Predictive Inequity in Object Detection

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.

References in corpus (1)

• Equality of Opportunity in Supervised Learning

Cited by in corpus (26)

• On the Opportunities and Risks of Foundation Models
• WILDS: A Benchmark of in-the-Wild Distribution Shifts
• Image Representations Learned With Unsupervised Pre-Training Contain Human-like Biases
• Robot Rights? Let's Talk about Human Welfare Instead
• Zeno: An Interactive Framework for Behavioral Evaluation of Machine Learning
• Robots Enact Malignant Stereotypes
• Symphony: Composing Interactive Interfaces for Machine Learning
• Large image datasets: A pyrrhic win for computer vision?
• Algorithmic Injustices: Towards a Relational Ethics
• Estimating Skin Tone and Effects on Classification Performance in Dermatology Datasets
• Balanced Datasets Are Not Enough: Estimating and Mitigating Gender Bias in Deep Image Representations
• Can Explainable AI Explain Unfairness? A Framework for Evaluating Explainable AI
• FairFaceGAN: Fairness-aware Facial Image-to-Image Translation
• Characterizing Bias in Classifiers using Generative Models
• Skin Deep: Investigating Subjectivity in Skin Tone Annotations for Computer Vision Benchmark Datasets
• Detecting disparities in police deployments using dashcam data
• FairVis: Visual Analytics for Discovering Intersectional Bias in Machine Learning
• Age and gender bias in pedestrian detection algorithms
• What do we teach to engineering students: embedded ethics, morality, and politics
• Making ethical decisions for the immersive web
• Does enforcing fairness mitigate biases caused by subpopulation shift?
• DeepFN: Towards Generalizable Facial Action Unit Recognition with Deep Face Normalization
• Robustness Enhancement of Object Detection in Advanced Driver Assistance Systems (ADAS)
• A Set of Distinct Facial Traits Learned by Machines Is Not Predictive of Appearance Bias in the Wild
• Free cyclic group actions on highly-connected $2n$-manifolds
• Human Imperceptible Attacks and Applications to Improve Fairness