paper-with-me

Object Detection 벤치마크

Object Detection on COCO-O

225개 결과 · ⬇ CSV · JSON

Average mAP

13.6 24.65 35.7 46.75 57.8 2015-06 2026-09 Faster R-CNN (ResNet-50-FPN) — 16.4 (2015-06-04) Faster R-CNN (ResNet-50-FPN) — 16.4 (2015-06-04) Faster R-CNN (ResNet-50-FPN) — 16.4 (2015-06-04) Faster R-CNN (ResNet-50-FPN) — 16.4 (2015-06-04) Faster R-CNN (ResNet-50-FPN) — 16.4 (2015-06-04) SSD (VGG-16) — 13.6 (2015-12-08) SSD (VGG-16) — 13.6 (2015-12-08) SSD (VGG-16) — 13.6 (2015-12-08) SSD (VGG-16) — 13.6 (2015-12-08) SSD (VGG-16) — 13.6 (2015-12-08) Mask R-CNN (ResNet-50) — 17.1 (2017-03-20) Mask R-CNN (ResNet-50) — 17.1 (2017-03-20) Mask R-CNN (ResNet-50) — 17.1 (2017-03-20) Mask R-CNN (ResNet-50) — 17.1 (2017-03-20) Mask R-CNN (ResNet-50) — 17.1 (2017-03-20) RetinaNet (ResNet-50) — 16.6 (2017-08-07) RetinaNet (ResNet-50) — 16.6 (2017-08-07) RetinaNet (ResNet-50) — 16.6 (2017-08-07) RetinaNet (ResNet-50) — 16.6 (2017-08-07) RetinaNet (ResNet-50) — 16.6 (2017-08-07) YOLOv3 (DarkNet-53) — 14.8 (2018-04-08) YOLOv3 (DarkNet-53) — 14.8 (2018-04-08) YOLOv3 (DarkNet-53) — 14.8 (2018-04-08) YOLOv3 (DarkNet-53) — 14.8 (2018-04-08) YOLOv3 (DarkNet-53) — 14.8 (2018-04-08) HTC (ResNet-50) — 19.1 (2019-01-22) HTC (ResNet-50) — 19.1 (2019-01-22) HTC (ResNet-50) — 19.1 (2019-01-22) HTC (ResNet-50) — 19.1 (2019-01-22) HTC (ResNet-50) — 19.1 (2019-01-22) FCOS (ResNet-50) — 16.7 (2019-04-02) FCOS (ResNet-50) — 16.7 (2019-04-02) FCOS (ResNet-50) — 16.7 (2019-04-02) FCOS (ResNet-50) — 16.7 (2019-04-02) FCOS (ResNet-50) — 16.7 (2019-04-02) GCNet (RX-101-32x4d-DCN) — 26.0 (2019-04-25) GCNet (RX-101-32x4d-DCN) — 26.0 (2019-04-25) GCNet (RX-101-32x4d-DCN) — 26.0 (2019-04-25) GCNet (RX-101-32x4d-DCN) — 26.0 (2019-04-25) GCNet (RX-101-32x4d-DCN) — 26.0 (2019-04-25) Cascade R-CNN (ResNet-50) — 18.2 (2019-06-24) Cascade R-CNN (ResNet-50) — 18.2 (2019-06-24) Cascade R-CNN (ResNet-50) — 18.2 (2019-06-24) Cascade R-CNN (ResNet-50) — 18.2 (2019-06-24) Cascade R-CNN (ResNet-50) — 18.2 (2019-06-24) EfficientDet-D5 (EfficientNet-B5) — 28.5 (2019-11-20) EfficientDet-D5 (EfficientNet-B5) — 28.5 (2019-11-20) EfficientDet-D5 (EfficientNet-B5) — 28.5 (2019-11-20) EfficientDet-D5 (EfficientNet-B5) — 28.5 (2019-11-20) EfficientDet-D5 (EfficientNet-B5) — 28.5 (2019-11-20) ATSS (ResNet-50) — 16.8 (2019-12-05) ATSS (ResNet-50) — 16.8 (2019-12-05) ATSS (ResNet-50) — 16.8 (2019-12-05) ATSS (ResNet-50) — 16.8 (2019-12-05) ATSS (ResNet-50) — 16.8 (2019-12-05) YOLOv4-P6 — 30.4 (2020-04-23) YOLOv4-P6 — 30.4 (2020-04-23) YOLOv4-P6 — 30.4 (2020-04-23) YOLOv4-P6 — 30.4 (2020-04-23) YOLOv4-P6 — 30.4 (2020-04-23) DETR (ResNet-50) — 17.1 (2020-05-26) DETR (ResNet-50) — 17.1 (2020-05-26) DETR (ResNet-50) — 17.1 (2020-05-26) DETR (ResNet-50) — 17.1 (2020-05-26) DETR (ResNet-50) — 17.1 (2020-05-26) RepPointsV2 (RX-101-64x4d-DCN) — 24.9 (2020-07-16) RepPointsV2 (RX-101-64x4d-DCN) — 24.9 (2020-07-16) RepPointsV2 (RX-101-64x4d-DCN) — 24.9 (2020-07-16) RepPointsV2 (RX-101-64x4d-DCN) — 24.9 (2020-07-16) RepPointsV2 (RX-101-64x4d-DCN) — 24.9 (2020-07-16) VFNet (RX-101-64x4d) — 28.0 (2020-08-31) VFNet (RX-101-64x4d) — 28.0 (2020-08-31) VFNet (RX-101-64x4d) — 