Event Date
Event Date
Miniaturized, mask-based integrated fluorescence microscopes offer a promising platform for endoscopic and implantable bio-imaging, enabling large field-of-view and single-shot 3D imaging. However, accurate and efficient 3D reconstruction from such systems remains a challenge due to the complexity of the imaging model and the ill-posed nature of the inverse problem. We present a physics-informed, unrolled neural network architecture that combines the underlying physics of the imaging system, the interpretability of model-based iterative optimization, and the efficiency of data-driven learning. Our approach enables real-time, high-resolution 3D reconstruction across large imaging volumes. We demonstrate its effectiveness on a range of challenging fluorescence samples, achieving high-quality reconstructions that outperform conventional methods in both speed and fidelity.
Presenter
Univ. of California, Davis (United States)
Dr. Yang is an associate professor in the department of electrical and computer engineering at the University of California, Davis (UC Davis). He received his undergraduate degree from Peking University in China and his Ph.D. from the University of California, Berkeley, both in electrical engineering. After postdoctoral training in neuroscience at Columbia University, he started his own laboratory at UC Davis in late 2017. His research group aims to develop advanced optical methods and neurotechnologies to interrogate and modulate brain activity, with the goal of understanding how neural circuits organize and function, and how behaviors emerge from neuronal activity.