# Shape Inpainting using 3D Generative Adversarial Network and Recurrent Convolutional Networks

arXiv: https://arxiv.org/abs/1711.06375 · HTML: https://mlsanity.com/paper/1711.06375/

## Key facts

- Cited by 23 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 17 Nov 2017.

## Built on by (most cited first, 23 of 23)

- [PoinTr: Diverse Point Cloud Completion with Geometry-Aware Transformers](https://mlsanity.com/paper/2108.08839/) (2021)
- [AdaPoinTr: Diverse Point Cloud Completion with Adaptive Geometry-Aware Transformers](https://mlsanity.com/paper/2301.04545/) (2023)
- [InFusion: Inpainting 3D Gaussians via Learning Depth Completion from Diffusion Prior](https://mlsanity.com/paper/2404.11613/) (2024)
- [Comprehensive Review of Deep Learning-Based 3D Point Cloud Completion Processing and Analysis](https://mlsanity.com/paper/2203.03311/) (2022)
- [Reconstructing Hand-Held Objects from Monocular Video](https://mlsanity.com/paper/2211.16835/) (2022)
- [Voxel-based Network for Shape Completion by Leveraging Edge Generation](https://mlsanity.com/paper/2108.09936/) (2021)
- [DepthLab: From Partial to Complete](https://mlsanity.com/paper/2412.18153/) (2024)
- [Advancements in Point Cloud Data Augmentation for Deep Learning: A Survey](https://mlsanity.com/paper/2308.12113/) (2023)
- [On the use of deep learning for phase recovery](https://mlsanity.com/paper/2308.00942/) (2023)
- [Mutually improved endoscopic image synthesis and landmark detection in unpaired image-to-image translation](https://mlsanity.com/paper/2107.06941/) (2021)
- [FakeNews: GAN-based generation of realistic 3D volumetric data -- A systematic review and taxonomy](https://mlsanity.com/paper/2207.01390/) (2022)
- [KTNet: Knowledge Transfer for Unpaired 3D Shape Completion](https://mlsanity.com/paper/2111.11976/) (2021)
- [T-CorresNet: Template Guided 3D Point Cloud Completion with Correspondence Pooling Query Generation Strategy](https://mlsanity.com/paper/2407.05008/) (2024)
- [Boosting Image Outpainting with Semantic Layout Prediction](https://mlsanity.com/paper/2110.09267/) (2021)
- [Multimodal Shape Completion via IMLE](https://mlsanity.com/paper/2106.16237/) (2021)
- [FruitNinja: 3D Object Interior Texture Generation with Gaussian Splatting](https://mlsanity.com/paper/2411.12089/) (2024)
- [Deep Mesh Prior: Unsupervised Mesh Restoration using Graph Convolutional Networks](https://mlsanity.com/paper/2107.02909/) (2021)
- [Complete Gaussian Splats from a Single Image with Denoising Diffusion Models](https://mlsanity.com/paper/2508.21542/) (2025)
- [3D GANs and Latent Space: A comprehensive survey](https://mlsanity.com/paper/2304.03932/) (2023)
- [Snapshot HDR Video Construction Using Coded Mask](https://mlsanity.com/paper/2112.02522/) (2021)
- [3D Colored Shape Reconstruction from a Single RGB Image through Diffusion](https://mlsanity.com/paper/2302.05573/) (2023)
- [UNeR3D: Versatile and Scalable 3D RGB Point Cloud Generation from 2D Images in Unsupervised Reconstruction](https://mlsanity.com/paper/2312.06706/) (2023)
- [RefComp: A Reference-guided Unified Framework for Unpaired Point Cloud Completion](https://mlsanity.com/paper/2504.13788/) (2025)

## Abstract (arXiv)

Recent advances in convolutional neural networks have shown promising results in 3D shape completion. But due to GPU memory limitations, these methods can only produce low-resolution outputs. To inpaint 3D models with semantic plausibility and contextual details, we introduce a hybrid framework that combines a 3D Encoder-Decoder Generative Adversarial Network (3D-ED-GAN) and a Long-term Recurrent Convolutional Network (LRCN). The 3D-ED-GAN is a 3D convolutional neural network trained with a generative adversarial paradigm to fill missing 3D data in low-resolution. LRCN adopts a recurrent neural network architecture to minimize GPU memory usage and incorporates an Encoder-Decoder pair into a Long Short-term Memory Network. By handling the 3D model as a sequence of 2D slices, LRCN transforms a coarse 3D shape into a more complete and higher resolution volume. While 3D-ED-GAN captures global contextual structure of the 3D shape, LRCN localizes the fine-grained details. Experimental results on both real-world and synthetic data show reconstructions from corrupted models result in complete and high-resolution 3D objects.