# Magnifying Subtle Facial Motions for Effective 4D Expression Recognition

arXiv: https://arxiv.org/abs/2105.02319 · HTML: https://mlsanity.com/paper/2105.02319/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 0 of its 45 references that are in the map.
- First posted on arXiv 5 May 2021.

## Built on by (most cited first, 5 of 5)

- [A Comprehensive Survey on Affective Computing; Challenges, Trends, Applications, and Future Directions](https://mlsanity.com/paper/2305.07665/) (2023)
- [Contrastive Language-Image Learning with Augmented Textual Prompts for 3D/4D FER Using Vision-Language Model](https://mlsanity.com/paper/2504.19739/) (2025)
- [Unsupervised Multiview Contrastive Language-Image Joint Learning with Pseudo-Labeled Prompts Via Vision-Language Model for 3D/4D Facial Expression Recognition](https://mlsanity.com/paper/2505.09336/) (2025)
- [Self-Supervised Multi-View Representation Learning using Vision-Language Model for 3D/4D Facial Expression Recognition](https://mlsanity.com/paper/2506.01203/) (2025)
- [Facial Emotion Learning with Text-Guided Multiview Fusion via Vision-Language Model for 3D/4D Facial Expression Recognition](https://mlsanity.com/paper/2507.01673/) (2025)

## Abstract (arXiv)

In this paper, an effective pipeline to automatic 4D Facial Expression Recognition (4D FER) is proposed. It combines two growing but disparate ideas in Computer Vision -- computing the spatial facial deformations using tools from Riemannian geometry and magnifying them using temporal filtering. The flow of 3D faces is first analyzed to capture the spatial deformations based on the recently-developed Riemannian approach, where registration and comparison of neighboring 3D faces are led jointly. Then, the obtained temporal evolution of these deformations are fed into a magnification method in order to amplify the facial activities over the time. The latter, main contribution of this paper, allows revealing subtle (hidden) deformations which enhance the emotion classification performance. We evaluated our approach on BU-4DFE dataset, the state-of-art 94.18% average performance and an improvement that exceeds 10% in classification accuracy, after magnifying extracted geometric features (deformations), are achieved.