Can We Read Speech Beyond the Lips? Rethinking RoI Selection for Deep Visual Speech Recognition

Benchmark Model Rank Results
lipreading-on-grid-corpus-mixed-speechLipNet (with Face Cutout)#2Word Error Rate (WER): 2.9
lipreading-on-lip-reading-in-the-wild3D Conv + ResNet-18 + Bi-GRU(Face Cutout)#11Top-1 Accuracy: 85.02
lipreading-on-lrw-10003D Conv + ResNet-18 + Bi-GRU (Face Cutout)#6Top-1 Accuracy: 45.24%