28.0 (2020-08-31) VFNet (RX-101-64x4d) — 28.0 (2020-08-31) VFNet (RX-101-64x4d) — 28.0 (2020-08-31) Deformable-DETR (ResNet-50) — 18.5 (2020-10-08) Deformable-DETR (ResNet-50) — 18.5 (2020-10-08) Deformable-DETR (ResNet-50) — 18.5 (2020-10-08) Deformable-DETR (ResNet-50) — 18.5 (2020-10-08) Deformable-DETR (ResNet-50) — 18.5 (2020-10-08) GFLv2 (R2-101-DCN) — 25.1 (2020-11-25) GFLv2 (R2-101-DCN) — 25.1 (2020-11-25) GFLv2 (R2-101-DCN) — 25.1 (2020-11-25) GFLv2 (R2-101-DCN) — 25.1 (2020-11-25) GFLv2 (R2-101-DCN) — 25.1 (2020-11-25) CenterNet2 (R2-101-DCN) — 29.5 (2021-03-12) CenterNet2 (R2-101-DCN) — 29.5 (2021-03-12) CenterNet2 (R2-101-DCN) — 29.5 (2021-03-12) CenterNet2 (R2-101-DCN) — 29.5 (2021-03-12) CenterNet2 (R2-101-DCN) — 29.5 (2021-03-12) Det-AdvProp (EfficientNet-B5) — 30.8 (2021-03-23) Det-AdvProp (EfficientNet-B5) — 30.8 (2021-03-23) Det-AdvProp (EfficientNet-B5) — 30.8 (2021-03-23) Det-AdvProp (EfficientNet-B5) — 30.8 (2021-03-23) Det-AdvProp (EfficientNet-B5) — 30.8 (2021-03-23) UniverseNet (R2-101-DCN) — 24.8 (2021-03-25) UniverseNet (R2-101-DCN) — 24.8 (2021-03-25) UniverseNet (R2-101-DCN) — 24.8 (2021-03-25) UniverseNet (R2-101-DCN) — 24.8 (2021-03-25) UniverseNet (R2-101-DCN) — 24.8 (2021-03-25) QueryInst (Swin-L) — 33.2 (2021-05-05) QueryInst (Swin-L) — 33.2 (2021-05-05) QueryInst (Swin-L) — 33.2 (2021-05-05) QueryInst (Swin-L) — 33.2 (2021-05-05) QueryInst (Swin-L) — 33.2 (2021-05-05) YOLOS-B (ViT-B) — 20.0 (2021-06-01) YOLOS-B (ViT-B) — 20.0 (2021-06-01) YOLOS-B (ViT-B) — 20.0 (2021-06-01) YOLOS-B (ViT-B) — 20.0 (2021-06-01) YOLOS-B (ViT-B) — 20.0 (2021-06-01) DyHead (Swin-L) — 35.3 (2021-06-15) DyHead (ResNet-50) — 19.3 (2021-06-15) DyHead (Swin-L) — 35.3 (2021-06-15) DyHead (ResNet-50) — 19.3 (2021-06-15) DyHead (Swin-L) — 35.3 (2021-06-15) DyHead (ResNet-50) — 19.3 (2021-06-15) DyHead (Swin-L) — 35.3 (2021-06-15) DyHead (ResNet-50) — 19.3 (2021-06-15) DyHead (Swin-L) — 35.3 (2021-06-15) DyHead (ResNet-50) — 19.3 (2021-06-15) PVTv2-B5 (Mask R-CNN) — 28.2 (2021-06-25) PVTv2-B5 (Mask R-CNN) — 28.2 (2021-06-25) PVTv2-B5 (Mask R-CNN) — 28.2 (2021-06-25) PVTv2-B5 (Mask R-CNN) — 28.2 (2021-06-25) PVTv2-B5 (Mask R-CNN) — 28.2 (2021-06-25) CBNetV2 (Swin-L) — 39.0 (2021-07-01) CBNetV2 (Swin-L) — 39.0 (2021-07-01) CBNetV2 (Swin-L) — 39.0 (2021-07-01) CBNetV2 (Swin-L) — 39.0 (2021-07-01) CBNetV2 (Swin-L) — 39.0 (2021-07-01) YOLOX-X — 30.3 (2021-07-18) YOLOX-S — 20.6 (2021-07-18) YOLOX-X — 30.3 (2021-07-18) YOLOX-S — 20.6 (2021-07-18) YOLOX-X — 30.3 (2021-07-18) YOLOX-S — 20.6 (2021-07-18) YOLOX-X — 30.3 (2021-07-18) YOLOX-S — 20.6 (2021-07-18) YOLOX-X — 30.3 (2021-07-18) YOLOX-S — 20.6 (2021-07-18) MViTV2-H (Cascade Mask R-CNN) — 30.9 (2021-12-02) MViTV2-H (Cascade Mask R-CNN) — 30.9 (2021-12-02) MViTV2-H (Cascade Mask R-CNN) — 30.9 (2021-12-02) MViTV2-H (Cascade Mask R-CNN) — 30.9 (2021-12-02) MViTV2-H (Cascade Mask R-CNN) — 30.9 (2021-12-02) GLIP-L (Swin-L) — 48.0 (2021-12-07) GLIP-T (Swin-T) — 29.1 (2021-12-07) GLIP-L (Swin-L) — 48.0 (2021-12-07) GLIP-T (Swin-T) — 29.1 (2021-12-07) GLIP-L (Swin-L) — 48.0 (2021-12-07) GLIP-T (Swin-T) — 29.1 (2021-12-07) GLIP-L (Swin-L) — 48.0 (2021-12-07) GLIP-T (Swin-T) — 29.1 (2021-12-07) GLIP-L (Swin-L) — 48.0 (2021-12-07) GLIP-T (Swin-T) — 29.1 (2021-12-07) ConvNeXt-XL (Cascade Mask R-CNN) — 37.5 (2022-01-10) ConvNeXt-XL (Cascade Mask R-CNN) — 37.5 (2022-01-10) ConvNeXt-XL (Cascade Mask R-CNN) — 37.5 (2022-01-10) ConvNeXt-XL (Cascade Mask R-CNN) — 37.5 (2022-01-10) ConvNeXt-XL (Cascade Mask R-CNN) — 37.5 (2022-01-10) DINO (Swin-L) — 42.1 (2022-03-07) DINO (Swin-L) — 42.1 (2022-03-07) DINO (Swin-L) — 42.1 (2022-03-07) DINO (Swin-L) — 42.1 (2022-03-07) DINO (Swin-L) — 42.1 (2022-03-07) ViTDet (ViT-H) — 34.3 (2022-03-30) ViTDet (ViT-H) — 34.3 (2022-03-30) ViTDet (ViT-H) — 34.3 (2022-03-30) ViTDet (ViT-H) — 34.3 (2022-03-30) ViTDet (ViT-H) — 34.3 (2022-03-30) ViT-Adapter (BEiTv2-L) — 34.25 (2022-05-17) ViT-Adapter (BEiTv2-L) — 34.25 (2022-05-17) ViT-Adapter (BEiTv2-L) — 34.25 (2022-05-17) ViT-Adapter (BEiTv2-L) — 34.25 (2022-05-17) ViT-Adapter (BEiTv2-L) — 34.25 (2022-05-17) FIBER-B (Swin-B) — 33.7 (2022-06-15) FIBER-B (Swin-B) — 33.7 (2022-06-15) FIBER-B (Swin-B) — 33.7 (2022-06-15) FIBER-B (Swin-B) — 33.7 (2022-06-15) FIBER-B (Swin-B) — 33.7 (2022-06-15) YOLOv7-E6E — 32.0 (2022-07-06) YOLOv7-E6E — 32.0 (2022-07-06) YOLOv7-E6E — 32.0 (2022-07-06) YOLOv7-E6E — 32.0 (2022-07-06) YOLOv7-E6E — 32.0 (2022-07-06) YOLOv6-L6 — 32.5 (2022-09-07) YOLOv6-L6 — 32.5 (2022-09-07) YOLOv6-L6 — 32.5 (2022-09-07) YOLOv6-L6 — 32.5 (2022-09-07) YOLOv6-L6 — 32.5 (2022-09-07) InternImage-L (Cascade Mask R-CNN) — 37.0 (2022-11-10) InternImage-L (Cascade Mask R-CNN) — 37.0 (2022-11-10) InternImage-L (Cascade Mask R-CNN) — 37.0 (2022-11-10) InternImage-L (Cascade Mask R-CNN) — 37.0 (2022-11-10) InternImage-L (Cascade Mask R-CNN) — 37.0 (2022-11-10) EVA — 57.8 (2022-11-14) EVA — 57.8 (2022-11-14) EVA — 57.8 (2022-11-14) EVA — 57.8 (2022-11-14) EVA — 57.8 (2022-11-14) GRiT (ViT-H) — 42.9 (2022-12-01) GRiT (ViT-H) — 42.9 (2022-12-01) GRiT (ViT-H) — 42.9 (2022-12-01) GRiT (ViT-H) — 42.9 (2022-12-01) GRiT (ViT-H) — 42.9 (2022-12-01) DETA (Swin-L) — 48.5 (2022-12-12) DETA (Swin-L) — 48.5 (2022-12-12) DETA (Swin-L) — 48.5 (2022-12-12) DETA (Swin-L) — 48.5 (2022-12-12) DETA (Swin-L) — 48.5 (2022-12-12) Faster R-CNN (ResNet-50-FPN) — 16.4 (2015-06-04) Mask R-CNN (ResNet-50) — 17.1 (2017-03-20) HTC (ResNet-50) — 19.1 (2019-01-22) GCNet (RX-101-32x4d-DCN) — 26.0 (2019-04-25) EfficientDet-D5 (EfficientNet-B5) — 28.5 (2019-11-20) YOLOv4-P6 — 30.4 (2020-04-23) Det-AdvProp (EfficientNet-B5) — 30.8 (2021-03-23) QueryInst (Swin-L) — 33.2 (2021-05-05) DyHead (Swin-L) — 35.3 (2021-06-15) CBNetV2 (Swin-L) — 39.0 (2021-07-01) GLIP-L (Swin-L) — 48.0 (2021-12-07) EVA — 57.8 (2022-11-14)
RankModel Average mAPEffective Robustness Extra Training Data PaperCodeYear
101 ViT-Adapter (BEiTv2-L) 34.257.79 Vision Transformer Adapter for Dense Predictions czczup/vit-adapter · chenller/mmseg-extension 2022
102 FIBER-B (Swin-B) 33.711.43 Coarse-to-Fine Vision-Language Pre-training with Fusion in the Backbone microsoft/fiber 2022
103 QueryInst (Swin-L) 33.28.26 Instances as Queries open-mmlab/mmdetection · hustvl/QueryInst · Bo396543018/picodet_repro · +2 2021
