Monte Carlo Dropout Ensembles for Robust Illumination Estimation

TL;DR

This paper introduces a robust illumination estimation method that combines multiple deep learning models weighted by their uncertainty estimates, significantly improving accuracy especially on challenging samples.

Contribution

It proposes a novel ensemble approach using Monte Carlo dropout to estimate uncertainty and aggregate multiple models for better illumination estimation.

Findings

Achieves state-of-the-art performance on INTEL-TAU dataset.

Effectively handles extreme samples with high errors.

Demonstrates improved robustness over individual models.

Abstract

Computational color constancy is a preprocessing step used in many camera systems. The main aim is to discount the effect of the illumination on the colors in the scene and restore the original colors of the objects. Recently, several deep learning-based approaches have been proposed to solve this problem and they often led to state-of-the-art performance in terms of average errors. However, for extreme samples, these methods fail and lead to high errors. In this paper, we address this limitation by proposing to aggregate different deep learning methods according to their output uncertainty. We estimate the relative uncertainty of each approach using Monte Carlo dropout and the final illumination estimate is obtained as the sum of the different model estimates weighted by the log-inverse of their corresponding uncertainties. The proposed framework leads to state-of-the-art performance on INTEL-TAU dataset.