WIDER-FAIR adds ethnicity and sex annotations to 16,256 images from WIDER-FACE and demonstrates its utility by showing lower detection performance for Black faces in YOLOv5 experiments.
Predictive Inequity in Object Detection
5 Pith papers cite this work, alongside 157 external citations. Polarity classification is still indexing.
5
Pith papers citing it
157
external citations · Pith
abstract
In this work, we investigate whether state-of-the-art object detection systems have equitable predictive performance on pedestrians with different skin tones. This work is motivated by many recent examples of ML and vision systems displaying higher error rates for certain demographic groups than others. We annotate an existing large scale dataset which contains pedestrians, BDD100K, with Fitzpatrick skin tones in ranges [1-3] or [4-6]. We then provide an in-depth comparative analysis of performance between these two skin tone groupings, finding that neither time of day nor occlusion explain this behavior, suggesting this disparity is not merely the result of pedestrians in the 4-6 range appearing in more difficult scenes for detection. We investigate to what extent time of day, occlusion, and reweighting the supervised loss during training affect this predictive bias.
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
Face-Feature Tuning is a label-free logit remapping method that reduces FPR/TPR gaps across groups in deepfake detection while preserving overall accuracy.
A plug-and-play Anonymizing Adapter Module removes private information from video latent features using self-supervised privacy objectives and consistency losses while retaining utility on action recognition, temporal detection, and anomaly tasks.
An empirical audit of one web-scraped ML training dataset reveals persistent PII after sanitization, which the authors combine with legal analysis to highlight privacy risks and advocate redefining 'publicly available' data for AI training.
All 24 top methods from the Caltech benchmark show significantly higher miss rates on children than adults in the age-and-gender-labeled INRIA Person Dataset.
citing papers explorer
-
WIDER-FAIR: An Annotated Version of the WIDER-FACE Dataset for Fairness Evaluation
WIDER-FAIR adds ethnicity and sex annotations to 16,256 images from WIDER-FACE and demonstrates its utility by showing lower detection performance for Black faces in YOLOv5 experiments.
-
Toward Calibrated, Fair, and accurate Deepfake Detection
Face-Feature Tuning is a label-free logit remapping method that reduces FPR/TPR gaps across groups in deepfake detection while preserving overall accuracy.
-
Privacy Beyond Pixels: Latent Anonymization for Privacy-Preserving Video Understanding
A plug-and-play Anonymizing Adapter Module removes private information from video latent features using self-supervised privacy objectives and consistency losses while retaining utility on action recognition, temporal detection, and anomaly tasks.
-
A Common Pool of Privacy Problems: Legal and Technical Lessons from a Large-Scale Web-Scraped Machine Learning Dataset
An empirical audit of one web-scraped ML training dataset reveals persistent PII after sanitization, which the authors combine with legal analysis to highlight privacy risks and advocate redefining 'publicly available' data for AI training.
-
Age and gender bias in pedestrian detection algorithms
All 24 top methods from the Caltech benchmark show significantly higher miss rates on children than adults in the age-and-gender-labeled INRIA Person Dataset.