Bias Mitigation Strategies for Facial Attribute Classification Leveraging Fine-grained Features and ChatGPT
Ayesha Manzoor · 2025
Published research highlights the presence of demographic bias in automated facial attribute classification algorithms, particularly impacting women and individuals with darker skin tones. Existing bias mitigation techniques typically require demographic annotations and often obtain a trade-off between fairness and accuracy, i.e., Pareto inefficiency. As a part of this thesis, we proposed novel approaches to fair facial attribute classification by (a) framing it as a fine-grained classification problem. Our approach effectively integrates both low-level local features (like edges and color) and high-level semantic features (like shapes and structures) through cross-layer mutual attention learning, and (b) proposing a novel training framework that integrates ChatGPT’s linguistic features with vision models to improve classification accuracy, fairness and introduce explainability. An exhaustive evaluation on facial attribute annotated datasets demonstrates that our proposed models improve model accuracy by 1.32% to 1.74% and fairness by 67% to 83.6%, over the SOTA bias mitigation techniques.