TasksSotADatasetsPapersMethodsSubmitAbout
Papers With Code 2

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

Explore

Notable BenchmarksAll SotADatasetsPapersMethods

Community

Submit ResultsAbout

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

Models/DAN

DAN

Reported on 157 benchmarks across 21 tasks · 10 papers · 93 SOTA

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

Computer Vision59 results

  • Face ReconstructiononRAF-DB
    Overall Accuracy· 2021-09-15
    89.7
    best: 94.76 (ResEmoteNet)
    SOTA
    Distract Your Attention: Multi-head Cross Attention Network for Facial Expression RecognitionarXiv:2109.07270
  • Facial Expression Recognition (FER)onRAF-DB
    Overall Accuracy· 2021-09-15
    89.7
    best: 94.76 (ResEmoteNet)
    SOTA
    Distract Your Attention: Multi-head Cross Attention Network for Facial Expression RecognitionarXiv:2109.07270
  • 3D Face ReconstructiononRAF-DB
    Overall Accuracy· 2021-09-15
    89.7
    best: 94.76 (ResEmoteNet)
    SOTA
    Distract Your Attention: Multi-head Cross Attention Network for Facial Expression RecognitionarXiv:2109.07270
  • Image RestorationonBSD100 - 2x upscaling
    PSNR· 2020-10-06
    31.76
    best: 32.04 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image RestorationonBSD100 - 2x upscaling
    SSIM· 2020-10-06
    0.8858
    best: 0.8907 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image RestorationonSet14 - 4x upscaling
    PSNR· 2020-10-06
    28.43
    best: 28.54 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image RestorationonSet14 - 4x upscaling
    SSIM· 2020-10-06
    0.7693
    best: 0.7728 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image RestorationonManga109 - 4x upscaling
    PSNR· 2020-10-06
    30.5
    best: 30.86 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image RestorationonManga109 - 4x upscaling
    SSIM· 2020-10-06
    0.9037
    best: 0.9086 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image RestorationonBSD100 - 4x upscaling
    PSNR· 2020-10-06
    27.51
    best: 27.6 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image RestorationonSet5 - 2x upscaling
    PSNR· 2020-10-06
    37.33
    best: 37.63 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image RestorationonSet5 - 4x upscaling
    PSNR· 2020-10-06
    31.89
    best: 32.12 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image RestorationonUrban100 - 2x upscaling
    PSNR· 2020-10-06
    30.6
    best: 31.69 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image RestorationonUrban100 - 2x upscaling
    SSIM· 2020-10-06
    0.902
    best: 0.9202 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image RestorationonSet14 - 2x upscaling
    PSNR· 2020-10-06
    33.07
    best: 33.46 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image RestorationonSet14 - 2x upscaling
    SSIM· 2020-10-06
    0.9068
    best: 0.9103 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image RestorationonUrban100 - 4x upscaling
    PSNR· 2020-10-06
    25.86
    best: 26.15 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image RestorationonDIV2KRK - 2x upscaling
    PSNR· 2020-10-06
    32.56
    best: 32.75 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image RestorationonDIV2KRK - 2x upscaling
    SSIM· 2020-10-06
    0.8997
    best: 0.9094 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image RestorationonManga109 - 2x upscaling
    PSNR· 2020-10-06
    37.23
