Towards Unsupervised Learning of Generative Models for 3D Controllable Image Synthesis

Keynote presented on June 14, 2020 at CVPR in the Joint Workshop on Long Term Visual Localization, Visual Odometry and Geometric and Learning-based SLAM Slides: Paper: Abstract: In recent years, Generative Adversarial Networks have achieved impressive results in photorealistic image synthesis. This progress nurtures hopes that one day the classical rendering pipeline can be replaced by efficient models that are learned directly from images. However, current image synthesis models operate in the 2D domain where disentangling 3D properties such as camera viewpoint or object pose is challenging. Furthermore, they lack an interpretable and controllable representation. The key hypothesis I like to advocate in this talk, is that the image generation process should be modeled in 3D space as the physical world surrounding us is intrinsically three-dimensional. I will define the task o