Universal-Scale Object Detection Benchmark | 2021-03-25 | Code |
| 20 | ResNeSt-200 (single-scale) | 55.17 | No | ResNeSt: Split-Attention Networks | 2020-04-19 | Code |
| 21 | Sparse R-CNN (PVTv2-B2) | 54.9 | No | PVT v2: Improved Baselines with Pyramid Vision T... | 2021-06-25 | Code |
| 22 | BoTNet 200 (Mask R-CNN, single scale, 72 epochs) | 54.6 | No | Bottleneck Transformers for Visual Recognition | 2021-01-27 | Code |
| 23 | BoTNet 152 (Mask R-CNN, single scale, 72 epochs) | 54.2 | No | Bottleneck Transformers for Visual Recognition | 2021-01-27 | Code |
| 24 | DN-Deformable-DETR-R50++ | 53.8 | No | DN-DETR: Accelerate DETR Training by Introducing... | 2022-03-02 | Code |
| 25 | REGO-Deformable DETR-X101 | 53.1 | No | Recurrent Glimpse-based Decoder for Detection wi... | 2021-12-09 | Code |
| 26 | Mask R-CNN (ResNeXt-152-FPN, cascade) | 52.9 | No | Rethinking ImageNet Pre-training | 2018-11-21 | Code |
| 27 | ELSA-S (Mask RCNN) | 52.9 | No | ELSA: Enhanced Local Self-Attention for Vision T... | 2021-12-23 | Code |
| 28 | UniverseNet-20.08 (Res2Net-50, DCN, single-scale) | 52.6 | No | USB: Universal-Scale Object Detection Benchmark | 2021-03-25 | Code |
| 29 | GCNet (ResNeXt-101 + DCN + cascade + GC r16) | 52.2 | No | GCNet: Non-local Networks Meet Squeeze-Excitatio... | 2019-04-25 | Code |
| 30 | MAE-Det(MAE-Det-L+GFLV2) | 52.2 | No | MAE-DET: Revisiting Maximum Entropy Principle in... | 2021-11-26 | Code |
| 31 | Res2Net101+HTC | 51.3 | No | Res2Net: A New Multi-scale Backbone Architecture | 2019-04-02 | Code |
| 32 | Mask R-CNN (ResNeXt-152-FPN) | 51.1 | No | Rethinking ImageNet Pre-training | 2018-11-21 | Code |
| 33 | Cascade Mask R-CNN (ResNet-50) | 50.5 | No | Deep Residual Learning for Image Recognition | 2015-12-10 | Code |
| 34 | HoughNet (HG-104, MS) | 50.3 | No | HoughNet: Integrating near and long-range eviden... | 2020-07-05 | Code |
| 35 | DAB-DETR-DC5-R101 | 50.2 | No | DAB-DETR: Dynamic Anchor Boxes are Better Querie... | 2022-01-28 | Code |
| 36 | Conditional DETR-DC5-R101 | 49.5 | No | Conditional DETR for Fast Training Convergence | 2021-08-13 | Code |
| 37 | Sparse R-CNN (ResNet-101, learnable proposals, random crop aug, FPN) | 49.5 | No | Sparse R-CNN: End-to-End Object Detection with L... | 2020-11-25 | Code |
| 38 | Mask R-CNN-FPN (AOGNet-40M) | 49.1 | No | Attentive Normalization | 2019-08-04 | Code |
| 39 | Mask R-CNN (ResNeXt-152 + 1 NL) | 48.9 | No | Non-local Neural Networks | 2017-11-21 | Code |
| 40 | R3-CNN (ResNet-50-FPN, DCN) | 48.9 | No | Recursively Refined R-CNN: Instance Segmentation... | 2021-04-03 | Code |
| 41 | Anchor DETR-DC5-R101 | 48.8 | No | Anchor DETR: Query Design for Transformer-Based ... | 2021-09-15 | Code |
| 42 | Cascade R-CNN (HRNetV2p-W48) | 48.7 | No | Deep High-Resolution Representation Learning for... | 2019-08-20 | Code |
| 43 | Pix2seq (R101-DC5) | 48.6 | No | Pix2seq: A Language Modeling Framework for Objec... | 2021-09-22 | Code |
| 44 | Conditional DETR-DC5-R50 | 48.5 | No | Conditional DETR for Fast Training Convergence | 2021-08-13 | Code |
| 45 | R3-CNN (ResNet-50-FPN, GC-Net) | 48.4 | No | Recursively Refined R-CNN: Instance Segmentation... | 2021-04-03 | Code |
| 46 | GFL (ResNet-50) | 48.3 | No | Deep Residual Learning for Image Recognition | 2015-12-10 | Code |
| 47 | Sparse R-CNN (ResNet-50, learnable proposals, random crop aug, FPN) | 48.2 | No | Sparse R-CNN: End-to-End Object Detection with L... | 2020-11-25 | Code |
