Towards Unsupervised Learning of Generative Models for 3D Controllable Image Synthesis
This work addresses the problem of interpretable and controllable image synthesis for computer vision and graphics applications, representing an incremental advance by extending generative models to 3D reasoning.
The paper tackles the challenge of disentangling 3D properties like camera viewpoint and object pose in image synthesis by proposing a model that operates in 3D space, enabling unsupervised learning from raw images and synthesizing scenes consistent with viewpoint or pose changes compared to 2D baselines.
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. Our key hypothesis is that the image generation process should be modeled in 3D space as the physical world surrounding us is intrinsically three-dimensional. We define the new task of 3D controllable image synthesis and propose an approach for solving it by reasoning both in 3D space and in the 2D image domain. We demonstrate that our model is able to disentangle latent 3D factors of simple multi-object scenes in an unsupervised fashion from raw images. Compared to pure 2D baselines, it allows for synthesizing scenes that are consistent wrt. changes in viewpoint or object pose. We further evaluate various 3D representations in terms of their usefulness for this challenging task.