Lip-reading Classification on LRW (test)
88.5AccuracyProposed Method (MVM)
Evaluation results
| Method | Links | |
|---|---|---|
| Proposed Method (MVM)Backbone / Architecture=R18 + MS-TCN / MVM2022.04 | 88.5 | |
| Ma et al.Backbone / Architecture=R18 + DC-TCN2022.04 | 88.4 | |
| Ma et al.Backbone / Architecture=R18 + MS-TCN / LIRA2022.04 | 88.1 | |
| Ma et al.Backbone / Architecture=R18 + MS-TCN / Born-Again2022.04 | 87.9 | |
| Kim et al.Backbone / Architecture=R18 + BIGRU / Mem2022.04 | 85.4 | |
| Multi-Scale TCNBackbone=ResNet18*2020.01 | 85.3 | |
| Martinez et al.Backbone / Architecture=R18 + MS-TCN2022.04 | 85.3 | |
| Zhang et al.Backbone / Architecture=R18 + BIGRU / Face Cutout2022.04 | 85 | |
| Xu et al.Backbone / Architecture=P3D R50 + BILSTM2022.04 | 84.8 | |
| GLMIMfull_method=Proposed GLMIM2020.03 | 84.41 | |
| Zhao et al.Backbone / Architecture=R18 + BIGRU + LSTM2022.04 | 84.4 | |
| DFTN2020.03 | 84.13 | |
| 2-stream 3DCNNBackbone=(3D ResNet34) x 22020.01 | 84.1 | |
| Weng and KitaniBackbone / Architecture=I3D + BiLSTM2022.04 | 84.1 | |
| Xiao et al.Backbone / Architecture=R18 + BiGRU2022.04 | 84.1 | |
| Weng192020.03 | 84.07 | |
| PCPG2020.03 | 83.5 | |
| Luo et al.Backbone / Architecture=R18 + BIGRU + GRU2022.04 | 83.5 | |
| End-to-end AVRBackbone=ResNet18*2020.01 | 83.4 | |
| WangYear=20192020.03 | 83.34 | |
| Wang2020.03 | 83.34 | |
| Baseline+LMIMmodule=LMIM2020.03 | 83.33 | |
| Multi-GrainedBackbone=ResNet34 + DenseNet3D2020.01 | 83.3 | |
| ResNet + LSTMBackbone=ResNet34*2020.01 | 83 | |
| StafylakisYear=20172020.03 | 83 | |
| Stafylakis172020.03 | 83 | |
| Stafylakis and Tzimiropoulos2020.03 | 83 | |
| Stafylakis et al.2020.03 | 82.9 | |
| Baseline2020.03 | 82.14 | |
| PetridisYear=20182020.03 | 82 | |
| Petridis et al.2020.03 | 82 | |
| Stafylakis17reproduced=true2020.03 | 77.8 | |
| WASBackbone=VGG-M2020.01 | 76.2 | |
| ChungYear=20172020.03 | 76.2 | |
| Chung172020.03 | 76.2 | |
| ChungYear=20182020.03 | 71.5 | |
| LRWBackbone=VGG-M2020.01 | 61.1 | |
| Chung162020.03 | 61.1 |