Supervised training of deep networks has led to remarkable successes in computer vision, for example on image classification or object detection problems. These successes are driven by the availability of large quantities of paired training data with manual ground truth annotations. For many photography or inverse graphics applications however, manual annotation of ground truth labels is not viable. Motivated by this, I propose several portable hardware prototypes that enable the collection of training data for applications ranging from non-line-of-sight imaging to relighting and dark-flash photography. In the second part of the talk, I discuss my work on differentiable rendering, which points to a less data-dependent approach to inverse graphics problems.
Thesis Supervisor: Prof. Fredo Durand
To attend this defense, please contact the doctoral candidate at lmurmann at mit dot edu