No Fear of the Dark: Image Retrieval under Varying Illumination Conditions

TL;DR

This paper introduces a learnable photometric normalisation method using U-Net for image retrieval across different illumination conditions, demonstrating improved robustness over traditional methods and competitive performance on daylight benchmarks.

Contribution

A novel learnable normalisation technique based on U-Net architecture for illumination-invariant image retrieval, trained on multi-exposure and landmark datasets.

Findings

Learnable normalisation outperforms traditional methods under varying illumination.

Both normalisation approaches achieve comparable results to state-of-the-art daylight benchmarks.

Photometric normalisation enhances image descriptor robustness across different lighting conditions.

Abstract

Image retrieval under varying illumination conditions, such as day and night images, is addressed by image preprocessing, both hand-crafted and learned. Prior to extracting image descriptors by a convolutional neural network, images are photometrically normalised in order to reduce the descriptor sensitivity to illumination changes. We propose a learnable normalisation based on the U-Net architecture, which is trained on a combination of single-camera multi-exposure images and a newly constructed collection of similar views of landmarks during day and night. We experimentally show that both hand-crafted normalisation based on local histogram equalisation and the learnable normalisation outperform standard approaches in varying illumination conditions, while staying on par with the state-of-the-art methods on daylight illumination benchmarks, such as Oxford or Paris datasets.