# Self-Supervised Point Cloud Completion via Inpainting

arXiv: https://arxiv.org/abs/2111.10701 · HTML: https://mlsanity.com/paper/2111.10701/

## Key facts

- Cited by 16 papers in the map (data through 14 Oct 2026).
- Builds on 23 of its 30 references that are in the map.
- First posted on arXiv 21 Nov 2021.

## Builds on (most cited first, 23 of 23)

- [ShapeNet: An Information-Rich 3D Model Repository](https://mlsanity.com/paper/1512.03012/) (2015)
- [Context Encoders: Feature Learning by Inpainting](https://mlsanity.com/paper/1604.07379/) (2016)
- [Generative Image Inpainting with Contextual Attention](https://mlsanity.com/paper/1801.07892/) (2018)
- [PCN: Point Completion Network](https://mlsanity.com/paper/1808.00671/) (2018)
- [GRNet: Gridding Residual Network for Dense Point Cloud Completion](https://mlsanity.com/paper/2006.03761/) (2020)
- [Semantic Image Inpainting with Deep Generative Models](https://mlsanity.com/paper/1607.07539/) (2016)
- [PF-Net: Point Fractal Network for 3D Point Cloud Completion](https://mlsanity.com/paper/2003.00410/) (2020)
- [Morphing and Sampling Network for Dense Point Cloud Completion](https://mlsanity.com/paper/1912.00280/) (2019)
- [Cascaded Refinement Network for Point Cloud Completion](https://mlsanity.com/paper/2004.03327/) (2020)
- [Perspective Transformer Nets: Learning Single-View 3D Object Reconstruction without 3D Supervision](https://mlsanity.com/paper/1612.00814/) (2016)
- [Point Cloud Completion by Skip-attention Network with Hierarchical Folding](https://mlsanity.com/paper/2005.03871/) (2020)
- [Rethinking Image Inpainting via a Mutual Encoder-Decoder with Feature Equalizations](https://mlsanity.com/paper/2007.06929/) (2020)
- [Self-Supervised Scene De-occlusion](https://mlsanity.com/paper/2004.02788/) (2020)
- [Unsupervised Learning of Shape and Pose with Differentiable Point Clouds](https://mlsanity.com/paper/1810.09381/) (2018)
- [Multi-view Consistency as Supervisory Signal for Learning Shape and Pose Prediction](https://mlsanity.com/paper/1801.03910/) (2018)
- [SoftPoolNet: Shape Descriptor for Point Cloud Completion and Classification](https://mlsanity.com/paper/2008.07358/) (2020)
- [Weakly-supervised 3D Shape Completion in the Wild](https://mlsanity.com/paper/2008.09110/) (2020)
- [Rethinking Reprojection: Closing the Loop for Pose-aware ShapeReconstruction from a Single Image](https://mlsanity.com/paper/1707.04682/) (2017)
- [Deep Fusion Network for Image Completion](https://mlsanity.com/paper/1904.08060/) (2019)
- [Point Cloud Completion by Learning Shape Priors](https://mlsanity.com/paper/2008.00394/) (2020)
- [Local Frequency Interpretation and Non-Local Self-Similarity on Graph for Point Cloud Inpainting](https://mlsanity.com/paper/1810.03973/) (2018)
- [Iterative Transformer Network for 3D Point Cloud](https://mlsanity.com/paper/1811.11209/) (2018)
- [3D Dynamic Point Cloud Inpainting via Temporal Consistency on Graphs](https://mlsanity.com/paper/1904.10795/) (2019)

## Built on by (most cited first, 16 of 16)

- [Cross-modal Learning for Image-Guided Point Cloud Shape Completion](https://mlsanity.com/paper/2209.09552/) (2022)
- [P2C: Self-Supervised Point Cloud Completion from Single Partial Clouds](https://mlsanity.com/paper/2307.14726/) (2023)
- [ACL-SPC: Adaptive Closed-Loop system for Self-Supervised Point Cloud Completion](https://mlsanity.com/paper/2303.01979/) (2023)
- [Clutter Detection and Removal in 3D Scenes with View-Consistent Inpainting](https://mlsanity.com/paper/2304.03763/) (2023)
- [GenPC: Zero-shot Point Cloud Completion via 3D Generative Priors](https://mlsanity.com/paper/2502.19896/) (2025)
- [Towards Learning to Complete Anything in Lidar](https://mlsanity.com/paper/2504.12264/) (2025)
- [3D Shape Completion on Unseen Categories:A Weakly-supervised Approach](https://mlsanity.com/paper/2401.10578/) (2024)
- [Unsupervised 3D Point Cloud Completion via Multi-view Adversarial Learning](https://mlsanity.com/paper/2407.09786/) (2024)
- [GS-RoadPatching: Inpainting Gaussians via 3D Searching and Placing for Driving Scenes](https://mlsanity.com/paper/2509.19937/) (2025)
- [CNC-Net: Self-Supervised Learning for CNC Machining Operations](https://mlsanity.com/paper/2312.09925/) (2023)
- [Hyperbolic Chamfer Distance for Point Cloud Completion and Beyond](https://mlsanity.com/paper/2412.17951/) (2024)
- [Learning Self-Prior for Mesh Inpainting Using Self-Supervised Graph Convolutional Networks](https://mlsanity.com/paper/2305.00635/) (2023)
- [Self-Supervised Large Scale Point Cloud Completion for Archaeological Site Restoration](https://mlsanity.com/paper/2503.04030/) (2025)
- [Are All Point Clouds Suitable for Completion? Weakly Supervised Quality Evaluation Network for Point Cloud Completion](https://mlsanity.com/paper/2303.01804/) (2023)
- [Defurnishing with X-Ray Vision: Joint Removal of Furniture from Panoramas and Mesh](https://mlsanity.com/paper/2506.05338/) (2025)
- [Self-Supervised Point Cloud Completion based on Multi-View Augmentations of Single Partial Point Cloud](https://mlsanity.com/paper/2509.22132/) (2025)

## Abstract (arXiv)

When navigating in urban environments, many of the objects that need to be tracked and avoided are heavily occluded. Planning and tracking using these partial scans can be challenging. The aim of this work is to learn to complete these partial point clouds, giving us a full understanding of the object's geometry using only partial observations. Previous methods achieve this with the help of complete, ground-truth annotations of the target objects, which are available only for simulated datasets. However, such ground truth is unavailable for real-world LiDAR data. In this work, we present a self-supervised point cloud completion algorithm, PointPnCNet, which is trained only on partial scans without assuming access to complete, ground-truth annotations. Our method achieves this via inpainting. We remove a portion of the input data and train the network to complete the missing region. As it is difficult to determine which regions were occluded in the initial cloud and which were synthetically removed, our network learns to complete the full cloud, including the missing regions in the initial partial cloud. We show that our method outperforms previous unsupervised and weakly-supervised methods on both the synthetic dataset, ShapeNet, and real-world LiDAR dataset, Semantic KITTI.