    best: 38.31 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image RestorationonManga109 - 2x upscaling
    SSIM· 2020-10-06
    0.971
    best: 0.974 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Face Reconstructionon300W Split 2
    NME (inter-ocular)· 2019-02-25
    4.3
    SOTA
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • 3D Face Reconstructionon300W Split 2
    NME (inter-ocular)· 2019-02-25
    4.3
    SOTA
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • Visual DialogonVisDial v0.9 val
    MRR· 2019-02-25
    66.38
    best: 68.92 (9xFGA (VGG))
    SOTA
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • Face Reconstructionon300W Split 2
    AUC@8 (inter-ocular)· 2017-06-06
    47
    best: 57.27 (SPIGA)
    SOTA
    Deep Alignment Network: A convolutional neural network for robust face alignmentarXiv:1706.01789
  • Face Reconstructionon300W Split 2
    FR@8 (inter-ocular)· 2017-06-06
    2.67
    SOTA
    Deep Alignment Network: A convolutional neural network for robust face alignmentarXiv:1706.01789
  • 3D Face Reconstructionon300W Split 2
    AUC@8 (inter-ocular)· 2017-06-06
    47
    best: 57.27 (SPIGA)
    SOTA
    Deep Alignment Network: A convolutional neural network for robust face alignmentarXiv:1706.01789
  • 3D Face Reconstructionon300W Split 2
    FR@8 (inter-ocular)· 2017-06-06
    2.67
    SOTA
    Deep Alignment Network: A convolutional neural network for robust face alignmentarXiv:1706.01789
  • Image RetrievalonFlickr30K 1K test
    R@1· 2016-11-02
    39.4
    best: 86.9 (X-VLM (base))
    SOTA
    Dual Attention Networks for Multimodal Reasoning and MatchingarXiv:1611.00471
  • Image RetrievalonFlickr30K 1K test
    R@10· 2016-11-02
    79.1
    best: 98.7 (X-VLM (base))
    SOTA
    Dual Attention Networks for Multimodal Reasoning and MatchingarXiv:1611.00471
  • Image RetrievalonFlickr30K 1K test
    R@5· 2016-11-02
    69.2
    best: 97.3 (X-VLM (base))
    SOTA
    Dual Attention Networks for Multimodal Reasoning and MatchingarXiv:1611.00471
  • Face ReconstructiononAffectNet
    Accuracy (7 emotion)· 2021-09-15
    65.69
    best: 72.93 (ResEmoteNet)
    Distract Your Attention: Multi-head Cross Attention Network for Facial Expression RecognitionarXiv:2109.07270
  • Face ReconstructiononAffectNet
    Accuracy (8 emotion)· 2021-09-15
    62.09
    best: 68.69 (Norface)
    Distract Your Attention: Multi-head Cross Attention Network for Facial Expression RecognitionarXiv:2109.07270
  • Facial Expression Recognition (FER)onAffectNet
    Accuracy (7 emotion)· 2021-09-15
    65.69
    best: 72.93 (ResEmoteNet)
    Distract Your Attention: Multi-head Cross Attention Network for Facial Expression RecognitionarXiv:2109.07270
  • Facial Expression Recognition (FER)onAffectNet
    Accuracy (8 emotion)· 2021-09-15
    62.09
    best: 68.69 (Norface)
    Distract Your Attention: Multi-head Cross Attention Network for Facial Expression RecognitionarXiv:2109.07270
  • 3D Face ReconstructiononAffectNet
    Accuracy (7 emotion)· 2021-09-15
    65.69
    best: 72.93 (ResEmoteNet)
    Distract Your Attention: Multi-head Cross Attention Network for Facial Expression RecognitionarXiv:2109.07270
  • 3D Face ReconstructiononAffectNet
    Accuracy (8 emotion)· 2021-09-15
    62.09
    best: 68.69 (Norface)
    Distract Your Attention: Multi-head Cross Attention Network for Facial Expression RecognitionarXiv:2109.07270
  • Image RestorationonBSD100 - 4x upscaling
    SSIM· 2020-10-06
    0.7248
