Tasks SotA Datasets Papers Methods Submit About

Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable Benchmarks All SotA Datasets Papers Methods

Community

Submit Results About

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

Models/SRN

SRN

Reported on 109 benchmarks across 21 tasks · 4 papers · 86 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Computer Vision42 results

Scene ParsingonSVT
Accuracy· 2020-03-27
91.5
best: 99.1 (CLIP4STR-H (DFN-5B))
SOTA
Towards Accurate Scene Text Recognition with Semantic Reasoning Networks arXiv:2003.12294
Scene ParsingonICDAR2013
Accuracy· 2020-03-27
95.5
best: 99.42 (CLIP4STR-L*)
SOTA
Towards Accurate Scene Text Recognition with Semantic Reasoning Networks arXiv:2003.12294
Scene Text RecognitiononSVT
Accuracy· 2020-03-27
91.5
best: 99.1 (CLIP4STR-H (DFN-5B))
SOTA
Towards Accurate Scene Text Recognition with Semantic Reasoning Networks arXiv:2003.12294
Scene Text RecognitiononICDAR2013
Accuracy· 2020-03-27
95.5
best: 99.42 (CLIP4STR-L*)
SOTA
Towards Accurate Scene Text Recognition with Semantic Reasoning Networks arXiv:2003.12294
Face DetectiononWIDER Face (Medium)
AP· 2018-09-07
0.948
best: 0.965 (ASFD-D6)
SOTA
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
Face DetectiononAnnotated Faces in the Wild
AP· 2018-09-07
0.9987
SOTA
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
Face DetectiononPASCAL Face
AP· 2018-09-07
0.9909
SOTA
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
Face ReconstructiononWIDER Face (Medium)
AP· 2018-09-07
0.948
best: 0.965 (ASFD-D6)
SOTA
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
Face ReconstructiononAnnotated Faces in the Wild
AP· 2018-09-07
0.9987
SOTA
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
Face ReconstructiononPASCAL Face
AP· 2018-09-07
0.9909
SOTA
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
3D Face ReconstructiononWIDER Face (Medium)
AP· 2018-09-07
0.948
best: 0.965 (ASFD-D6)
SOTA
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
3D Face ReconstructiononAnnotated Faces in the Wild
AP· 2018-09-07
0.9987
SOTA
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
3D Face ReconstructiononPASCAL Face
AP· 2018-09-07
0.9909
SOTA
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
DeblurringonRealBlur-J
PSNR (sRGB)· 2018-02-06
31.38
best: 33.96 (AdaRevD)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
DeblurringonRealBlur-J
Params(M)· 2018-02-06
8.06
best: 22.2 (MAXIM)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
DeblurringonRealBlur-J
SSIM (sRGB)· 2018-02-06
0.909
best: 0.946 (ALGNet)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
DeblurringonRealBlur-R
PSNR (sRGB)· 2018-02-06
38.65
best: 41.19 (AdaRevD)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
DeblurringonRealBlur-R
Params· 2018-02-06
8.06
best: 20 (Stripformer)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
DeblurringonRealBlur-R
SSIM (sRGB)· 2018-02-06
0.965
best: 0.981 (ALGNet)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
DeblurringonGoPro
SSIM· uses extra data· 2018-02-06
0.9342
best: 0.9792 (BSSTNet)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
DeblurringonRealBlur-R (trained on GoPro)
SSIM (sRGB)· 2018-02-06
0.947
best: 0.961 (ALGNet)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
DeblurringonRealBlur-J (trained on GoPro)
PSNR (sRGB)· 2018-02-06
28.56
best: 30.12 (AdaRevD)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
DeblurringonHIDE (trained on GOPRO)
PSNR (sRGB)· 2018-02-06
28.36
best: 32.83 (MAXIM)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
DeblurringonHIDE (trained on GOPRO)
Params (M)· 2018-02-06
8.06
best: 50.88 (Uformer-B)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
DeblurringonHIDE (trained on GOPRO)
SSIM (sRGB)· 2018-02-06
0.915
best: 0.956 (MAXIM)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
Image Enhancementon VIDIT’20 validation set
LPIPS· 2018-02-06
0.4319
best: 0.2733 (OIDDR-Net)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
Image Enhancementon VIDIT’20 validation set
MPS· 2018-02-06
0.567
best: 0.6956 (OIDDR-Net)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
Image Enhancementon VIDIT’20 validation set
PSNR· 2018-02-06
16.94
best: 17.62 (OIDDR-Net)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
Image Enhancementon VIDIT’20 validation set
Runtime(s)· 2018-02-06
0.87
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
Image Enhancementon VIDIT’20 validation set
SSIM· 2018-02-06
0.566
best: 0.6645 (OIDDR-Net)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
Image DeblurringonGoPro
Params (M)· uses extra data· 2018-02-06
8.06
best: 80.3 (UFPDeblur)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
Image DeblurringonGoPro
SSIM· uses extra data· 2018-02-06
0.9342
best: 0.972 (AdaRevD)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
Face DetectiononWIDER Face (Easy)
AP· 2018-09-07
0.959
best: 0.972 (ASFD-D6)
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
Face DetectiononFDDB
AP· 2018-09-07
0.988
best: 0.991 (DSFD)
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
Face DetectiononWIDER Face (Hard)
AP· 2018-09-07
0.896
best: 0.934 (TinaFace(ResNet-50))
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
Face ReconstructiononWIDER Face (Easy)
AP· 2018-09-07
0.959
best: 0.972 (ASFD-D6)
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
Face ReconstructiononFDDB
AP· 2018-09-07
0.988
best: 0.991 (DSFD)
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
Face ReconstructiononWIDER Face (Hard)
AP· 2018-09-07
0.896
best: 0.934 (TinaFace(ResNet-50))
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
3D Face ReconstructiononWIDER Face (Easy)
AP· 2018-09-07
0.959
best: 0.972 (ASFD-D6)
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
3D Face ReconstructiononFDDB
AP· 2018-09-07
0.988
best: 0.991 (DSFD)
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
3D Face ReconstructiononWIDER Face (Hard)
AP· 2018-09-07
0.896
best: 0.934 (TinaFace(ResNet-50))
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
Seeing Beyond the VisibleonKITTI360-EX
Average PSNR
16.1
best: 20.5 (FlowLens)

