Predictive Inequity in Object Detection

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.

[1]  Marcus A. MaloofDepartment A Machine Learning Researcher's Foray into Recidivism Prediction , 1999 .

[2]  Luc Van Gool,et al.  A mobile vision system for robust multi-person tracking , 2008, 2008 IEEE Conference on Computer Vision and Pattern Recognition.

[3]  Luc Van Gool,et al.  The Pascal Visual Object Classes (VOC) Challenge , 2010, International Journal of Computer Vision.

[4]  R. Berk,et al.  Small Area Estimation of the Homeless in Los Angeles: An Application of Cost-Sensitive stochastic Gradient Boosting , 2010, 1011.2890.

[5]  Pietro Perona,et al.  Pedestrian Detection: An Evaluation of the State of the Art , 2012, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[6]  Anil K. Jain,et al.  Face Recognition Performance: Role of Demographic Information , 2012, IEEE Transactions on Information Forensics and Security.

[7]  Toniann Pitassi,et al.  Fairness through awareness , 2011, ITCS '12.

[8]  Andreas Geiger,et al.  Are we ready for autonomous driving? The KITTI vision benchmark suite , 2012, 2012 IEEE Conference on Computer Vision and Pattern Recognition.

[9]  Latanya Sweeney,et al.  Discrimination in online ad delivery , 2013, CACM.

[10]  Pietro Perona,et al.  Microsoft COCO: Common Objects in Context , 2014, ECCV.

[11]  Kaiming He,et al.  Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks , 2015, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[12]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[13]  K. Lum,et al.  To predict and serve? , 2016 .

[14]  Liang Lin,et al.  Is Faster R-CNN Doing Well for Pedestrian Detection? , 2016, ECCV.

[15]  Sebastian Ramos,et al.  The Cityscapes Dataset for Semantic Urban Scene Understanding , 2016, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[16]  Nathan Srebro,et al.  Equality of Opportunity in Supervised Learning , 2016, NIPS.

[17]  Alexandra Chouldechova,et al.  Fair prediction with disparate impact: A study of bias in recidivism prediction instruments , 2016, Big Data.

[18]  Andrew D. Selbst Disparate Impact in Big Data Policing , 2017 .

[19]  Jon M. Kleinberg,et al.  Inherent Trade-Offs in the Fair Determination of Risk Scores , 2016, ITCS.

[20]  Timnit Gebru,et al.  Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification , 2018, FAT.

[21]  Jian Yang,et al.  Occluded Pedestrian Detection Through Guided Attention in CNNs , 2018, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[22]  Alexandra Chouldechova,et al.  A case study of algorithm-assisted decision making in child maltreatment hotline screening decisions , 2018, FAT.

[23]  Michael Carl Tschantz,et al.  Discrimination in Online Advertising: A Multidisciplinary Inquiry , 2018 .

[24]  Trevor Darrell,et al.  BDD100K: A Diverse Driving Video Database with Scalable Annotation Tooling , 2018, ArXiv.

[25]  Shuicheng Yan,et al.  Scale-Aware Fast R-CNN for Pedestrian Detection , 2015, IEEE Transactions on Multimedia.