104 YOLOv6-L6 32.56.73 YOLOv6: A Single-Stage Object Detection Framework for Industrial Applications PaddlePaddle/PaddleDetection · meituan/yolov6 · open-mmlab/mmyolo · +4 2022
105 YOLOv7-E6E 32.06.42 YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors pjreddie/darknet · AlexeyAB/darknet · wongkinyiu/yolov7 · +18 2022
106 MViTV2-H (Cascade Mask R-CNN) 30.95.62 MViTv2: Improved Multiscale Vision Transformers for Classification and Detection rwightman/pytorch-image-models · facebookresearch/detectron2 · facebookresearch/SlowFast · +6 2021
107 Det-AdvProp (EfficientNet-B5) 30.87.34 Robust and Accurate Object Detection via Adversarial Learning google/automl · MindSpore-scientific-2/code-5 · MindSpore-scientific-2/code-4 2021
108 YOLOv4-P6 30.45.89 YOLOv4: Optimal Speed and Accuracy of Object Detection tensorflow/models · pjreddie/darknet · AlexeyAB/darknet · +220 2020
109 YOLOX-X 30.37.26 YOLOX: Exceeding YOLO Series in 2021 open-mmlab/mmdetection · PaddlePaddle/PaddleDetection · Megvii-BaseDetection/YOLOX · +39 2021
110 CenterNet2 (R2-101-DCN) 29.54.29 Probabilistic two-stage detection xingyizhou/CenterNet2 · smart-car-lab/Centernet2-mmdetction · aim-uofa/DiverGen 2021
111 GLIP-T (Swin-T) 29.18.11 Grounded Language-Image Pre-training microsoft/GLIP · brown-palm/ObjectPrompt · rsCPSyEu/ovd_cod 2021
112 EfficientDet-D5 (EfficientNet-B5) 28.55.44 EfficientDet: Scalable and Efficient Object Detection tensorflow/models · PaddlePaddle/PaddleDetection · google/automl · +61 2019
113 PVTv2-B5 (Mask R-CNN) 28.26.85 PVT v2: Improved Baselines with Pyramid Vision Transformer rwightman/pytorch-image-models · open-mmlab/mmdetection · open-mmlab/mmpose · +15 2021
114 VFNet (RX-101-64x4d) 28.05.27 VarifocalNet: An IoU-aware Dense Object Detector open-mmlab/mmdetection · hyz-xmaster/VarifocalNet · fcakyon/sahi-benchmark · +1 2020
115 GCNet (RX-101-32x4d-DCN) 26.04.38 GCNet: Non-local Networks Meet Squeeze-Excitation Networks and Beyond open-mmlab/mmdetection · open-mmlab/mmsegmentation · PaddlePaddle/PaddleSeg · +6 2019
116 GFLv2 (R2-101-DCN) 25.12.6 Generalized Focal Loss V2: Learning Reliable Localization Quality Estimation for Dense Object Detection PaddlePaddle/PaddleDetection · implus/GFocalV2 · shinya7y/UniverseNet · +2 2020
117 RepPointsV2 (RX-101-64x4d-DCN) 24.92.7 RepPoints V2: Verification Meets Regression for Object Detection Scalsol/RepPointsV2 2020
118 UniverseNet (R2-101-DCN) 24.8– USB: Universal-Scale Object Detection Benchmark shinya7y/UniverseNet 2021
119 YOLOX-S 20.62.48 YOLOX: Exceeding YOLO Series in 2021 open-mmlab/mmdetection · PaddlePaddle/PaddleDetection · Megvii-BaseDetection/YOLOX · +39 2021
120 YOLOS-B (ViT-B) 20.01.05 You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection huggingface/transformers · hustvl/YOLOS 2021
121 DyHead (ResNet-50) 19.30.16 Dynamic Head: Unifying Object Detection Heads with Attentions open-mmlab/mmdetection · microsoft/DynamicHead · Coldestadam/DynamicHead 2021
122 HTC (ResNet-50) 19.10.08 Hybrid Task Cascade for Instance Segmentation open-mmlab/mmdetection · PaddlePaddle/PaddleDetection · amirassov/kaggle-imaterialist · +2 2019