    best: 0.8014 (IKC)
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image RestorationonSet5 - 2x upscaling
    SSIM· 2020-10-06
    0.9526
    best: 0.9658 (IKC)
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image RestorationonSet5 - 4x upscaling
    SSIM· 2020-10-06
    0.8864
    best: 0.9278 (IKC)
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image RestorationonUrban100 - 4x upscaling
    SSIM· 2020-10-06
    0.7721
    best: 0.7809 (DCLS)
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Scene ParsingonSVT
    Accuracy· 2019-12-21
    89.2
    best: 99.1 (CLIP4STR-H (DFN-5B))
    Decoupled Attention Network for Text RecognitionarXiv:1912.10205
  • Scene ParsingonICDAR2015
    Accuracy· 2019-12-21
    74.5
    best: 93.5 (DTrOCR 105M)
    Decoupled Attention Network for Text RecognitionarXiv:1912.10205
  • Scene ParsingonICDAR 2003
    Accuracy· 2019-12-21
    95
    best: 97.1 (Yet Another Text Recognizer)
    Decoupled Attention Network for Text RecognitionarXiv:1912.10205
  • Scene ParsingonICDAR2013
    Accuracy· 2019-12-21
    93.9
    best: 99.42 (CLIP4STR-L*)
    Decoupled Attention Network for Text RecognitionarXiv:1912.10205
  • Scene Text RecognitiononSVT
    Accuracy· 2019-12-21
    89.2
    best: 99.1 (CLIP4STR-H (DFN-5B))
    Decoupled Attention Network for Text RecognitionarXiv:1912.10205
  • Scene Text RecognitiononICDAR2015
    Accuracy· 2019-12-21
    74.5
    best: 93.5 (DTrOCR 105M)
    Decoupled Attention Network for Text RecognitionarXiv:1912.10205
  • Scene Text RecognitiononICDAR 2003
    Accuracy· 2019-12-21
    95
    best: 97.1 (Yet Another Text Recognizer)
    Decoupled Attention Network for Text RecognitionarXiv:1912.10205
  • Scene Text RecognitiononICDAR2013
    Accuracy· 2019-12-21
    93.9
    best: 99.42 (CLIP4STR-L*)
    Decoupled Attention Network for Text RecognitionarXiv:1912.10205
  • Visual DialogonVisDial v0.9 val
    Mean Rank· 2019-02-25
    4.04
    best: 5.84 (HieCoAtt-QI)
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • Visual DialogonVisDial v0.9 val
    R@1· 2019-02-25
    53.33
    best: 55.16 (9xFGA (VGG))
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • Visual DialogonVisDial v0.9 val
    R@10· 2019-02-25
    90.38
    best: 92.95 (9xFGA (VGG))
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • Visual DialogonVisDial v0.9 val
    R@5· 2019-02-25
    82.42
    best: 86.26 (9xFGA (VGG))
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • Visual DialogonVisual Dialog v1.0 test-std
    MRR (x 100)· 2019-02-25
    63.2
    best: 71.24 (MRR ensemble (Naive))
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • Visual DialogonVisual Dialog v1.0 test-std
    Mean· 2019-02-25
    4.3
    best: 49.61 (qqhe)
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • Visual DialogonVisual Dialog v1.0 test-std
    NDCG (x 100)· 2019-02-25
    57.59
    best: 78.7 (Single)
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • Visual DialogonVisual Dialog v1.0 test-std
    R@1· 2019-02-25
    49.63
    best: 58.3 (2 Step: Factor Graph Attention + VD-Bert)
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • Visual DialogonVisual Dialog v1.0 test-std
    R@10· 2019-02-25
    89.35
    best: 95.08 (Ensemble FGA + BERT)
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • Visual DialogonVisual Dialog v1.0 test-std
    R@5· 2019-02-25
    79.75
    best: 88.42 (Ensemble FGA + BERT)
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368