| 48 | Faster RCNN-R101-FPN+ | 47.8 | No | End-to-End Object Detection with Transformers | 2020-05-26 | Code |
| 49 | DETR-DC5 (ResNet-101) | 47.7 | No | End-to-End Object Detection with Transformers | 2020-05-26 | Code |
| 50 | Cascade R-CNN (HRNetV2p-W32) | 47.7 | No | Deep High-Resolution Representation Learning for... | 2019-08-20 | Code |
| 51 | RetinaNet (ViL-Base, multi-scale, 3x) | 47.6 | No | Multi-Scale Vision Longformer: A New Vision Tran... | 2021-03-29 | Code |
| 52 | Conditional DETR-R101 | 47.5 | No | Conditional DETR for Fast Training Convergence | 2021-08-13 | Code |
| 53 | Anchor DETR-DC5-R50 | 47.5 | No | Anchor DETR: Query Design for Transformer-Based ... | 2021-09-15 | Code |
| 54 | DAB-DETR-R101 | 47.2 | No | DAB-DETR: Dynamic Anchor Boxes are Better Querie... | 2022-01-28 | Code |
| 55 | Sparse R-CNN (ResNet-101, FPN) | 47.2 | No | Sparse R-CNN: End-to-End Object Detection with L... | 2020-11-25 | Code |
| 56 | Mask R-CNN-FPN (ResNeXt-101, GN+WS) | 47.11 | No | Micro-Batch Training with Batch-Channel Normaliz... | 2019-03-25 | Code |
| 57 | RetinaNet (ViL-Base) | 47.1 | No | Multi-Scale Vision Longformer: A New Vision Tran... | 2021-03-29 | Code |
| 58 | ATSS (ResNet-50) | 47 | No | Deep Residual Learning for Image Recognition | 2015-12-10 | Code |
| 59 | HoughNet (HG-104) | 46.9 | No | HoughNet: Integrating near and long-range eviden... | 2020-07-05 | Code |
| 60 | ExtremeNet (Hourglass-104, multi-scale) | 46.8 | No | Bottom-up Object Detection by Grouping Extreme a... | 2019-01-23 | Code |
| 61 | Cascade R-CNN (ResNet-101-FPN+, cascade) | 46.6 | No | Cascade R-CNN: Delving into High Quality Object ... | 2017-12-03 | Code |
| 62 | Faster R-CNN (FPN, X-volution) | 46.4 | No | X-volution: On the unification of convolution an... | 2021-06-04 | - |
| 63 | R3-CNN (ResNet-50-FPN) | 46.3 | No | Recursively Refined R-CNN: Instance Segmentation... | 2021-04-03 | Code |
| 64 | Mask R-CNN (ResNet-101-FPN, GroupNorm, long) | 46.2 | No | Group Normalization | 2018-03-22 | Code |
| 65 | PVT-Large (RetinaNet 3x,MS) | 46.1 | No | Pyramid Vision Transformer: A Versatile Backbone... | 2021-02-24 | Code |
| 66 | Pix2seq (R50-DC5 ) | 46.1 | No | Pix2seq: A Language Modeling Framework for Objec... | 2021-09-22 | Code |
| 67 | Faster R-CNN (HRNetV2p-W48) | 45.9 | No | Deep High-Resolution Representation Learning for... | 2019-08-20 | Code |
| 68 | Conditional DETR-R50 | 45.7 | No | Conditional DETR for Fast Training Convergence | 2021-08-13 | Code |
| 69 | Sparse R-CNN (ResNet-50, FPN) | 45.7 | No | Sparse R-CNN: End-to-End Object Detection with L... | 2020-11-25 | Code |
| 70 | Faster R-CNN (LIP-ResNet-101) | 45.6 | No | LIP: Local Importance-based Pooling | 2019-08-12 | Code |
| 71 | R3-CNN (ResNet-50-FPN, GRoIE) | 45.6 | No | Recursively Refined R-CNN: Instance Segmentation... | 2021-04-03 | Code |
| 72 | TridentNet (ResNet-101) | 45.5 | No | Scale-Aware Trident Networks for Object Detection | 2019-01-07 | Code |
| 73 | PVT-Large (RetinaNet 1x) | 45.4 | No | Pyramid Vision Transformer: A Versatile Backbone... | 2021-02-24 | Code |
| 74 | Cascade R-CNN (HRNetV2p-W18) | 44.9 | No | Deep High-Resolution Representation Learning for... | 2019-08-20 | Code |
| 75 | PoolFormer-S36 (Mask R-CNN) | 44.8 | No | MetaFormer Is Actually What You Need for Vision | 2021-11-22 | Code |
| 76 | Faster R-CNN (HRNetV2p-W32) | 44.8 | No | Deep High-Resolution Representation Learning for... | 2019-08-20 | Code |