Methodology22 results

Optical Character Recognition (OCR)onBenchmarking Chinese Text Recognition: Datasets, Baselines, and an Empirical Study
Accuracy (%)· 2020-03-27
65
best: 89.6 (DTrOCR)
SOTA
Towards Accurate Scene Text Recognition with Semantic Reasoning Networks arXiv:2003.12294
3DonWIDER Face (Medium)
AP· 2018-09-07
0.948
best: 0.965 (ASFD-D6)
SOTA
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
3DonAnnotated Faces in the Wild
AP· 2018-09-07
0.9987
SOTA
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
3DonPASCAL Face
AP· 2018-09-07
0.9909
SOTA
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
2D ClassificationonRealBlur-J
PSNR (sRGB)· 2018-02-06
31.38
best: 33.96 (AdaRevD)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
2D ClassificationonRealBlur-J
Params(M)· 2018-02-06
8.06
best: 22.2 (MAXIM)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
2D ClassificationonRealBlur-J
SSIM (sRGB)· 2018-02-06
0.909
best: 0.946 (ALGNet)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
2D ClassificationonRealBlur-R
PSNR (sRGB)· 2018-02-06
38.65
best: 41.19 (AdaRevD)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
2D ClassificationonRealBlur-R
Params· 2018-02-06
8.06
best: 20 (Stripformer)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
2D ClassificationonRealBlur-R
SSIM (sRGB)· 2018-02-06
0.965
best: 0.981 (ALGNet)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
2D ClassificationonGoPro
SSIM· uses extra data· 2018-02-06
0.9342
best: 0.9792 (BSSTNet)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
2D ClassificationonRealBlur-R (trained on GoPro)
SSIM (sRGB)· 2018-02-06
0.947
best: 0.961 (ALGNet)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
2D ClassificationonRealBlur-J (trained on GoPro)
PSNR (sRGB)· 2018-02-06
28.56
best: 30.12 (AdaRevD)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
2D ClassificationonHIDE (trained on GOPRO)
PSNR (sRGB)· 2018-02-06
28.36
best: 32.83 (MAXIM)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
2D ClassificationonHIDE (trained on GOPRO)
Params (M)· 2018-02-06
8.06
best: 50.88 (Uformer-B)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
2D ClassificationonHIDE (trained on GOPRO)
SSIM (sRGB)· 2018-02-06
0.915
best: 0.956 (MAXIM)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
16konGoPro
Params (M)· uses extra data· 2018-02-06
8.06
best: 80.3 (UFPDeblur)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
16konGoPro
SSIM· uses extra data· 2018-02-06
0.9342
best: 0.972 (AdaRevD)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
Multi-Label ClassificationonNUS-WIDE
MAP· 2017-02-20
62
best: 70.1 (Q2L-CvT(resolution 384, ImageNet-21K pretrained))
SOTA
Learning Spatial Regularization with Image-level Supervisions for Multi-label Image Classification arXiv:1702.05891
3DonWIDER Face (Easy)
AP· 2018-09-07
0.959
best: 0.972 (ASFD-D6)
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
3DonFDDB
AP· 2018-09-07
0.988
best: 0.991 (DSFD)
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
3DonWIDER Face (Hard)
AP· 2018-09-07
0.896
best: 0.934 (TinaFace(ResNet-50))
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693

