Flowed Time of Flight Radiance Fields

Mikhail Okunev*, Marc Mapeke, Benjamin Attal, Christian Richardt, Matthew O'Toole, James Tompkin ;

"Flowed time of flight radiance fields () is a method to correct for motion artifacts in continuous-wave time of flight imaging (C-ToF). As C-ToF cameras must capture multiple exposures over time to derive depth, any moving object will exhibit depth errors. We formulate an optimization problem to reconstruct the raw frames captured by the camera via an underlying 4D volumetric scene and a physically-based differentiable C-ToF simulator. With weak optical flow supervision, we can infer a 3D volume with scene flow that explains the raw captures, even though any particular time instant does not provide sufficient constraints upon the depth or motion. On synthetic sequences, we find that our approach reduces depth errors on dynamic objects by up to 20× compared to C-ToF, particularly for axial motions and large disparities (≥ 25 pixels) between raw frames. On real-world sequences, we see qualitatively similar gains with artifacts resolved on falling pillows and swinging baseball bats. All experiments were performed by university authors; CR’s role was as an advisor."