# Emotion Recognition with Spatial Attention and Temporal Softmax Pooling

arXiv: https://arxiv.org/abs/1910.01254 · HTML: https://mlsanity.com/paper/1910.01254/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 2 Oct 2019.

## Built on by (most cited first, 6 of 6)

- [Facial expression and attributes recognition based on multi-task learning of lightweight neural networks](https://mlsanity.com/paper/2103.17107/) (2021)
- [Expression Snippet Transformer for Robust Video-based Facial Expression Recognition](https://mlsanity.com/paper/2109.08409/) (2021)
- [NR-DFERNet: Noise-Robust Network for Dynamic Facial Expression Recognition](https://mlsanity.com/paper/2206.04975/) (2022)
- [MIDAS: Mixing Ambiguous Data with Soft Labels for Dynamic Facial Expression Recognition](https://mlsanity.com/paper/2503.00202/) (2025)
- [Emotion Recognition from the perspective of Activity Recognition](https://mlsanity.com/paper/2403.16263/) (2024)
- [Enhancing Ambiguous Dynamic Facial Expression Recognition with Soft Label-based Data Augmentation](https://mlsanity.com/paper/2506.20867/) (2025)

## Abstract (arXiv)

Video-based emotion recognition is a challenging task because it requires to distinguish the small deformations of the human face that represent emotions, while being invariant to stronger visual differences due to different identities. State-of-the-art methods normally use complex deep learning models such as recurrent neural networks (RNNs, LSTMs, GRUs), convolutional neural networks (CNNs, C3D, residual networks) and their combination. In this paper, we propose a simpler approach that combines a CNN pre-trained on a public dataset of facial images with (1) a spatial attention mechanism, to localize the most important regions of the face for a given emotion, and (2) temporal softmax pooling, to select the most important frames of the given video. Results on the challenging EmotiW dataset show that this approach can achieve higher accuracy than more complex approaches.