The third is a collaboration with KAIST & MSRA involving Seung-Hwan Baek, Tizian Zeltner, Hyun Jin Ku, Inseung Hwang, Xin Tong, and Min Kim. We’ve put together the first comprehensive database of BRDF measurements that captures changes in polarization state of light. 1/3

If you are @ #SIGGRAPH2023 , come to our course tomorrow on Polarization-based Visual computing with Achuta Kadambi Jinwei Ye George Chen & Ramesh Raskar to learn how this invisible light property enhances graphics & vision! hgchen.com/pol_vis_tutori… s2023.siggraph.org/presentation/?…

Does the RGB radiance field encompass all exploitable visual information? In our SIGA23 paper, Neural Spectro-Polarimetric Fields (NeSpoFs), we delve into light's wave properties, polarization and spectrum, that are beyond human vision capability. Project: youngchan-k.github.io/nespof/

We present our CVPR'24 paper, differentiable point-based inverse rendering (DPIR) that achieves accurate reconstruction with small memory footprint and fast reconstruction speed. Our key idea is to (1) use differentiable point-based forward rendering in an analysis-by-synthesis

What are the optimal illumination patterns for 3D reconstruction? In our CVPR'24 paper, we propose learning display illumination patterns for photometric stereo using real-world training datasets created with 3D-printed objects. Our method, titled "Differentiable Display

RGB image datasets have fueled innovation in computer vision, graphics, and AI. However, it is the RGB intensity format that restricts us to the level of human vision. In our CVPR '24 highlight paper, we explore polarization and spectrum, wave properties of light that animals in

Modern 3D imaging methods have made 3D vision applications accessible. However, analyzing detailed material properties in 3D has been challenging due to the high cost and large-form factor of accurate hyperspectral cameras. In our CVPR '24 paper, we use diffraction-grating film

Sensing and illuminating your 360 environment is crucial for robot/VR-AR/HCI/display applications. In our #NaturePhotonics paper, we computationally design metasurface-driven 360 structured light optimized for a downstream task. This interdisciplinary research, combining computer

📢📢📢 Come and submit to our workshop on Physics-inspired 3D Vision and Imaging at CVPR 2025! Speakers 🗣️ include Ioannis Gkioulekas, Laura Waller, Berthy Feng, Seung-Hwan Baek and Gordon Wetzstein! Thanks to coorganizers Benjamin Attal, Cecilia Zhang, Dor Verbin and James Tompkin!

We have significantly enhanced our previously published hyperspectral 3D imaging approach, Dispersed Structured Light (DSL), presented at #CVPR2024, to better accommodate dynamic scenes. Our latest work, Dense Dispersed Structured Light (DDSL), which will be presented at

We're excited to announce the release of PADO (Pytorch Automatic Differentiable Optics), a python library designed to facilitate differentiable wave-optics simulations. PADO is tailored specifically for researchers and developers engaged in computational imaging, computational