Training Against Disguises: Addressing and Mitigating Bias in Facial Emotion Recognition with Synthetic Data

• Published in: IEEE International Conference and Workshops on Automatic Face and Gesture Recognition : FG pp. 1 - 6
• Main Authors: Sukumar, Aadith, Desai, Aditya, Singhal, Peeyush, Gokhale, Sai, Jain, Deepak Kumar, Walambe, Rahee, Kotecha, Ketan
• Format: Conference Proceeding
• Language: English
• Published: IEEE 27.05.2024
• Subjects: Access control; Benchmark testing; disguise; Emotion recognition; Face recognition; Facial Expression Recognition; good-image bias; Mitigating Bias; Prevention and mitigation; Rendering (computer graphics); security; Synthetic data; Training

Facial Emotion Recognition (FER) is a challenging problem due to various challenges such as variability in expressions and ambiguity in data. Several popular benchmarking datasets, specifically employed for FER tasks exhibit bias towards ethnicity, demography and image capture mechanisms. More speci...