Visual Speech Recognition on DVSpeaker Matched-scene
100SI Accuracy (0°)Get
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Get#Subjects=50, Mode=Dyn.+Sta., Backbone=Transformer, Representation=Group Token2026.04 | 100 | 100 | 99.6 | 99.8 | |
| EV-Gait#Subjects=50, Mode=Dyn.+Sta., Backbone=CNN, Representation=Frame2026.04 | 100 | 100 | 100 | 100 | |
| EST#Subjects=50, Mode=Dyn.+Sta., Backbone=VGG, Representation=Spike Tensor2026.04 | 100 | 99.9 | 100 | 100 | |
| Tin#Subjects=50, Mode=Dyn.+Sta., Backbone=ResNet, Representation=Frame2026.04 | 100 | 100 | 99.8 | 99.4 | |
| Slowonly#Subjects=50, Mode=Dyn.+Sta., Backbone=ResNet, Representation=Frame2026.04 | 100 | 100 | 99.9 | 100 | |
| R2plus1d#Subjects=50, Mode=Dyn.+Sta., Backbone=ResNet, Representation=Frame2026.04 | 100 | 99.9 | 100 | 100 | |
| Tsn#Subjects=50, Mode=Dyn.+Sta., Backbone=ResNet, Representation=Frame2026.04 | 100 | 100 | 100 | 100 | |
| C3d#Subjects=50, Mode=Dyn.+Sta., Backbone=CNN, Representation=Frame2026.04 | 100 | 100 | 100 | 100 | |
| NeuroLip#Subjects=50, Mode=Dyn.+Sta., Backbone=ResNet, Representation=Weighted Voxel2026.04 | 100 | 100 | 100 | 100 | |
| Slowfast#Subjects=50, Mode=Dyn.+Sta., Backbone=ResNet, Representation=Frame2026.04 | 99.9 | 99.8 | 99.2 | 100 | |
| Tsm#Subjects=50, Mode=Dyn.+Sta., Backbone=ResNet, Representation=Frame2026.04 | 99.9 | 99.5 | 99.9 | 99.9 | |
| MSTP#Subjects=50, Mode=Dyn.+Sta., Backbone=ResNet, Representation=Frame2026.04 | 99.8 | 99.4 | 99.1 | 99.8 | |
| Tanet#Subjects=50, Mode=Dyn.+Sta., Backbone=ResNet, Representation=Frame2026.04 | 99.8 | 99.9 | 99.8 | 99.9 | |
| C2d#Subjects=50, Mode=Dyn.+Sta., Backbone=ResNet, Representation=Frame2026.04 | 99.8 | 98.9 | 99.1 | 100 | |
| Trn#Subjects=50, Mode=Dyn.+Sta., Backbone=Inception, Representation=Frame2026.04 | 97.9 | 99.3 | 99.8 | 97.9 | |
| Timesformer#Subjects=50, Mode=Dyn.+Sta., Backbone=Transformer, Representation=Frame2026.04 | 95 | 97.5 | 96.6 | 89.9 | |
| MTGA#Subjects=50, Mode=Dyn.+Sta., Backbone=Transformer, Representation=Frame+Graph2026.04 | 93.85 | 97.58 | 96.98 | 94.56 | |
| ANN#Subjects=50, Mode=Dyn.+Sta., Backbone=BiGRU, Representation=Frame2026.04 | 93.3 | 92.4 | 96.6 | 81.8 | |
| SpikGRU2+#Subjects=50, Mode=Dyn.+Sta., Backbone=Spiking GRU, Representation=Raw2026.04 | 88.4 | 87.5 | 94.4 | 85.5 | |
| EgoEvGesture#Subjects=50, Mode=Dyn.+Sta., Backbone=ResNet, Representation=Event Surfaces2026.04 | 80.7 | 83.9 | 83.6 | 71.6 | |
| Uniformerv2#Subjects=50, Mode=Dyn.+Sta., Backbone=Transformer, Representation=Frame2026.04 | 35 | 51.9 | 40.6 | 25.6 |