I’m a Ph.D. candidate in Computer Science at Northwestern University specializing in AI-driven algorithms for Computational Photography, Display and Imaging. I’m mentored by my main advisor Prof. Oliver Cossairt and co-advised by Prof. Aggelos Katsaggelos. My PhD thesis focuses on optimization-based methods in next-generation 3D displays, and I also hold a keen interest in exploring topics in Computer Vision, Inverse Problems, and Medical Imaging.
Before delving into my Ph.D., I earned my Masters in Physics in Erlangen and worked at Siemens Healthineers. During my PhD, I had the opportunity to work with Reality Labs, Meta, where I extensively collaborated with Doug Lanman, Grace Kuo, Felix Heide and Nathan Matsuda to contribute to the evolution of holographic near-eye displays. Additionally, I have had the enriching experience of being a full instructor for several courses at Northwestern.
- Computational Photography/Imaging/Display
- Generative AI and Machine Learning
- Computer Vision
PhD in Computer Science, 2024
Northwestern University
MSc in Physics, 2018
FAU Erlangen, Germany
MSc in Advanced Optical Technologies, 2017
FAU Erlangen, Germany
Selected Publications
MultiSource Holography Grace Kuo, Florian Schiffers, Douglas Lanman, Oliver Cossairt, and Nathan Matsuda SIGGRAPH ASIA 2023 arXiv | Supplement We propose an architecture for speckle reduction in holographic displays that uses an array of mutually incoherent sources and two sequential spatial light modulators. Multisource holography can suppress speckle in a single frame without sacrificing resolution. | |
Stochastic Lightfield Holography Florian Schiffers, Praneeth Chakravarthula, N. Matsuda, G. Kuo, E. Tseng, D. Lanman, F. Heide, Oliver Cossairt ICCP 2023 arXiv | Supplement Our novel hologram generation algorithm enhances the realism of near-eye displays by matching the projection operators of incoherent (Light Field) and coherent (Wigner Function) light transport. By supervising hologram computation with on-the-fly synthesized photographs using Light Field refocusing, our method significantly improves image quality and viewing experience across diverse pupil states. | |
2-Step Sparse-View CT Reconstruction with a Domain-Specific Perceptual Network Florian Schiffers*, Haoyu Wei*, Tobias Wuerfl, Daming Shen, Daniel Kim, Aggelos Katsaggelos, Oliver Cossairt arXiv | Supplement Our novel framework for sparse-view computed tomography overcomes the challenges of angular undersampling by using a super-resolution network and a refinement network, resulting in high-quality reconstructions with significantly reduced streak artifacts. Our method enhances domain-specific information and demonstrates a 4 dB improvement over current solutions. | |
Selfvi Self-Supervised Light-Field Video Reconstruction from Stereo Video Prasan Shedligeri, Florian Schiffers, Sushobhan Ghosh, Oliver Cossairt, Kaushik Mitra ICCV 2021 paper | Supplement Light- field imaging is attractive for mobile devices due to its intuitive post-capture processing, though acquiring high-quality LF data is challenging with space constraints. We propose a self-supervised algorithm for reconstructing high-fidelity LF videos from stereo videos, leveraging geometric and temporal information and enables applications like post-capture focus control. |