Controllable Data Augmentation Through Deep Relighting
This work addresses the need for better generalization in deep learning models, particularly for learned descriptors in computer vision, but it is incremental as it builds on existing data augmentation methods.
The paper tackles the problem of improving model invariance to illumination changes by developing a controllable data augmentation tool for relighting images, resulting in higher performance on localization benchmarks.
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.