Carlos Caetano, Jessica Sena, François Brémond, Jefersson A. dos Santos, William Robson Schwartz
Due to the availability of large-scale skeleton datasets, 3D human action recognition has recently called the attention of computer vision community. Many works have focused on encoding skeleton data as skeleton image representations based on spatial structure of the skeleton joints, in which the temporal dynamics of the sequence is encoded as variations in columns and the spatial structure of each frame is represented as rows of a matrix. To further improve such representations, we introduce a novel skeleton image representation to be used as input of Convolutional Neural Networks (CNNs), named SkeleMotion. The proposed approach encodes the temporal dynamics by explicitly computing the magnitude and orientation values of the skeleton joints. Different temporal scales are employed to compute motion values to aggregate more temporal dynamics to the representation making it able to capture longrange joint interactions involved in actions as well as filtering noisy motion values. Experimental results demonstrate the effectiveness of the proposed representation on 3D action recognition outperforming the state-of-the-art on NTU RGB+D 120 dataset.
| Task | Dataset | Metric | Value | Model |
|---|---|---|---|---|
| Video | NTU RGB+D | Accuracy (CS) | 76.5 | Skelemotion + Yang et al. |
| Video | NTU RGB+D | Accuracy (CV) | 84.7 | Skelemotion + Yang et al. |
| Temporal Action Localization | NTU RGB+D | Accuracy (CS) | 76.5 | Skelemotion + Yang et al. |
| Temporal Action Localization | NTU RGB+D | Accuracy (CV) | 84.7 | Skelemotion + Yang et al. |
| Zero-Shot Learning | NTU RGB+D | Accuracy (CS) | 76.5 | Skelemotion + Yang et al. |
| Zero-Shot Learning | NTU RGB+D | Accuracy (CV) | 84.7 | Skelemotion + Yang et al. |
| Activity Recognition | NTU RGB+D | Accuracy (CS) | 76.5 | Skelemotion + Yang et al. (Skeleton only) |
| Activity Recognition | NTU RGB+D | Accuracy (CV) | 84.7 | Skelemotion + Yang et al. (Skeleton only) |
| Activity Recognition | NTU RGB+D 120 | Accuracy (Cross-Setup) | 66.9 | Skelemotion + Yang et al. (skeleton only) |
| Activity Recognition | NTU RGB+D 120 | Accuracy (Cross-Subject) | 67.7 | Skelemotion + Yang et al. (skeleton only) |
| Activity Recognition | NTU RGB+D | Accuracy (CS) | 76.5 | Skelemotion + Yang et al. |
| Activity Recognition | NTU RGB+D | Accuracy (CV) | 84.7 | Skelemotion + Yang et al. |
| Action Localization | NTU RGB+D | Accuracy (CS) | 76.5 | Skelemotion + Yang et al. |
| Action Localization | NTU RGB+D | Accuracy (CV) | 84.7 | Skelemotion + Yang et al. |
| Action Detection | NTU RGB+D | Accuracy (CS) | 76.5 | Skelemotion + Yang et al. |
| Action Detection | NTU RGB+D | Accuracy (CV) | 84.7 | Skelemotion + Yang et al. |
| 3D Action Recognition | NTU RGB+D | Accuracy (CS) | 76.5 | Skelemotion + Yang et al. |
| 3D Action Recognition | NTU RGB+D | Accuracy (CV) | 84.7 | Skelemotion + Yang et al. |
| Action Recognition | NTU RGB+D | Accuracy (CS) | 76.5 | Skelemotion + Yang et al. (Skeleton only) |
| Action Recognition | NTU RGB+D | Accuracy (CV) | 84.7 | Skelemotion + Yang et al. (Skeleton only) |
| Action Recognition | NTU RGB+D 120 | Accuracy (Cross-Setup) | 66.9 | Skelemotion + Yang et al. (skeleton only) |
| Action Recognition | NTU RGB+D 120 | Accuracy (Cross-Subject) | 67.7 | Skelemotion + Yang et al. (skeleton only) |
| Action Recognition | NTU RGB+D | Accuracy (CS) | 76.5 | Skelemotion + Yang et al. |
| Action Recognition | NTU RGB+D | Accuracy (CV) | 84.7 | Skelemotion + Yang et al. |