# GRASS: Generative Recursive Autoencoders for Shape Structures

arXiv: https://arxiv.org/abs/1705.02090 · HTML: https://mlsanity.com/paper/1705.02090/

## Key facts

- Cited by 24 papers in the map (data through 7 Oct 2026).
- First posted on arXiv 5 May 2017.

## Built on by (most cited first, 24 of 24)

- [Dream3D: Zero-Shot Text-to-3D Synthesis Using 3D Shape Prior and Text-to-Image Diffusion Models](https://mlsanity.com/paper/2212.14704/) (2022)
- [Deep Generative Models in Engineering Design: A Review](https://mlsanity.com/paper/2110.10863/) (2021)
- [Neural Parts: Learning Expressive 3D Shape Abstractions with Invertible Neural Networks](https://mlsanity.com/paper/2103.10429/) (2021)
- [CAPRI-Net: Learning Compact CAD Shapes with Adaptive Primitive Assembly](https://mlsanity.com/paper/2104.05652/) (2021)
- [StructFormer: Learning Spatial Structure for Language-Guided Semantic Rearrangement of Novel Objects](https://mlsanity.com/paper/2110.10189/) (2021)
- [Unsupervised Learning for Cuboid Shape Abstraction via Joint Segmentation from Point Clouds](https://mlsanity.com/paper/2106.03437/) (2021)
- [DeepMetaHandles: Learning Deformation Meta-Handles of 3D Meshes with Biharmonic Coordinates](https://mlsanity.com/paper/2102.09105/) (2021)
- [PartNeRF: Generating Part-Aware Editable 3D Shapes without 3D Supervision](https://mlsanity.com/paper/2303.09554/) (2023)
- [RGL-NET: A Recurrent Graph Learning framework for Progressive Part Assembly](https://mlsanity.com/paper/2107.12859/) (2021)
- [Point Scene Understanding via Disentangled Instance Mesh Reconstruction](https://mlsanity.com/paper/2203.16832/) (2022)
- [Flattening-Net: Deep Regular 2D Representation for 3D Point Cloud Analysis](https://mlsanity.com/paper/2212.08892/) (2022)
- [3D Part Assembly Generation with Instance Encoded Transformer](https://mlsanity.com/paper/2207.01779/) (2022)
- [Hierarchical Neural Semantic Representation for 3D Semantic Correspondence](https://mlsanity.com/paper/2509.17431/) (2025)
- [Fostering Generalization in Single-view 3D Reconstruction by Learning a Hierarchy of Local and Global Shape Priors](https://mlsanity.com/paper/2104.00476/) (2021)
- [Representing Shape Collections with Alignment-Aware Linear Models](https://mlsanity.com/paper/2109.01605/) (2021)
- [Unsupervised 3D Shape Reconstruction by Part Retrieval and Assembly](https://mlsanity.com/paper/2303.01999/) (2023)
- [Neural Shape Compiler: A Unified Framework for Transforming between Text, Point Cloud, and Program](https://mlsanity.com/paper/2212.12952/) (2022)
- [Learning Mesh Representations via Binary Space Partitioning Tree Networks](https://mlsanity.com/paper/2106.14274/) (2021)
- [Learning Body-Aware 3D Shape Generative Models](https://mlsanity.com/paper/2112.07022/) (2021)
- [ShapeMOD: Macro Operation Discovery for 3D Shape Programs](https://mlsanity.com/paper/2104.06392/) (2021)
- [Text-guided 3D Human Generation from 2D Collections](https://mlsanity.com/paper/2305.14312/) (2023)
- [ReGroup: Recursive Neural Networks for Hierarchical Grouping of Vector Graphic Primitives](https://mlsanity.com/paper/2111.11759/) (2021)
- [Learning Implicit Functions for Dense 3D Shape Correspondence of Generic Objects](https://mlsanity.com/paper/2212.14276/) (2022)
- [Evaluating 3D Shape Analysis Methods for Robustness to Rotation Invariance](https://mlsanity.com/paper/2305.18557/) (2023)

## Abstract (arXiv)

We introduce a novel neural network architecture for encoding and synthesis of 3D shapes, particularly their structures. Our key insight is that 3D shapes are effectively characterized by their hierarchical organization of parts, which reflects fundamental intra-shape relationships such as adjacency and symmetry. We develop a recursive neural net (RvNN) based autoencoder to map a flat, unlabeled, arbitrary part layout to a compact code. The code effectively captures hierarchical structures of man-made 3D objects of varying structural complexities despite being fixed-dimensional: an associated decoder maps a code back to a full hierarchy. The learned bidirectional mapping is further tuned using an adversarial setup to yield a generative model of plausible structures, from which novel structures can be sampled. Finally, our structure synthesis framework is augmented by a second trained module that produces fine-grained part geometry, conditioned on global and local structural context, leading to a full generative pipeline for 3D shapes. We demonstrate that without supervision, our network learns meaningful structural hierarchies adhering to perceptual grouping principles, produces compact codes which enable applications such as shape classification and partial matching, and supports shape synthesis and interpolation with significant variations in topology and geometry.