Controllable Data Augmentation Through Deep Relighting

TL;DR

This paper introduces a deep relighting data augmentation method that enhances model invariance to illumination changes, improving localization performance by generating varied lighting conditions using an encoder-decoder network.

Contribution

The authors develop a controllable relighting tool for data augmentation that allows parameterized illumination variations, improving model robustness to lighting changes.

Findings

Enhanced localization accuracy on benchmarks.

Effective augmentation of datasets with diverse lighting conditions.

Improved model invariance to illumination variations.

Abstract

At the heart of the success of deep learning is the quality of the data. Through data augmentation, one can train models with better generalization capabilities and thus achieve greater results in their field of interest. In this work, we explore how to augment a varied set of image datasets through relighting so as to improve the ability of existing models to be invariant to illumination changes, namely for learned descriptors. We develop a tool, based on an encoder-decoder network, that is able to quickly generate multiple variations of the illumination of various input scenes whilst also allowing the user to define parameters such as the angle of incidence and intensity. We demonstrate that by training models on datasets that have been augmented with our pipeline, it is possible to achieve higher performance on localization benchmarks.