Related to the Smart Asset re-Use in Creative Environments (SAUCE) project
Published in Proceedings of the European Conference on Visual Media Production 2019, 2019
In recent years, there has been an increase in the popularity of Light Field (LF) imaging technology with the increase in availability of LF camera devices such as the Lytro, as well as an increase in the use of LF camera arrays . However, both camera arrays and LF cameras can create views that exhibit colour discrepancies, and previous work has already tackled similar issues when data is captured using the Lytro camera  or multiview camera systems . In , an L2 based colour transfer method is applied in an iterative approach to recolour the views of a Lytro light field to correct the colour fading that occurs on outer views of the LF. This method was based on earlier L2 based methods proposed by Grogan et al.[1, 2]. In this paper, we propose to combine similar aspects of the L2 based cost function proposed in  with some of the cost function constraints proposed in , and propose a new propagation scheme so that this L2 based framework can be extended to colour correcting LF arrays. We also take advantage of a colour chart captured in the scene to not only ensure that colours are consistent across the LF, but also match the ground truth colour chart.