Face Recognition: Too Bias, or Not Too Bias?
This addresses bias issues in facial recognition, which is critical for fairness in applications like security and surveillance, though it is incremental as it builds on existing threshold optimization methods.
The paper tackles bias in facial recognition systems by introducing a Balanced Faces In the Wild (BFW) dataset balanced for gender and ethnicity, showing that using subgroup-specific thresholds instead of a global one reduces performance gaps and boosts overall accuracy.
We reveal critical insights into problems of bias in state-of-the-art facial recognition (FR) systems using a novel Balanced Faces In the Wild (BFW) dataset: data balanced for gender and ethnic groups. We show variations in the optimal scoring threshold for face-pairs across different subgroups. Thus, the conventional approach of learning a global threshold for all pairs resulting in performance gaps among subgroups. By learning subgroup-specific thresholds, we not only mitigate problems in performance gaps but also show a notable boost in the overall performance. Furthermore, we do a human evaluation to measure the bias in humans, which supports the hypothesis that such a bias exists in human perception. For the BFW database, source code, and more, visit github.com/visionjo/facerec-bias-bfw.