Medical28 results

  • 3D Face ModellingonRAF-DB
    Overall Accuracy· 2021-09-15
    89.7
    best: 94.76 (ResEmoteNet)
    SOTA
    Distract Your Attention: Multi-head Cross Attention Network for Facial Expression RecognitionarXiv:2109.07270
  • Image ReconstructiononBSD100 - 2x upscaling
    PSNR· 2020-10-06
    31.76
    best: 32.04 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image ReconstructiononBSD100 - 2x upscaling
    SSIM· 2020-10-06
    0.8858
    best: 0.8907 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image ReconstructiononSet14 - 4x upscaling
    PSNR· 2020-10-06
    28.43
    best: 28.54 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image ReconstructiononSet14 - 4x upscaling
    SSIM· 2020-10-06
    0.7693
    best: 0.7728 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image ReconstructiononManga109 - 4x upscaling
    PSNR· 2020-10-06
    30.5
    best: 30.86 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image ReconstructiononManga109 - 4x upscaling
    SSIM· 2020-10-06
    0.9037
    best: 0.9086 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image ReconstructiononBSD100 - 4x upscaling
    PSNR· 2020-10-06
    27.51
    best: 27.6 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image ReconstructiononSet5 - 2x upscaling
    PSNR· 2020-10-06
    37.33
    best: 37.63 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image ReconstructiononSet5 - 4x upscaling
    PSNR· 2020-10-06
    31.89
    best: 32.12 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image ReconstructiononUrban100 - 2x upscaling
    PSNR· 2020-10-06
    30.6
    best: 31.69 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image ReconstructiononUrban100 - 2x upscaling
    SSIM· 2020-10-06
    0.902
    best: 0.9202 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image ReconstructiononSet14 - 2x upscaling
    PSNR· 2020-10-06
    33.07
    best: 33.46 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image ReconstructiononSet14 - 2x upscaling
    SSIM· 2020-10-06
    0.9068
    best: 0.9103 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image ReconstructiononUrban100 - 4x upscaling
    PSNR· 2020-10-06
    25.86
    best: 26.15 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image ReconstructiononDIV2KRK - 2x upscaling
    PSNR· 2020-10-06
    32.56
    best: 32.75 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image ReconstructiononDIV2KRK - 2x upscaling
    SSIM· 2020-10-06
    0.8997
    best: 0.9094 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image ReconstructiononManga109 - 2x upscaling
    PSNR· 2020-10-06
    37.23
    best: 38.31 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image ReconstructiononManga109 - 2x upscaling
    SSIM· 2020-10-06
    0.971
    best: 0.974 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 3D Face Modellingon300W Split 2
    NME (inter-ocular)· 2019-02-25
    4.3
    SOTA
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • 3D Face Modellingon300W Split 2
    AUC@8 (inter-ocular)· 2017-06-06
    47
    best: 57.27 (SPIGA)
    SOTA
    Deep Alignment Network: A convolutional neural network for robust face alignmentarXiv:1706.01789
  • 3D Face Modellingon300W Split 2
    FR@8 (inter-ocular)· 2017-06-06
    2.67
    SOTA
    Deep Alignment Network: A convolutional neural network for robust face alignmentarXiv:1706.01789
  • 3D Face ModellingonAffectNet
    Accuracy (7 emotion)· 2021-09-15
    65.69
    best: 72.93 (ResEmoteNet)
    Distract Your Attention: Multi-head Cross Attention Network for Facial Expression RecognitionarXiv:2109.07270
  • 3D Face ModellingonAffectNet
    Accuracy (8 emotion)· 2021-09-15
    62.09
    best: 68.69 (Norface)
    Distract Your Attention: Multi-head Cross Attention Network for Facial Expression RecognitionarXiv:2109.07270
  • Image ReconstructiononBSD100 - 4x upscaling
    SSIM· 2020-10-06
    0.7248
    best: 0.8014 (IKC)
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image ReconstructiononSet5 - 2x upscaling
    SSIM· 2020-10-06
    0.9526
    best: 0.9658 (IKC)
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image ReconstructiononSet5 - 4x upscaling
    SSIM· 2020-10-06
    0.8864
    best: 0.9278 (IKC)
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • Image ReconstructiononUrban100 - 4x upscaling
    SSIM· 2020-10-06
    0.7721
    best: 0.7809 (DCLS)
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631