123 Deformable-DETR (ResNet-50) 18.5-1.49 Deformable DETR: Deformable Transformers for End-to-End Object Detection PaddlePaddle/PaddleDetection · fundamentalvision/Deformable-DETR · roboflow/rf-detr · +17 2020
124 Cascade R-CNN (ResNet-50) 18.20.02 Cascade R-CNN: High Quality Object Detection and Instance Segmentation open-mmlab/mmdetection · zhaoweicai/cascade-rcnn · zhaoweicai/Detectron-Cascade-RCNN · +1 2019
125 Mask R-CNN (ResNet-50) 17.1– Mask R-CNN tensorflow/models · facebookresearch/detectron2 · facebookresearch/detectron · +176 2017
125 DETR (ResNet-50) 17.1-1.82 End-to-End Object Detection with Transformers huggingface/transformers · tensorflow/models · open-mmlab/mmdetection · +34 2020
127 ATSS (ResNet-50) 16.8-0.91 Bridging the Gap Between Anchor-based and Anchor-free Detection via Adaptive Training Sample Selection open-mmlab/mmdetection · RangiLyu/nanodet · open-edge-platform/training_extensions · +10 2019
128 FCOS (ResNet-50) 16.70.25 FCOS: Fully Convolutional One-Stage Object Detection open-mmlab/mmdetection · pytorch/vision · PaddlePaddle/PaddleDetection · +84 2019
129 RetinaNet (ResNet-50) 16.60.18 Focal Loss for Dense Object Detection tensorflow/models · facebookresearch/detectron2 · open-mmlab/mmdetection · +231 2017
130 Faster R-CNN (ResNet-50-FPN) 16.4-0.41 ✓ Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks facebookresearch/detectron2 · open-mmlab/mmdetection · facebookresearch/detectron · +193 2015
131 YOLOv3 (DarkNet-53) 14.8-0.37 YOLOv3: An Incremental Improvement open-mmlab/mmdetection · ultralytics/yolov3 · qqwweee/keras-yolo3 · +308 2018
132 SSD (VGG-16) 13.60.36 SSD: Single Shot MultiBox Detector open-mmlab/mmdetection · serengil/deepface · pytorch/vision · +218 2015
133 ViTDet (ViT-H) –7.89 Exploring Plain Vision Transformer Backbones for Object Detection facebookresearch/detectron2 · PaddlePaddle/PaddleDetection · alibaba/EasyCV · +8 2022
134 UniverseNet (R2-101-DCN) –1.86 USB: Universal-Scale Object Detection Benchmark shinya7y/UniverseNet 2021
135 Mask R-CNN (ResNet-50) –-0.11 Mask R-CNN tensorflow/models · facebookresearch/detectron2 · facebookresearch/detectron · +176 2017
136 EVA 57.828.86 EVA: Exploring the Limits of Masked Visual Representation Learning at Scale rwightman/pytorch-image-models · open-mmlab/mmselfsup · baaivision/eva · +3 2022
137 DETA (Swin-L) 48.520.15 NMS Strikes Back jozhang97/deta 2022
138 GLIP-L (Swin-L) 48.024.89 Grounded Language-Image Pre-training microsoft/GLIP · brown-palm/ObjectPrompt · rsCPSyEu/ovd_cod 2021
139 GRiT (ViT-H) 42.915.72 GRiT: A Generative Region-to-text Transformer for Object Understanding JialianW/GRiT 2022
140 DINO (Swin-L) 42.115.76 DINO: DETR with Improved DeNoising Anchor Boxes for End-to-End Object Detection IDEA-Research/Grounded-Segment-Anything · PaddlePaddle/PaddleDetection · lucasjinreal/yolov7_d2 · +13 2022
141 CBNetV2 (Swin-L) 39.012.36 CBNet: A Composite Backbone Network Architecture for Object Detection PaddlePaddle/PaddleDetection · shinya7y/UniverseNet · VDIGPKU/CBNetV2 · +1 2021
142 ConvNeXt-XL (Cascade Mask R-CNN) 37.512.68 A ConvNet for the 2020s keras-team/keras · rwightman/pytorch-image-models · pytorch/vision · +51 2022
