OOWL500: Overcoming Dataset Collection Bias in the Wild

TL;DR

This paper introduces OOWL500, a large, unbiased 'in the lab' image dataset collected via drone, which helps identify and reduce biases in object recognition models and improves their robustness.

Contribution

The paper presents a scalable drone-based data collection method, creating the OOWL500 dataset, and demonstrates its effectiveness in reducing biases and enhancing object recognition.

Findings

OOWL500 contains 120,000 images of 500 objects, making it the largest of its kind.

Augmenting wild datasets with in-lab data reduces biases and improves generalization.

Camera shake and pose diversity are crucial for robust object recognition.

Abstract

The hypothesis that image datasets gathered online "in the wild" can produce biased object recognizers, e.g. preferring professional photography or certain viewing angles, is studied. A new "in the lab" data collection infrastructure is proposed consisting of a drone which captures images as it circles around objects. Crucially, the control provided by this setup and the natural camera shake inherent to flight mitigate many biases. It's inexpensive and easily replicable nature may also potentially lead to a scalable data collection effort by the vision community. The procedure's usefulness is demonstrated by creating a dataset of Objects Obtained With fLight (OOWL). Denoted as OOWL500, it contains 120,000 images of 500 objects and is the largest "in the lab" image dataset available when both number of classes and objects per class are considered. Furthermore, it has enabled several of new insights on object recognition. First, a novel adversarial attack strategy is proposed, where image perturbations are defined in terms of semantic properties such as camera shake and pose. Indeed, experiments have shown that ImageNet has considerable amounts of pose and professional photography bias. Second, it is used to show that the augmentation of in the wild datasets, such as ImageNet, with in the lab data, such as OOWL500, can significantly decrease these biases, leading to object recognizers of improved generalization. Third, the dataset is used to study questions on "best procedures" for dataset collection. It is revealed that data augmentation with synthetic images does not suffice to eliminate in the wild datasets biases, and that camera shake and pose diversity play a more important role in object recognition robustness than previously thought.