A Step Toward More Inclusive People Annotations for Fairness

TL;DR

This paper introduces the MIAP subset with comprehensive annotations for all people in selected images from the Open Images Dataset, aiming to facilitate fairness research in computer vision.

Contribution

It presents a new exhaustive annotation methodology for the person class in Open Images, enabling fairness analysis and understanding annotation biases.

Findings

MIAP enables fairness research in computer vision.

Analysis of annotation patterns reveals biases affecting models.

Exhaustive annotations help study systematic effects in training data.

Abstract

The Open Images Dataset contains approximately 9 million images and is a widely accepted dataset for computer vision research. As is common practice for large datasets, the annotations are not exhaustive, with bounding boxes and attribute labels for only a subset of the classes in each image. In this paper, we present a new set of annotations on a subset of the Open Images dataset called the MIAP (More Inclusive Annotations for People) subset, containing bounding boxes and attributes for all of the people visible in those images. The attributes and labeling methodology for the MIAP subset were designed to enable research into model fairness. In addition, we analyze the original annotation methodology for the person class and its subclasses, discussing the resulting patterns in order to inform future annotation efforts. By considering both the original and exhaustive annotation sets, researchers can also now study how systematic patterns in training annotations affect modeling.