Real-World Computational Aberration Correction via Quantized Domain-Mixing Representation

arXiv:2403.10012v1 Announce Type: new
Abstract: Relying on paired synthetic data, existing learning-based Computational Aberration Correction (CAC) methods are confronted with the intricate and multifaceted synthetic-to-real domain gap, which leads to suboptimal performance in real-world applications. In this paper, in contrast to improving the simulation pipeline, we deliver a novel insight into real-world CAC from the perspective of Unsupervised Domain Adaptation (UDA). By incorporating readily accessible unpaired real-world data into training, we formalize the Domain Adaptive CAC (DACAC) task, and then …

abstract applications arxiv cac computational contrast cs.cv cs.ro data domain eess.iv gap insight leads novel paper performance physics.optics pipeline representation simulation synthetic synthetic data the simulation type via world