The Mission of Our Lab
In visual computing, encompassing computer graphics and computer vision, we have traditionally analyzed rendered/captured images from virtual/real cameras that were developed in isolation from visual-computing downstream tasks. Our lab focuses on differentiable visual computing, a field in which we strive to jointly design and optimize the end-to-end pipeline of visual computing: light, lens, sensors, (neural) representations, and (neural) reconstructions. By bringing these components together, we could achieve something that was previously unattainable. As such, our research lies at the intersection of computer graphics, computer vision, optics, imaging, display, and machine learning.
In regards to differentiable visual computing, we are currently working on the following specific topics
Neural scene representation, reflectance, BRDF, polarization, spectrum, plenoptic function, light field, ...
Vision under extreme condition
Metasurface imaging, 360 imaging, seeing through fog, low-light imaging, ...
Neural scene reconstruction, shape and reflectance imaging, multi-view imaging, ...
Metasurface display, virtual and augmented reality, ...
Recent Presentations on Youtube