Benjamin Attal

I am a Postdoctoral Fellow at the University of Toronto Computational with Prof. David Lindell. I am generously supported by a Schmidt AI In Science Fellowship. Previously, I was a PhD student at Carnegie Mellon University, where I was advised by Prof. Matthew O'Toole and supported by a Meta PhD Fellowship. I received my Bachelor's degree in Computer Science and Applied Math from Brown University. I received my Master's degree in Applied Math from Brown University.

Email  /  CV  /  Scholar

Research

My research is at the intersection of computational imaging, computer graphics, and machine learning. I am interested in using physics-based light transport simulation and data-driven methods to design robust systems for inverse rendering and 3D reconstruction.

Time-resolved relighting and geometry estimation through radiance caching.

A more physically-accurate inverse rendering system based on radiance caching for recovering geometry, materials, and lighting from RGB images of an object or scene.

C-ToF depth cameras can't reconstruct dynamic objects well. We fix that with our NeRF model that takes raw ToF signal and reconstructs motion along with the depth.

We apply a neural volume rendering framework to the raw images from structured-light sensors in order to achieve high-quality 3D reconstruction.

A 6-DoF video pipeline based on neural radiance fields that achieves a good trade-off between speed, quality, and memory efficiency. It excels at representing challenging view-dependent effects such as reflections and refractions.

A fast and compact neural field representation for light fields.

A practical coded diffraction imaging framework that can decouple mutually incoherent mixed-states, such as different wavelengths. Applications in computational microscopy.

We apply a phasor volume rendering model to the raw images from C-ToF sensors in order to achieve high-quality 3D torfstruction of static and dynamic scenes.

We build a real-time inference and rendering framework for 6-DoF video based on multi-sphere images.