Lipreading On Lrw 1000 1
評価指標
Top-1 Accuracy
評価結果
このベンチマークにおける各モデルのパフォーマンス結果
モデル名 | Top-1 Accuracy | Paper Title | Repository |
|---|---|---|---|
| DenseNet3D + Bi-GRU | 34.76% | LRW-1000: A Naturally-Distributed Large-Scale Benchmark for Lip Reading in the Wild | |
| 3D Conv + ResNet-34 + Bi-GRU | 38.19% | LRW-1000: A Naturally-Distributed Large-Scale Benchmark for Lip Reading in the Wild | |
| 3D Conv + ResNet-18 + MS-TCN | 41.4% | Lipreading using Temporal Convolutional Networks | |
| Multi-Tower LSTM-5 | 25.76% | LRW-1000: A Naturally-Distributed Large-Scale Benchmark for Lip Reading in the Wild |
0 of 4 row(s) selected.