Action Recognition on Ego4D v1 (test)
25.1Top-1 Accuracy (Verb)EgoPack
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| EgoPackTrained on frozen features=true2024.03 | 25.1 | 31.1 | — | |
| Temporal GraphTrained on frozen features=true2024.03 | 24.25 | 30.43 | — | |
| MLPTrained on frozen features=true2024.03 | 24.08 | 30.45 | — | |
| Task Translation+Trained on frozen features=true2024.03 | 23.68 | 28.28 | — | |
| RepLAILAVC=true, LASTC=true, MoI=true, AVC Pretraining=true2022.09 | 23.1 | 15.8 | — | |
| EgoT2sTrained on frozen features=false2024.03 | 23.04 | 23.28 | — | |
| RepLAI w/o AVCLAVC=false, LASTC=true, MoI=true, AVC Pretraining=true2022.09 | 22.7 | 14 | — | |
| RepLAI (scratch)LAVC=true, LASTC=true, MoI=true, AVC Pretraining=false2022.09 | 22.5 | 15 | — | |
| RepLAILAVC=true, LASTC=true, MoI=true, AVC Pretraining=true2022.09 | 22.5 | 14.7 | — | |
| RepLAI (scratch)LAVC=true, LASTC=true, MoI=true, AVC Pretraining=false2022.09 | 22.2 | 14.1 | — | |
| Ego4D BaselinesTrained on frozen features=false2024.03 | 22.18 | 21.55 | — | |
| Multi-Task LearningTrained on frozen features=true2024.03 | 22.05 | 29.44 | — | |
| RepLAI w/o AStCLAVC=true, LASTC=false, MoI=true, AVC Pretraining=true2022.09 | 21.4 | 13 | — | |
| RepLAI w/o AStCLAVC=true, LASTC=false, MoI=true, AVC Pretraining=true2022.09 | 21.1 | 13.5 | — | |
| RepLAI w/o MoILAVC=true, LASTC=true, MoI=false, AVC Pretraining=true2022.09 | 20.8 | 11.7 | — | |
| RepLAI w/o AVCLAVC=false, LASTC=true, MoI=true, AVC Pretraining=true2022.09 | 20.3 | 12.4 | — | |
| AVIDLAVC=false, LASTC=false, MoI=false, AVC Pretraining=true2022.09 | 19.8 | 12.3 | — | |
| RepLAI w/o MoILAVC=true, LASTC=true, MoI=false, AVC Pretraining=true2022.09 | 19.8 | 11.2 | — | |
| XDCLAVC=false, LASTC=false, MoI=false, AVC Pretraining=false2022.09 | 19.1 | 10.2 | — | |
| RandomLAVC=false, LASTC=false, MoI=false, AVC Pretraining=false2022.09 | 18.9 | 9.5 | — | |
| AVIDLAVC=false, LASTC=false, MoI=false, AVC Pretraining=true2022.09 | 18.3 | 10.7 | — | |
| XDCLAVC=false, LASTC=false, MoI=false, AVC Pretraining=true2022.09 | 17.9 | 8.7 | — | |
| RandomLAVC=false, LASTC=false, MoI=false, AVC Pretraining=false2022.09 | 17.4 | 7.7 | — | |
| Our multimodal modelEval. Setting=Regular, Train Modalities=Video, Audio, IMU, Test Modalities=Video, Audio, IMU2023.05 | — | — | 55.66 | |
| Our multimodal modelEval. Setting=Missing Modality, Train Modalities=Video, Audio, IMU, Test Modalities=Video, Audio2023.05 | — | — | 55.47 | |
| Our multimodal modelEval. Setting=Missing Modality, Train Modalities=Video, Audio, IMU, Test Modalities=Audio, IMU2023.05 | — | — | 37.07 | |
| Our multimodal modelEval. Setting=Missing Modality, Train Modalities=Video, Audio, IMU, Test Modalities=Video, IMU2023.05 | — | — | 54.57 | |
| Our multimodal modelEval. Setting=Zero-Shot Evaluation, Train Modalities=Video, Test Modalities=Audio2023.05 | — | — | 30.98 | |
| Our multimodal modelEval. Setting=Zero-Shot Evaluation, Train Modalities=Video, Test Modalities=IMU2023.05 | — | — | 20 | |
| Our multimodal modelEval. Setting=Zero-Shot Evaluation, Train Modalities=Video, Test Modalities=Audio, IMU2023.05 | — | — | 25.03 | |
| Our multimodal modelEval. Setting=Zero-Shot Evaluation, Train Modalities=Audio, IMU, Test Modalities=Video2023.05 | — | — | 43.43 | |
| Our multimodal modelEval. Setting=Zero-Shot Evaluation, Train Modalities=Audio, Test Modalities=IMU2023.05 | — | — | 35.67 | |
| Our multimodal modelEval. Setting=Zero-Shot Evaluation, Train Modalities=IMU, Test Modalities=Audio2023.05 | — | — | 41.02 |