# A Deeper Look at Facial Expression Dataset Bias

arXiv: https://arxiv.org/abs/1904.11150 · HTML: https://mlsanity.com/paper/1904.11150/

## Key facts

- Cited by 25 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 25 Apr 2019.

## Built on by (most cited first, 25 of 25)

- [Domain-Incremental Continual Learning for Mitigating Bias in Facial Expression and Action Unit Recognition](https://mlsanity.com/paper/2103.08637/) (2021)
- [Boosting Facial Expression Recognition by A Semi-Supervised Progressive Teacher](https://mlsanity.com/paper/2205.14361/) (2022)
- [Disentangled Source-Free Personalization for Facial Expression Recognition with Neutral Target Data](https://mlsanity.com/paper/2503.20771/) (2025)
- [Personalized Feature Translation for Expression Recognition: An Efficient Source-Free Domain Adaptation Method](https://mlsanity.com/paper/2508.09202/) (2025)
- [A Comprehensive Survey on Affective Computing; Challenges, Trends, Applications, and Future Directions](https://mlsanity.com/paper/2305.07665/) (2023)
- [A call for embodied AI](https://mlsanity.com/paper/2402.03824/) (2024)
- [Faces of Fairness: Examining Bias in Facial Expression Recognition Datasets and Models](https://mlsanity.com/paper/2502.11049/) (2025)
- [Cluster-level pseudo-labelling for source-free cross-domain facial expression recognition](https://mlsanity.com/paper/2210.05246/) (2022)
- [Critically examining the Domain Generalizability of Facial Expression Recognition models](https://mlsanity.com/paper/2106.15453/) (2021)
- [AU-Aware Vision Transformers for Biased Facial Expression Recognition](https://mlsanity.com/paper/2211.06609/) (2022)
- [Adaptive Global-Local Representation Learning and Selection for Cross-Domain Facial Expression Recognition](https://mlsanity.com/paper/2401.11085/) (2024)
- [Towards Fair Affective Robotics: Continual Learning for Mitigating Bias in Facial Expression and Action Unit Recognition](https://mlsanity.com/paper/2103.09233/) (2021)
- [Causal-Ex: Causal Graph-based Micro and Macro Expression Spotting](https://mlsanity.com/paper/2503.09098/) (2025)
- [Automated Deepfake Detection](https://mlsanity.com/paper/2106.10705/) (2021)
- [Going Deeper than Tracking: a Survey of Computer-Vision Based Recognition of Animal Pain and Affective States](https://mlsanity.com/paper/2206.08405/) (2022)
- [Efficient Neural Architecture Search for Emotion Recognition](https://mlsanity.com/paper/2303.13653/) (2023)
- [Fair Uncertainty Quantification for Depression Prediction](https://mlsanity.com/paper/2505.04931/) (2025)
- [Graph-Attention Network with Adversarial Domain Alignment for Robust Cross-Domain Facial Expression Recognition](https://mlsanity.com/paper/2512.00641/) (2025)
- [Hypergraph-Guided Disentangled Spectrum Transformer Networks for Near-Infrared Facial Expression Recognition](https://mlsanity.com/paper/2312.05907/) (2023)
- [FastPacket: Towards Pre-trained Packets Embedding based on FastText for next-generation NIDS](https://mlsanity.com/paper/2209.14727/) (2022)
- [An experimental study in Real-time Facial Emotion Recognition on new 3RL dataset](https://mlsanity.com/paper/2304.03064/) (2023)
- [Vulnerability Detection Using Two-Stage Deep Learning Models](https://mlsanity.com/paper/2305.09673/) (2023)
- [Multi-Energy Guided Image Translation with Stochastic Differential Equations for Near-Infrared Facial Expression Recognition](https://mlsanity.com/paper/2312.05908/) (2023)
- [MotivNet: Evolving Meta-Sapiens into an Emotionally Intelligent Foundation Model](https://mlsanity.com/paper/2512.24231/) (2025)
- [Evaluating Test-Time Adaptation For Facial Expression Recognition Under Natural Cross-Dataset Distribution Shifts](https://mlsanity.com/paper/2603.19994/) (2026)

## Abstract (arXiv)

Datasets play an important role in the progress of facial expression recognition algorithms, but they may suffer from obvious biases caused by different cultures and collection conditions. To look deeper into this bias, we first conduct comprehensive experiments on dataset recognition and crossdataset generalization tasks, and for the first time explore the intrinsic causes of the dataset discrepancy. The results quantitatively verify that current datasets have a strong buildin bias and corresponding analyses indicate that the conditional probability distributions between source and target datasets are different. However, previous researches are mainly based on shallow features with limited discriminative ability under the assumption that the conditional distribution remains unchanged across domains. To address these issues, we further propose a novel deep Emotion-Conditional Adaption Network (ECAN) to learn domain-invariant and discriminative feature representations, which can match both the marginal and the conditional distributions across domains simultaneously. In addition, the largely ignored expression class distribution bias is also addressed by a learnable re-weighting parameter, so that the training and testing domains can share similar class distribution. Extensive cross-database experiments on both lab-controlled datasets (CK+, JAFFE, MMI and Oulu-CASIA) and real-world databases (AffectNet, FER2013, RAF-DB 2.0 and SFEW 2.0) demonstrate that our ECAN can yield competitive performances across various facial expression transfer tasks and outperform the state-of-theart methods.