| 77 | Mask R-CNN (ResNet-101 + 1 NL) | 44.5 | No | Non-local Neural Networks | 2017-11-21 | Code |
| 78 | Grid R-CNN (ResNet-101-FPN) | 44.4 | No | Grid R-CNN | 2018-11-29 | Code |
| 79 | Mask R-CNN (ResNet-50-FPN, GroupNorm, long) | 44.4 | No | Group Normalization | 2018-03-22 | Code |
| 80 | PPDet (ResNet-101-FPN) | 44.2 | No | Reducing Label Noise in Anchor-Free Object Detec... | 2020-08-03 | Code |
| 81 | RetinaMask (ResNet-101-FPN) | 44.1 | No | RetinaMask: Learning to predict masks improves s... | 2019-01-10 | Code |
| 82 | GCnet (ResNet-50-FPN, GRoIE) | 44 | No | GCNet: Non-local Networks Meet Squeeze-Excitatio... | 2019-04-25 | Code |
| 83 | Mask R-CNN (ResNet-50-FPN, GroupNorm) | 44 | No | Group Normalization | 2018-03-22 | Code |
| 84 | CenterNet511 (Hourglass-52) | 43.9 | No | CenterNet: Keypoint Triplets for Object Detection | 2019-04-17 | Code |
| 85 | Cascade R-CNN (ResNet-50-FPN+) | 43.7 | No | Cascade R-CNN: Delving into High Quality Object ... | 2017-12-03 | Code |
| 86 | ExtremeNet (Hourglass-104, single-scale) | 43.7 | No | Bottom-up Object Detection by Grouping Extreme a... | 2019-01-23 | Code |
| 87 | Faster R-CNN+aLRP Loss (ResNet-50, 500 scale) | 43.3 | No | A Ranking-based, Balanced Loss Function Unifying... | 2020-09-28 | Code |
| 88 | FPN+ | 43.3 | No | Feature Pyramid Networks for Object Detection | 2016-12-09 | Code |
| 89 | Grid R-CNN (ResNet-50-FPN) | 42.4 | No | Grid R-CNN | 2018-11-29 | Code |
| 90 | RetinaNet+aLRP Loss (ResNet-50, 500 scale) | 42.3 | No | A Ranking-based, Balanced Loss Function Unifying... | 2020-09-28 | Code |
| 91 | Libra R-CNN (ResNet-50 FPN) | 42 | No | Libra R-CNN: Towards Balanced Learning for Objec... | 2019-04-04 | Code |
| 92 | Mask R-CNN (ResNet-50 + 1 NL) | 41.9 | No | Non-local Neural Networks | 2017-11-21 | Code |
| 93 | Mask R-CNN (ResNet-50-FPN, GRoIE) | 41.7 | No | A novel Region of Interest Extraction Layer for ... | 2020-04-28 | Code |
| 94 | FoveaBox+aLRP Loss (ResNet-50, 500 scale) | 41.5 | No | A Ranking-based, Balanced Loss Function Unifying... | 2020-09-28 | Code |
| 95 | FoveaBox (ResNet-101-FPN, 800x800) | 41.5 | No | FoveaBox: Beyond Anchor-based Object Detector | 2019-04-08 | Code |
| 96 | Faster R-CNN (HRNetV2p-W18) | 41.5 | No | Deep High-Resolution Representation Learning for... | 2019-08-20 | Code |
| 97 | FCOS (ResNet-50-FPN + improvements) | 41.4 | No | FCOS: Fully Convolutional One-Stage Object Detec... | 2019-04-02 | Code |
| 98 | CornerNet511 (Hourglass-104) | 40.9 | No | CornerNet: Detecting Objects as Paired Keypoints | 2018-08-03 | Code |
| 99 | HTC (cascade) | 40.7 | No | Hybrid Task Cascade for Instance Segmentation | 2019-01-22 | Code |
| 100 | Faster R-CNN (ResNet-50-FPN, GRoIE) | 40.6 | No | A novel Region of Interest Extraction Layer for ... | 2020-04-28 | Code |
| 101 | FoveaBox+Retina (ResNet-50) | 40.5 | No | FoveaBox: Beyond Anchor-based Object Detector | 2019-04-08 | Code |
| 102 | FoveaBox (ResNet-101-FPN, 600x600) | 40.2 | No | FoveaBox: Beyond Anchor-based Object Detector | 2019-04-08 | Code |
| 103 | Mask R-CNN (ResNeXt-101-FPN) | 38.9 | No | Mask R-CNN | 2017-03-20 | Code |
| 104 | GHM-C + GHM-R (RetinaNet-FPN-ResNet-50, M=30) | 38.1 | No | Gradient Harmonized Single-stage Detector | 2018-11-13 | Code |
| 105 | FoveaBox (ResNet-50-FPN, 600x600) | 37.9 | No | FoveaBox: Beyond Anchor-based Object Detector | 2019-04-08 | Code |
| 106 | FSAF (ResNet-50) | 37.9 | No | Feature Selective Anchor-Free Module for Single-... | 2019-03-02 | Code |