143 InternImage-L (Cascade Mask R-CNN) 37.011.72 InternImage: Exploring Large-Scale Vision Foundation Models with Deformable Convolutions opengvlab/internimage · OpenGVLab/M3I-Pretraining · chenller/mmseg-extension 2022
144 DyHead (Swin-L) 35.310.00 Dynamic Head: Unifying Object Detection Heads with Attentions open-mmlab/mmdetection · microsoft/DynamicHead · Coldestadam/DynamicHead 2021
145 ViTDet (ViT-H) 34.3– Exploring Plain Vision Transformer Backbones for Object Detection facebookresearch/detectron2 · PaddlePaddle/PaddleDetection · alibaba/EasyCV · +8 2022
146 ViT-Adapter (BEiTv2-L) 34.257.79 Vision Transformer Adapter for Dense Predictions czczup/vit-adapter · chenller/mmseg-extension 2022
147 FIBER-B (Swin-B) 33.711.43 Coarse-to-Fine Vision-Language Pre-training with Fusion in the Backbone microsoft/fiber 2022
148 QueryInst (Swin-L) 33.28.26 Instances as Queries open-mmlab/mmdetection · hustvl/QueryInst · Bo396543018/picodet_repro · +2 2021
149 YOLOv6-L6 32.56.73 YOLOv6: A Single-Stage Object Detection Framework for Industrial Applications PaddlePaddle/PaddleDetection · meituan/yolov6 · open-mmlab/mmyolo · +4 2022
150 YOLOv7-E6E 32.06.42 YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors pjreddie/darknet · AlexeyAB/darknet · wongkinyiu/yolov7 · +18 2022
151 MViTV2-H (Cascade Mask R-CNN) 30.95.62 MViTv2: Improved Multiscale Vision Transformers for Classification and Detection rwightman/pytorch-image-models · facebookresearch/detectron2 · facebookresearch/SlowFast · +6 2021
152 Det-AdvProp (EfficientNet-B5) 30.87.34 Robust and Accurate Object Detection via Adversarial Learning google/automl · MindSpore-scientific-2/code-5 · MindSpore-scientific-2/code-4 2021
153 YOLOv4-P6 30.45.89 YOLOv4: Optimal Speed and Accuracy of Object Detection tensorflow/models · pjreddie/darknet · AlexeyAB/darknet · +220 2020
154 YOLOX-X 30.37.26 YOLOX: Exceeding YOLO Series in 2021 open-mmlab/mmdetection · PaddlePaddle/PaddleDetection · Megvii-BaseDetection/YOLOX · +39 2021
155 CenterNet2 (R2-101-DCN) 29.54.29 Probabilistic two-stage detection xingyizhou/CenterNet2 · smart-car-lab/Centernet2-mmdetction · aim-uofa/DiverGen 2021
156 GLIP-T (Swin-T) 29.18.11 Grounded Language-Image Pre-training microsoft/GLIP · brown-palm/ObjectPrompt · rsCPSyEu/ovd_cod 2021
157 EfficientDet-D5 (EfficientNet-B5) 28.55.44 EfficientDet: Scalable and Efficient Object Detection tensorflow/models · PaddlePaddle/PaddleDetection · google/automl · +61 2019
158 PVTv2-B5 (Mask R-CNN) 28.26.85 PVT v2: Improved Baselines with Pyramid Vision Transformer rwightman/pytorch-image-models · open-mmlab/mmdetection · open-mmlab/mmpose · +15 2021
159 VFNet (RX-101-64x4d) 28.05.27 VarifocalNet: An IoU-aware Dense Object Detector open-mmlab/mmdetection · hyz-xmaster/VarifocalNet · fcakyon/sahi-benchmark · +1 2020
160 GCNet (RX-101-32x4d-DCN) 26.04.38 GCNet: Non-local Networks Meet Squeeze-Excitation Networks and Beyond open-mmlab/mmdetection · open-mmlab/mmsegmentation · PaddlePaddle/PaddleSeg · +6 2019
161 GFLv2 (R2-101-DCN) 25.12.6 Generalized Focal Loss V2: Learning Reliable Localization Quality Estimation for Dense Object Detection PaddlePaddle/PaddleDetection · implus/GFocalV2 · shinya7y/UniverseNet · +2 2020
