{"id":8988,"date":"2024-08-21T15:06:37","date_gmt":"2024-08-21T19:06:37","guid":{"rendered":"https:\/\/portfolios.cs.earlham.edu\/?page_id=8988"},"modified":"2025-05-14T20:25:22","modified_gmt":"2025-05-15T00:25:22","slug":"martin-olate-pinochet","status":"publish","type":"page","link":"https:\/\/portfolios.cs.earlham.edu\/index.php\/students\/2024-2\/cs488\/martin-olate-pinochet\/","title":{"rendered":"Martin Olate"},"content":{"rendered":"

Facial Emotion Recognition (FER) with Bias Mitigation<\/h1>\n
\n

Introduction<\/h2>\n

I am a Computer Science major at Earlham College with a strong interest in Artificial Intelligence and ethical applications of machine learning. My capstone project focuses on developing a Facial Emotion Recognition (FER) system<\/strong> with bias mitigation techniques<\/strong> to ensure fairness across demographic groups.<\/p>\n<\/div>\n

\n

Abstract<\/h2>\n

Facial emotion recognition (FER) models often exhibit uneven performance across demographic groups, reinforcing existing social biases. This capstone quantifies and mitigates those disparities by building a pipeline that spans data cleaning through advanced bias-aware training. Using three benchmark datasets\u2014FER-2013, RAF-DB, and CK+\u2014we first trained a 48 \u00d7 48-pixel custom CNN baseline that reached 59.6. I then fine-tuned an ImageNet-pre-trained ResNet-50 and applied three complementary mitigation strategies: inverse-frequency sample re-weighting, focal-loss augmentation (\u03b3 = 2), and adversarial debiasing with a gradient-reversal race classifier. The best reweighted ResNet-50 improved overall accuracy to 66.5. The results demonstrate that simple weighting and adversarial objectives can substantially improve cross-group fairness without sacrificing overall performance, offering a practical recipe for bias-aware FER in real-time applications.<\/p>\n<\/div>\n

\n

Graphical Abstract<\/h2>\n<\/div>\n\n\n
\"\"<\/figure>\n\n\n
\n

Data Architecture Diagram<\/h2>\n<\/div>\n\n\n
\"\"<\/figure>\n\n\n\n

Technical Report<\/h2>\n\n\n\n
technicalreportFinal<\/a>Download<\/a><\/div>\n\n\n
\n

Demo Video<\/h2>\n