Audio20 results

2D Semantic SegmentationonSVT
Accuracy· 2020-03-27
91.5
best: 99.1 (CLIP4STR-H (DFN-5B))
SOTA
Towards Accurate Scene Text Recognition with Semantic Reasoning Networks arXiv:2003.12294
2D Semantic SegmentationonICDAR2013
Accuracy· 2020-03-27
95.5
best: 99.42 (CLIP4STR-L*)
SOTA
Towards Accurate Scene Text Recognition with Semantic Reasoning Networks arXiv:2003.12294
10-shot image generationonRealBlur-J
PSNR (sRGB)· 2018-02-06
31.38
best: 33.96 (AdaRevD)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
10-shot image generationonRealBlur-J
Params(M)· 2018-02-06
8.06
best: 22.2 (MAXIM)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
10-shot image generationonRealBlur-J
SSIM (sRGB)· 2018-02-06
0.909
best: 0.946 (ALGNet)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
10-shot image generationonRealBlur-R
PSNR (sRGB)· 2018-02-06
38.65
best: 41.19 (AdaRevD)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
10-shot image generationonRealBlur-R
Params· 2018-02-06
8.06
best: 20 (Stripformer)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
10-shot image generationonRealBlur-R
SSIM (sRGB)· 2018-02-06
0.965
best: 0.981 (ALGNet)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
10-shot image generationonGoPro
SSIM· uses extra data· 2018-02-06
0.9342
best: 0.9792 (BSSTNet)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
10-shot image generationonRealBlur-R (trained on GoPro)
SSIM (sRGB)· 2018-02-06
0.947
best: 0.961 (ALGNet)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
10-shot image generationonRealBlur-J (trained on GoPro)
PSNR (sRGB)· 2018-02-06
28.56
best: 30.12 (AdaRevD)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
10-shot image generationonHIDE (trained on GOPRO)
PSNR (sRGB)· 2018-02-06
28.36
best: 32.83 (MAXIM)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
10-shot image generationonHIDE (trained on GOPRO)
Params (M)· 2018-02-06
8.06
best: 50.88 (Uformer-B)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
10-shot image generationonHIDE (trained on GOPRO)
SSIM (sRGB)· 2018-02-06
0.915
best: 0.956 (MAXIM)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
10-shot image generationonGoPro
Params (M)· uses extra data· 2018-02-06
8.06
best: 80.3 (UFPDeblur)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
1 Image, 2*2 StitchionGoPro
Params (M)· uses extra data· 2018-02-06
8.06
best: 80.3 (UFPDeblur)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
1 Image, 2*2 StitchionGoPro
SSIM· uses extra data· 2018-02-06
0.9342
best: 0.972 (AdaRevD)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
10-shot image generationonGoPro
SSIM· uses extra data· 2018-02-06
0.9342
best: 0.9792 (BSSTNet)
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
10-shot image generationonShapeNet-Part
Class Average IoU
82.2
best: 87.7 (Feature Geometric Net (FG-Net))
10-shot image generationonShapeNet-Part
Instance Average IoU
85.3
best: 89.1 (GeomGCNN)