162 RepPointsV2 (RX-101-64x4d-DCN) 24.92.7 RepPoints V2: Verification Meets Regression for Object Detection Scalsol/RepPointsV2 2020
163 UniverseNet (R2-101-DCN) 24.8– USB: Universal-Scale Object Detection Benchmark shinya7y/UniverseNet 2021
164 YOLOX-S 20.62.48 YOLOX: Exceeding YOLO Series in 2021 open-mmlab/mmdetection · PaddlePaddle/PaddleDetection · Megvii-BaseDetection/YOLOX · +39 2021
165 YOLOS-B (ViT-B) 20.01.05 You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection huggingface/transformers · hustvl/YOLOS 2021
166 DyHead (ResNet-50) 19.30.16 Dynamic Head: Unifying Object Detection Heads with Attentions open-mmlab/mmdetection · microsoft/DynamicHead · Coldestadam/DynamicHead 2021
167 HTC (ResNet-50) 19.10.08 Hybrid Task Cascade for Instance Segmentation open-mmlab/mmdetection · PaddlePaddle/PaddleDetection · amirassov/kaggle-imaterialist · +2 2019
168 Deformable-DETR (ResNet-50) 18.5-1.49 Deformable DETR: Deformable Transformers for End-to-End Object Detection PaddlePaddle/PaddleDetection · fundamentalvision/Deformable-DETR · roboflow/rf-detr · +17 2020
169 Cascade R-CNN (ResNet-50) 18.20.02 Cascade R-CNN: High Quality Object Detection and Instance Segmentation open-mmlab/mmdetection · zhaoweicai/cascade-rcnn · zhaoweicai/Detectron-Cascade-RCNN · +1 2019
170 Mask R-CNN (ResNet-50) 17.1– Mask R-CNN tensorflow/models · facebookresearch/detectron2 · facebookresearch/detectron · +176 2017
170 DETR (ResNet-50) 17.1-1.82 End-to-End Object Detection with Transformers huggingface/transformers · tensorflow/models · open-mmlab/mmdetection · +34 2020
172 ATSS (ResNet-50) 16.8-0.91 Bridging the Gap Between Anchor-based and Anchor-free Detection via Adaptive Training Sample Selection open-mmlab/mmdetection · RangiLyu/nanodet · open-edge-platform/training_extensions · +10 2019
173 FCOS (ResNet-50) 16.70.25 FCOS: Fully Convolutional One-Stage Object Detection open-mmlab/mmdetection · pytorch/vision · PaddlePaddle/PaddleDetection · +84 2019
174 RetinaNet (ResNet-50) 16.60.18 Focal Loss for Dense Object Detection tensorflow/models · facebookresearch/detectron2 · open-mmlab/mmdetection · +231 2017
175 Faster R-CNN (ResNet-50-FPN) 16.4-0.41 ✓ Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks facebookresearch/detectron2 · open-mmlab/mmdetection · facebookresearch/detectron · +193 2015
176 YOLOv3 (DarkNet-53) 14.8-0.37 YOLOv3: An Incremental Improvement open-mmlab/mmdetection · ultralytics/yolov3 · qqwweee/keras-yolo3 · +308 2018
177 SSD (VGG-16) 13.60.36 SSD: Single Shot MultiBox Detector open-mmlab/mmdetection · serengil/deepface · pytorch/vision · +218 2015
178 ViTDet (ViT-H) –7.89 Exploring Plain Vision Transformer Backbones for Object Detection facebookresearch/detectron2 · PaddlePaddle/PaddleDetection · alibaba/EasyCV · +8 2022
179 UniverseNet (R2-101-DCN) –1.86 USB: Universal-Scale Object Detection Benchmark shinya7y/UniverseNet 2021
180 Mask R-CNN (ResNet-50) –-0.11 Mask R-CNN tensorflow/models · facebookresearch/detectron2 · facebookresearch/detectron · +176 2017
