List your company
ResearchFormerly WizWand

Lip Reading on LRW Word-level (test)

88.8AccuracyStafylakis et al. (WB)
59.99267.47174.9582.429Nov 15, 2020
Updated 8mo ago

Evaluation results

MethodLinks
Stafylakis et al. (WB)Year=2018, Frontend=ResNet-18, Backend=BiLSTM, Data Type=Lip + Word Boundary, Input Size=112x1122020.1188.8
Refined Network (SE+MixUp+Cosine LR+LS+WB)Year=2021, Frontend=SE-ResNet-18, Backend=BiGRU, Data Type=Aligned Lip + Word Boundary, Input Size=88x882020.1188.4
Multi-Stage Distillation (Ma et al.)Year=2020, Frontend=ResNet-18, Backend=MS-TCN, Data Type=Aligned Lip, Input Size=88x882020.1187.7
Temporal Convolution (Martinez et al.)Year=2020, Frontend=ResNet-18, Backend=MS-TCN, Data Type=Aligned Lip, Input Size=88x882020.1185.3
Face Cutout (Zhang et al.)Year=2020, Frontend=ResNet-18, Backend=BiGRU, Data Type=Aligned Face, Input Size=112x1122020.1185
Refined Network (SE+MixUp+Cosine LR+LS+WB)Year=2021, Frontend=SE-ResNet-18, Backend=BiGRU, Data Type=Aligned Lip, Input Size=88x882020.1185
Mutual Information (Zhao et al.)Year=2020, Frontend=ResNet-18, Backend=BiGRU, Data Type=Lip, Input Size=88x882020.1184.4
Two Stream (Weng et al.)Year=2019, Frontend=Two-Stream ResNet-18, Backend=BiLSTM, Data Type=Lip, Input Size=112x1122020.1184.1
Deformation Flow (Xiao et al.)Year=2020, Frontend=ResNet-18, Backend=BiGRU, Data Type=Lip, Input Size=88x882020.1184.1
Stafylakis et al.Year=2017, Frontend=ResNet-34, Backend=BiLSTM, Data Type=Lip, Input Size=112x1122020.1183.5
Policy Gradient (Luo et al.)Year=2020, Frontend=ResNet-18, Backend=BiGRU, Data Type=Lip, Input Size=88x882020.1183.5
Multi-Grained (Wang et al.)Year=2019, Frontend=Multi-Grained ResNet-18, Backend=Conv BiLSTM, Data Type=Lip, Input Size=88x882020.1183.3
Chung et al.Year=2016, Frontend=VGGM, Data Type=Lip, Input Size=112x1122020.1161.1