Audio26 results

  • 10-shot image generationonBSD100 - 2x upscaling
    PSNR· 2020-10-06
    31.76
    best: 32.04 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonBSD100 - 2x upscaling
    SSIM· 2020-10-06
    0.8858
    best: 0.8907 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonSet14 - 4x upscaling
    PSNR· 2020-10-06
    28.43
    best: 28.54 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonSet14 - 4x upscaling
    SSIM· 2020-10-06
    0.7693
    best: 0.7728 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonManga109 - 4x upscaling
    PSNR· 2020-10-06
    30.5
    best: 30.86 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonManga109 - 4x upscaling
    SSIM· 2020-10-06
    0.9037
    best: 0.9086 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonBSD100 - 4x upscaling
    PSNR· 2020-10-06
    27.51
    best: 27.6 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonSet5 - 2x upscaling
    PSNR· 2020-10-06
    37.33
    best: 37.63 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonSet5 - 4x upscaling
    PSNR· 2020-10-06
    31.89
    best: 32.12 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonUrban100 - 2x upscaling
    PSNR· 2020-10-06
    30.6
    best: 31.69 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonUrban100 - 2x upscaling
    SSIM· 2020-10-06
    0.902
    best: 0.9202 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonSet14 - 2x upscaling
    PSNR· 2020-10-06
    33.07
    best: 33.46 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonSet14 - 2x upscaling
    SSIM· 2020-10-06
    0.9068
    best: 0.9103 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonUrban100 - 4x upscaling
    PSNR· 2020-10-06
    25.86
    best: 26.15 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonDIV2KRK - 2x upscaling
    PSNR· 2020-10-06
    32.56
    best: 32.75 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonDIV2KRK - 2x upscaling
    SSIM· 2020-10-06
    0.8997
    best: 0.9094 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonManga109 - 2x upscaling
    PSNR· 2020-10-06
    37.23
    best: 38.31 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonManga109 - 2x upscaling
    SSIM· 2020-10-06
    0.971
    best: 0.974 (DCLS)
    SOTA
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonBSD100 - 4x upscaling
    SSIM· 2020-10-06
    0.7248
    best: 0.8014 (IKC)
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonSet5 - 2x upscaling
    SSIM· 2020-10-06
    0.9526
    best: 0.9658 (IKC)
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonSet5 - 4x upscaling
    SSIM· 2020-10-06
    0.8864
    best: 0.9278 (IKC)
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 10-shot image generationonUrban100 - 4x upscaling
    SSIM· 2020-10-06
    0.7721
    best: 0.7809 (DCLS)
    Unfolding the Alternating Optimization for Blind Super ResolutionarXiv:2010.02631
  • 2D Semantic SegmentationonSVT
    Accuracy· 2019-12-21
    89.2
    best: 99.1 (CLIP4STR-H (DFN-5B))
    Decoupled Attention Network for Text RecognitionarXiv:1912.10205
  • 2D Semantic SegmentationonICDAR2015
    Accuracy· 2019-12-21
    74.5
    best: 93.5 (DTrOCR 105M)
    Decoupled Attention Network for Text RecognitionarXiv:1912.10205
  • 2D Semantic SegmentationonICDAR 2003
    Accuracy· 2019-12-21
    95
    best: 97.1 (Yet Another Text Recognizer)
    Decoupled Attention Network for Text RecognitionarXiv:1912.10205
  • 2D Semantic SegmentationonICDAR2013
    Accuracy· 2019-12-21
    93.9
    best: 99.42 (CLIP4STR-L*)
    Decoupled Attention Network for Text RecognitionarXiv:1912.10205

