Tilted Cross Entropy (TCE): Promoting Fairness in Semantic Segmentation. (arXiv:2103.14051v1 [cs.CV])

Traditional empirical risk minimization (ERM) for semantic segmentation can
disproportionately advantage or disadvantage certain target classes in favor of
an (unfair but) improved overall performance. Inspired by the recently
introduced tilted ERM (TERM), we propose tilted cross-entropy (TCE) loss and
adapt it to the semantic segmentation setting to minimize performance disparity
among target classes and promote fairness. Through quantitative and qualitative
performance analyses, we demonstrate that the proposed Stochastic TCE for
semantic segmentation can efficiently improve the low-performing classes of
Cityscapes and ADE20k datasets trained with multi-class cross-entropy (MCCE),
and also results in improved overall fairness.

Source: https://arxiv.org/abs/2103.14051


Related post