Skip to content

Opening book details…

About this document

A Feature-Level Bias Evaluation Framework For Facial Expression Recognition Models by hahorec703 is a document available to read on EtoBox.

The paper presents a feature-level bias evaluation framework for Facial Expression Recognition (FER) models that addresses the limitations of existing methods relying on pseudo-demographic labels. It introduces a statistical module to ensure the significance of bias evaluation results and conducts a comprehensive analysis across various demographic attributes and facial expressions. The findings reveal significant demographic biases in FER, highlighting the need for fairer network architectures and robust e

Author
hahorec703
Language
EN