Action Recognition on IkeaFA
72Top-1 AccMAE-Align w/ real humans
Evaluation Results
| Method | Links | |
|---|---|---|
| MAE-Align w/ real humansPrivacy Preserving=false, Stage 1: MAE=Kinetics, Stage 2: Alignment=Kinetics, Evaluation Protocol=Fine-tuning, Backbone=ViT-B2023.11 | 72 | |
| Privacy-Preserving MAE-Align (PPMA)Privacy Preserving=true, Stage 1: MAE=NH Kinetics, Stage 2: Alignment=NH Kinetics + Synthetic, Evaluation Protocol=Fine-tuning, Backbone=ViT-B2023.11 | 67.9 | |
| MAE-Align w/ SyntheticPrivacy Preserving=true, Stage 1: MAE=Synthetic, Stage 2: Alignment=Synthetic, Evaluation Protocol=Fine-tuning, Backbone=ViT-B2023.11 | 67.3 | |
| TimeSformer SyntheticPrivacy Preserving=false, Stage 2: Alignment=Synthetic, Evaluation Protocol=Fine-tuning, Backbone=ViT-B, Pre-trained on ImageNet-21K=true2023.11 | 63.6 | |
| TimeSformer KineticsPrivacy Preserving=false, Stage 2: Alignment=Kinetics, Evaluation Protocol=Fine-tuning, Backbone=ViT-B, Pre-trained on ImageNet-21K=true2023.11 | 61.9 | |
| MAE-Align w/ real humansPrivacy Preserving=false, Stage 1: MAE=Kinetics, Stage 2: Alignment=Kinetics, Evaluation Protocol=Linear Probing, Backbone=ViT-B2023.11 | 58.3 | |
| MAE-Align w/ SyntheticPrivacy Preserving=true, Stage 1: MAE=Synthetic, Stage 2: Alignment=Synthetic, Evaluation Protocol=Linear Probing, Backbone=ViT-B2023.11 | 57.7 | |
| Privacy-Preserving MAE-Align (PPMA)Privacy Preserving=true, Stage 1: MAE=NH Kinetics, Stage 2: Alignment=NH Kinetics + Synthetic, Evaluation Protocol=Linear Probing, Backbone=ViT-B2023.11 | 57.7 | |
| TimeSformer KineticsPrivacy Preserving=false, Stage 2: Alignment=Kinetics, Evaluation Protocol=Linear Probing, Backbone=ViT-B, Pre-trained on ImageNet-21K=true2023.11 | 47.7 | |
| TimeSformer SyntheticPrivacy Preserving=false, Stage 2: Alignment=Synthetic, Evaluation Protocol=Linear Probing, Backbone=ViT-B, Pre-trained on ImageNet-21K=true2023.11 | 45.5 | |
| TSN (RN50 backbone)Privacy Preserving=true, Stage 2: Alignment=Synthetic, Evaluation Protocol=Fine-tuning, Backbone=ResNet-502023.11 | 42.7 | |
| I3D (RN50 backbone)Privacy Preserving=true, Stage 2: Alignment=Synthetic, Evaluation Protocol=Fine-tuning, Backbone=ResNet-502023.11 | 42.7 | |
| R(2+1)D (RN50 backbone)Privacy Preserving=true, Stage 2: Alignment=Synthetic, Evaluation Protocol=Fine-tuning, Backbone=ResNet-502023.11 | 41.5 | |
| TSN (RN50 backbone)Privacy Preserving=true, Stage 2: Alignment=Synthetic, Evaluation Protocol=Linear Probing, Backbone=ResNet-502023.11 | 36 | |
| R(2+1)D (RN50 backbone)Privacy Preserving=true, Stage 2: Alignment=Synthetic, Evaluation Protocol=Linear Probing, Backbone=ResNet-502023.11 | 35.7 | |
| I3D (RN50 backbone)Privacy Preserving=true, Stage 2: Alignment=Synthetic, Evaluation Protocol=Linear Probing, Backbone=ResNet-502023.11 | 33.2 | |
| ScratchPrivacy Preserving=false, Evaluation Protocol=Fine-tuning, Backbone=ViT-B2023.11 | 8.4 |