Tolerance-Aware Deep Optics

Deep optics has emerged as a promising approach by co-designing optical elements with deep learning algorithms. However, current research typically overlooks the analysis and optimization of manufacturing and assembly tolerances. This oversight creates a significant performance gap between designed and fabricated optical systems. To address this challenge, we present the first end-to-end tolerance-aware optimization framework that incorporates multiple tolerance types into the deep optics design pipeline. Our method combines physics-informed modelling with data-driven training to enhance optical design by accounting for and compensating for structural deviations in manufacturing and assembly. We validate our approach through computational imaging applications, demonstrating results in both simulations and real-world experiments. We further examine how our proposed solution improves the robustness of optical systems and vision algorithms against tolerances through qualitative and quantitative analyses.

• We first implement a tolerance sampler that can convert random tolerances into variations in optical parameters and spatial transformations. The tolerance sampler is capable of applying realistic tolerance perturbations to lens systems, thereby reducing the gap between design and manufacture.
• We design two loss functions aimed at improving the optical tolerance robustness: PSF Loss and Spot Loss. The combined use of these two losses significantly enhances the stability of end-to-end deep optical training.