Methodology16 results

  • Optical Character Recognition (OCR)onSIMARA
    CER (%)· 2023-04-26
    6.46
    SOTA
    SIMARA: a database for key-value information extraction from full pagesarXiv:2304.13606
  • Optical Character Recognition (OCR)onSIMARA
    WER (%)· 2023-04-26
    14.79
    SOTA
    SIMARA: a database for key-value information extraction from full pagesarXiv:2304.13606
  • Optical Character Recognition (OCR)onREAD 2016
    CER (%)· 2022-03-23
    3.22
    SOTA
    DAN: a Segmentation-free Document Attention Network for Handwritten Document RecognitionarXiv:2203.12273
  • Optical Character Recognition (OCR)onREAD 2016
    WER (%)· 2022-03-23
    13.63
    SOTA
    DAN: a Segmentation-free Document Attention Network for Handwritten Document RecognitionarXiv:2203.12273
  • 3DonRAF-DB
    Overall Accuracy· 2021-09-15
    89.7
    best: 94.76 (ResEmoteNet)
    SOTA
    Distract Your Attention: Multi-head Cross Attention Network for Facial Expression RecognitionarXiv:2109.07270
  • 3Don300W Split 2
    NME (inter-ocular)· 2019-02-25
    4.3
    SOTA
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • 3Don300W Split 2
    AUC@8 (inter-ocular)· 2017-06-06
    47
    best: 57.27 (SPIGA)
    SOTA
    Deep Alignment Network: A convolutional neural network for robust face alignmentarXiv:1706.01789
  • 3Don300W Split 2
    FR@8 (inter-ocular)· 2017-06-06
    2.67
    SOTA
    Deep Alignment Network: A convolutional neural network for robust face alignmentarXiv:1706.01789
  • Continual Learningonvisual domain decathlon (10 tasks)
    Avg. Accuracy· 2017-05-11
    77.01
    best: 79.64 (NetTailor)
    SOTA
    Incremental Learning Through Deep AdaptationarXiv:1705.04228
  • Continual Learningonvisual domain decathlon (10 tasks)
    decathlon discipline (Score)· 2017-05-11
    2851
    best: 3744 (NetTailor)
    SOTA
    Incremental Learning Through Deep AdaptationarXiv:1705.04228
  • Domain AdaptationonSYNSIG-to-GTSRB
    Accuracy· 2015-02-10
    91.1
    best: 97.5 (DFA-MCD)
    SOTA
    Learning Transferable Features with Deep Adaptation NetworksarXiv:1502.02791
  • Domain AdaptationonImageCLEF-DA
    Accuracy· 2015-02-10
    76.9
    best: 94.3 (CMKD)
    SOTA
    Learning Transferable Features with Deep Adaptation NetworksarXiv:1502.02791
  • Optical Character Recognition (OCR)onREAD2016(line-level)
    Test CER· 2022-03-23
    4.1
    best: 3.9 (HTR-VT)
    DAN: a Segmentation-free Document Attention Network for Handwritten Document RecognitionarXiv:2203.12273
  • Optical Character Recognition (OCR)onREAD2016(line-level)
    Test WER· 2022-03-23
    17.6
    best: 16.3 (VAN)
    DAN: a Segmentation-free Document Attention Network for Handwritten Document RecognitionarXiv:2203.12273
  • 3DonAffectNet
    Accuracy (7 emotion)· 2021-09-15
    65.69
    best: 72.93 (ResEmoteNet)
    Distract Your Attention: Multi-head Cross Attention Network for Facial Expression RecognitionarXiv:2109.07270
  • 3DonAffectNet
    Accuracy (8 emotion)· 2021-09-15
    62.09
    best: 68.69 (Norface)
    Distract Your Attention: Multi-head Cross Attention Network for Facial Expression RecognitionarXiv:2109.07270

Speech11 results

  • DialogueonVisDial v0.9 val
    MRR· 2019-02-25
    66.38
    best: 68.92 (9xFGA (VGG))
    SOTA
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • DialogueonVisDial v0.9 val
    Mean Rank· 2019-02-25
    4.04
    best: 5.84 (HieCoAtt-QI)
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • DialogueonVisDial v0.9 val
    R@1· 2019-02-25
    53.33
    best: 55.16 (9xFGA (VGG))
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • DialogueonVisDial v0.9 val
    R@10· 2019-02-25
    90.38
    best: 92.95 (9xFGA (VGG))
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • DialogueonVisDial v0.9 val
    R@5· 2019-02-25
    82.42
    best: 86.26 (9xFGA (VGG))
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • DialogueonVisual Dialog v1.0 test-std
    MRR (x 100)· 2019-02-25
    63.2
    best: 71.24 (MRR ensemble (Naive))
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • DialogueonVisual Dialog v1.0 test-std
    Mean· 2019-02-25
    4.3
    best: 49.61 (qqhe)
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • DialogueonVisual Dialog v1.0 test-std
    NDCG (x 100)· 2019-02-25
    57.59
    best: 78.7 (Single)
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • DialogueonVisual Dialog v1.0 test-std
    R@1· 2019-02-25
    49.63
    best: 58.3 (2 Step: Factor Graph Attention + VD-Bert)
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • DialogueonVisual Dialog v1.0 test-std
    R@10· 2019-02-25
    89.35
    best: 95.08 (Ensemble FGA + BERT)
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • DialogueonVisual Dialog v1.0 test-std
    R@5· 2019-02-25
    79.75
    best: 88.42 (Ensemble FGA + BERT)
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368

