%0 Generic %1 richter2018matryoshka %A Richter, Stephan R. %A Roth, Stefan %D 2018 %K 2018 3D arxiv cvpr deep-learning paper %T Matryoshka Networks: Predicting 3D Geometry via Nested Shape Layers %U http://arxiv.org/abs/1804.10975 %X In this paper, we develop novel, efficient 2D encodings for 3D geometry, which enable reconstructing full 3D shapes from a single image at high resolution. The key idea is to pose 3D shape reconstruction as a 2D prediction problem. To that end, we first develop a simple baseline network that predicts entire voxel tubes at each pixel of a reference view. By leveraging well-proven architectures for 2D pixel-prediction tasks, we attain state-of-the-art results, clearly outperforming purely voxel-based approaches. We scale this baseline to higher resolutions by proposing a memory-efficient shape encoding, which recursively decomposes a 3D shape into nested shape layers, similar to the pieces of a Matryoshka doll. This allows reconstructing highly detailed shapes with complex topology, as demonstrated in extensive experiments; we clearly outperform previous octree-based approaches despite having a much simpler architecture using standard network components. Our Matryoshka networks further enable reconstructing shapes from IDs or shape similarity, as well as shape sampling.

@misc{richter2018matryoshka, abstract = {In this paper, we develop novel, efficient 2D encodings for 3D geometry, which enable reconstructing full 3D shapes from a single image at high resolution. The key idea is to pose 3D shape reconstruction as a 2D prediction problem. To that end, we first develop a simple baseline network that predicts entire voxel tubes at each pixel of a reference view. By leveraging well-proven architectures for 2D pixel-prediction tasks, we attain state-of-the-art results, clearly outperforming purely voxel-based approaches. We scale this baseline to higher resolutions by proposing a memory-efficient shape encoding, which recursively decomposes a 3D shape into nested shape layers, similar to the pieces of a Matryoshka doll. This allows reconstructing highly detailed shapes with complex topology, as demonstrated in extensive experiments; we clearly outperform previous octree-based approaches despite having a much simpler architecture using standard network components. Our Matryoshka networks further enable reconstructing shapes from IDs or shape similarity, as well as shape sampling.}, added-at = {2018-08-01T17:48:50.000+0200}, author = {Richter, Stephan R. and Roth, Stefan}, biburl = {https://www.bibsonomy.org/bibtex/2d1a99f699856d288d19e9745f0b26c6e/analyst}, description = {[1804.10975] Matryoshka Networks: Predicting 3D Geometry via Nested Shape Layers}, interhash = {6bf8ffd92d4743de0d06e08aabefe765}, intrahash = {d1a99f699856d288d19e9745f0b26c6e}, keywords = {2018 3D arxiv cvpr deep-learning paper}, note = {cite arxiv:1804.10975Comment: Published at the Conference on Computer Vision and Pattern Recognition (CVPR 2018)}, timestamp = {2018-08-01T17:48:50.000+0200}, title = {Matryoshka Networks: Predicting 3D Geometry via Nested Shape Layers}, url = {http://arxiv.org/abs/1804.10975}, year = 2018 }