BlockGAN: Learning 3D Object-aware Scene Representations from Unlabelled Images
This addresses the problem of scene compositionality for computer vision and graphics researchers, offering a novel approach to 3D representation learning from 2D data, though it builds on existing generative models.
BlockGAN tackles learning 3D object-aware scene representations from unlabelled 2D images by generating 3D features for background and foreground objects, combining them, and rendering realistic images, resulting in disentangled representations for objects and their properties without needing 3D geometry or labels.
We present BlockGAN, an image generative model that learns object-aware 3D scene representations directly from unlabelled 2D images. Current work on scene representation learning either ignores scene background or treats the whole scene as one object. Meanwhile, work that considers scene compositionality treats scene objects only as image patches or 2D layers with alpha maps. Inspired by the computer graphics pipeline, we design BlockGAN to learn to first generate 3D features of background and foreground objects, then combine them into 3D features for the wholes cene, and finally render them into realistic images. This allows BlockGAN to reason over occlusion and interaction between objects' appearance, such as shadow and lighting, and provides control over each object's 3D pose and identity, while maintaining image realism. BlockGAN is trained end-to-end, using only unlabelled single images, without the need for 3D geometry, pose labels, object masks, or multiple views of the same scene. Our experiments show that using explicit 3D features to represent objects allows BlockGAN to learn disentangled representations both in terms of objects (foreground and background) and their properties (pose and identity).