Adversarial6 results

  • Handwritten Text RecognitiononSIMARA
    CER (%)· 2023-04-26
    6.46
    SOTA
    SIMARA: a database for key-value information extraction from full pagesarXiv:2304.13606
  • Handwritten Text RecognitiononSIMARA
    WER (%)· 2023-04-26
    14.79
    SOTA
    SIMARA: a database for key-value information extraction from full pagesarXiv:2304.13606
  • Handwritten Text RecognitiononREAD 2016
    CER (%)· 2022-03-23
    3.22
    SOTA
    DAN: a Segmentation-free Document Attention Network for Handwritten Document RecognitionarXiv:2203.12273
  • Handwritten Text RecognitiononREAD 2016
    WER (%)· 2022-03-23
    13.63
    SOTA
    DAN: a Segmentation-free Document Attention Network for Handwritten Document RecognitionarXiv:2203.12273
  • Handwritten Text RecognitiononREAD2016(line-level)
    Test CER· 2022-03-23
    4.1
    best: 3.9 (HTR-VT)
    DAN: a Segmentation-free Document Attention Network for Handwritten Document RecognitionarXiv:2203.12273
  • Handwritten Text RecognitiononREAD2016(line-level)
    Test WER· 2022-03-23
    17.6
    best: 16.3 (VAN)
    DAN: a Segmentation-free Document Attention Network for Handwritten Document RecognitionarXiv:2203.12273

Music6 results

  • Facial Recognition and ModellingonRAF-DB
    Overall Accuracy· 2021-09-15
    89.7
    best: 94.76 (ResEmoteNet)
    SOTA
    Distract Your Attention: Multi-head Cross Attention Network for Facial Expression RecognitionarXiv:2109.07270
  • Facial Recognition and Modellingon300W Split 2
    NME (inter-ocular)· 2019-02-25
    4.3
    SOTA
    Dual Attention Networks for Visual Reference Resolution in Visual DialogarXiv:1902.09368
  • Facial Recognition and Modellingon300W Split 2
    AUC@8 (inter-ocular)· 2017-06-06
    47
    best: 57.27 (SPIGA)
    SOTA
    Deep Alignment Network: A convolutional neural network for robust face alignmentarXiv:1706.01789
  • Facial Recognition and Modellingon300W Split 2
    FR@8 (inter-ocular)· 2017-06-06
    2.67
    SOTA
    Deep Alignment Network: A convolutional neural network for robust face alignmentarXiv:1706.01789
  • Facial Recognition and ModellingonAffectNet
    Accuracy (7 emotion)· 2021-09-15
    65.69
    best: 72.93 (ResEmoteNet)
    Distract Your Attention: Multi-head Cross Attention Network for Facial Expression RecognitionarXiv:2109.07270
  • Facial Recognition and ModellingonAffectNet
    Accuracy (8 emotion)· 2021-09-15
    62.09
    best: 68.69 (Norface)
    Distract Your Attention: Multi-head Cross Attention Network for Facial Expression RecognitionarXiv:2109.07270

Robots4 results

  • Robot NavigationonHabitat 2020 Point Nav minival
    DISTANCE_TO_GOAL
    0.27
    best: 4.33 (RandomAgent)
  • Robot NavigationonHabitat 2020 Point Nav minival
    SOFT_SPL
    0.53
    best: 0.74 (VO)
  • Robot NavigationonHabitat 2020 Point Nav minival
    SPL
    0.53
    best: 0.61 (VO)
  • Robot NavigationonHabitat 2020 Point Nav minival
    SUCCESS
    0.93

Natural Language Processing1 result

  • Key Information ExtractiononSIMARA
    F1 (%)· 2023-04-26
    95.05
    SOTA
    SIMARA: a database for key-value information extraction from full pagesarXiv:2304.13606