Computer Code12 results

Blind Image DeblurringonRealBlur-J
PSNR (sRGB)· 2018-02-06
31.38
best: 33.96 (AdaRevD)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
Blind Image DeblurringonRealBlur-J
Params(M)· 2018-02-06
8.06
best: 22.2 (MAXIM)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
Blind Image DeblurringonRealBlur-J
SSIM (sRGB)· 2018-02-06
0.909
best: 0.946 (ALGNet)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
Blind Image DeblurringonRealBlur-R
PSNR (sRGB)· 2018-02-06
38.65
best: 41.19 (AdaRevD)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
Blind Image DeblurringonRealBlur-R
Params· 2018-02-06
8.06
best: 20 (Stripformer)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
Blind Image DeblurringonRealBlur-R
SSIM (sRGB)· 2018-02-06
0.965
best: 0.981 (ALGNet)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
Blind Image DeblurringonGoPro
SSIM· uses extra data· 2018-02-06
0.9342
best: 0.9792 (BSSTNet)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
Blind Image DeblurringonRealBlur-R (trained on GoPro)
SSIM (sRGB)· 2018-02-06
0.947
best: 0.961 (ALGNet)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
Blind Image DeblurringonRealBlur-J (trained on GoPro)
PSNR (sRGB)· 2018-02-06
28.56
best: 30.12 (AdaRevD)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
Blind Image DeblurringonHIDE (trained on GOPRO)
PSNR (sRGB)· 2018-02-06
28.36
best: 32.83 (MAXIM)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
Blind Image DeblurringonHIDE (trained on GOPRO)
Params (M)· 2018-02-06
8.06
best: 50.88 (Uformer-B)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770
Blind Image DeblurringonHIDE (trained on GOPRO)
SSIM (sRGB)· 2018-02-06
0.915
best: 0.956 (MAXIM)
SOTA
Scale-recurrent Network for Deep Image Deblurring arXiv:1802.01770

Medical8 results

3D Face ModellingonWIDER Face (Medium)
AP· 2018-09-07
0.948
best: 0.965 (ASFD-D6)
SOTA
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
3D Face ModellingonAnnotated Faces in the Wild
AP· 2018-09-07
0.9987
SOTA
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
3D Face ModellingonPASCAL Face
AP· 2018-09-07
0.9909
SOTA
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
3D Face ModellingonWIDER Face (Easy)
AP· 2018-09-07
0.959
best: 0.972 (ASFD-D6)
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
3D Face ModellingonFDDB
AP· 2018-09-07
0.988
best: 0.991 (DSFD)
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
3D Face ModellingonWIDER Face (Hard)
AP· 2018-09-07
0.896
best: 0.934 (TinaFace(ResNet-50))
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
Semantic SegmentationonShapeNet-Part
Class Average IoU
82.2
best: 87.7 (Feature Geometric Net (FG-Net))
Semantic SegmentationonShapeNet-Part
Instance Average IoU
85.3
best: 89.1 (GeomGCNN)

Music6 results

Facial Recognition and ModellingonWIDER Face (Medium)
AP· 2018-09-07
0.948
best: 0.965 (ASFD-D6)
SOTA
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
Facial Recognition and ModellingonAnnotated Faces in the Wild
AP· 2018-09-07
0.9987
SOTA
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
Facial Recognition and ModellingonPASCAL Face
AP· 2018-09-07
0.9909
SOTA
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
Facial Recognition and ModellingonWIDER Face (Easy)
AP· 2018-09-07
0.959
best: 0.972 (ASFD-D6)
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
Facial Recognition and ModellingonFDDB
AP· 2018-09-07
0.988
best: 0.991 (DSFD)
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693
Facial Recognition and ModellingonWIDER Face (Hard)
AP· 2018-09-07
0.896
best: 0.934 (TinaFace(ResNet-50))
Selective Refinement Network for High Performance Face Detection arXiv:1809.02693