Optical computing systems provide high-speed and low-energy data processing but face deficiencies in computationally demanding training and simulation-to-reality gaps. We propose a gradient-based model-free optimization (G-MFO) method based on a Monte Carlo gradient estimation algorithm for computationally efficient in situ training of optical computing systems. This approach treats an optical computing system as a black box and back-propagates the loss directly to the optical computing weights' probability distributions, circumventing the need for a computationally heavy and biased system simulation. Our experiments on diffractive optical computing systems show that G-MFO outperforms hybrid training on the MNIST and FMNIST datasets. Furthermore, we demonstrate image-free and high-speed classification of cells from their marker-free phase maps. Our method's model-free and high-performance nature, combined with its low demand for computational resources, paves the way for accelerating the transition of optical computing from laboratory demonstrations to practical, real-world applications.

We introduce neural lithography to address the ‘design-to-manufacturing’ gap in computational optics. Computational optics with large design degrees of freedom enable advanced functionalities and performance beyond traditional optics. However, the existing design approaches often overlook the numerical modeling of the manufacturing process, which can result in significant performance deviation between the design and the fabricated optics. To bridge this gap, we, for the first time, propose a fully differentiable design framework that integrates a pre-trained photolithography simulator into the model-based optical design loop. Leveraging a blend of physics-informed modeling and data-driven training using experimentally collected datasets, our photolithography simulator serves as a regularizer on fabrication feasibility during design, compensating for structure discrepancies introduced in the lithography process. We demonstrate the effectiveness of our approach through two typical tasks in computational optics, where we design and fabricate a holographic optical element (HOE) and a multi-level diffractive lens (MDL) using a two-photon lithography system, showcasing improved optical performance on the task-specific metrics. The source code for this work is available on the project page: https://neural-litho.github.io.

Cracks are one of the earliest indications of structure degradation relevant to buildings, roads, or widgets. How-ever, false alarms and misjudgments are easy to occur when segmenting cracks from large scenes. Thus, robustly detecting and precisely segmenting non-salient cracks are always difficult, especially for large oblique drone images. To handle the problems above, we propose a novel scene-aware semantic crack segmentation model, M-CrackNet. The contributions of the proposed algorithm can be mainly concluded as: (1) it decomposes the task of building crack detection into crack related scene activation and semantic crack segmentation so that the method can be transferred to other scenes with cracks; (2) by parallelly processing images with two decomposed modules, the computational cost has been largely reduced. Meanwhile, most of the false alarms and misjudgments are avoided, enhancing the accuracy of semantic crack segmentation. The effectiveness of the proposed method is evaluated upon CPRI, CrackForest, and the CrackODI proposed in this paper. Compared with the state-of-the-art methods, the mIoU and BFScore of the M-CrackNet are improved by 0.64%∼8.01% and 3.12%∼33.29% on datasets, which demonstrate that the M-CrackNet is effective for semantic crack segmentation.