181 EVA 57.828.86 EVA: Exploring the Limits of Masked Visual Representation Learning at Scale rwightman/pytorch-image-models · open-mmlab/mmselfsup · baaivision/eva · +3 2022
182 DETA (Swin-L) 48.520.15 NMS Strikes Back jozhang97/deta 2022
183 GLIP-L (Swin-L) 48.024.89 Grounded Language-Image Pre-training microsoft/GLIP · brown-palm/ObjectPrompt · rsCPSyEu/ovd_cod 2021
184 GRiT (ViT-H) 42.915.72 GRiT: A Generative Region-to-text Transformer for Object Understanding JialianW/GRiT 2022
185 DINO (Swin-L) 42.115.76 DINO: DETR with Improved DeNoising Anchor Boxes for End-to-End Object Detection IDEA-Research/Grounded-Segment-Anything · PaddlePaddle/PaddleDetection · lucasjinreal/yolov7_d2 · +13 2022
186 CBNetV2 (Swin-L) 39.012.36 CBNet: A Composite Backbone Network Architecture for Object Detection PaddlePaddle/PaddleDetection · shinya7y/UniverseNet · VDIGPKU/CBNetV2 · +1 2021
187 ConvNeXt-XL (Cascade Mask R-CNN) 37.512.68 A ConvNet for the 2020s keras-team/keras · rwightman/pytorch-image-models · pytorch/vision · +51 2022
188 InternImage-L (Cascade Mask R-CNN) 37.011.72 InternImage: Exploring Large-Scale Vision Foundation Models with Deformable Convolutions opengvlab/internimage · OpenGVLab/M3I-Pretraining · chenller/mmseg-extension 2022
189 DyHead (Swin-L) 35.310.00 Dynamic Head: Unifying Object Detection Heads with Attentions open-mmlab/mmdetection · microsoft/DynamicHead · Coldestadam/DynamicHead 2021
190 ViTDet (ViT-H) 34.3– Exploring Plain Vision Transformer Backbones for Object Detection facebookresearch/detectron2 · PaddlePaddle/PaddleDetection · alibaba/EasyCV · +8 2022
191 ViT-Adapter (BEiTv2-L) 34.257.79 Vision Transformer Adapter for Dense Predictions czczup/vit-adapter · chenller/mmseg-extension 2022
192 FIBER-B (Swin-B) 33.711.43 Coarse-to-Fine Vision-Language Pre-training with Fusion in the Backbone microsoft/fiber 2022
193 QueryInst (Swin-L) 33.28.26 Instances as Queries open-mmlab/mmdetection · hustvl/QueryInst · Bo396543018/picodet_repro · +2 2021
194 YOLOv6-L6 32.56.73 YOLOv6: A Single-Stage Object Detection Framework for Industrial Applications PaddlePaddle/PaddleDetection · meituan/yolov6 · open-mmlab/mmyolo · +4 2022
195 YOLOv7-E6E 32.06.42 YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors pjreddie/darknet · AlexeyAB/darknet · wongkinyiu/yolov7 · +18 2022
196 MViTV2-H (Cascade Mask R-CNN) 30.95.62 MViTv2: Improved Multiscale Vision Transformers for Classification and Detection rwightman/pytorch-image-models · facebookresearch/detectron2 · facebookresearch/SlowFast · +6 2021
197 Det-AdvProp (EfficientNet-B5) 30.87.34 Robust and Accurate Object Detection via Adversarial Learning google/automl · MindSpore-scientific-2/code-5 · MindSpore-scientific-2/code-4 2021
198 YOLOv4-P6 30.45.89 YOLOv4: Optimal Speed and Accuracy of Object Detection tensorflow/models · pjreddie/darknet · AlexeyAB/darknet · +220 2020
199 YOLOX-X 30.37.26 YOLOX: Exceeding YOLO Series in 2021 open-mmlab/mmdetection · PaddlePaddle/PaddleDetection · Megvii-BaseDetection/YOLOX · +39 2021
200 CenterNet2 (R2-101-DCN) 29.54.29 Probabilistic two-stage detection xingyizhou/CenterNet2 · smart-car-lab/Centernet2-mmdetction · aim-uofa/DiverGen 2021
← 이전 101–200 / 225 다음 → 페이지당 10 20 50 100