# Learning from 2D: Contrastive Pixel-to-Point Knowledge Transfer for 3D Pretraining

arXiv: https://arxiv.org/abs/2104.04687 · HTML: https://mlsanity.com/paper/2104.04687/

## Key facts

- Cited by 33 papers in the map (data through 14 Oct 2026).
- Builds on 65 of its 87 references that are in the map.
- First posted on arXiv 10 Apr 2021.

## Builds on (most cited first, 30 of 65)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [A Simple Framework for Contrastive Learning of Visual Representations](https://mlsanity.com/paper/2002.05709/) (2020)
- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Representation Learning with Contrastive Predictive Coding](https://mlsanity.com/paper/1807.03748/) (2018)
- [Momentum Contrast for Unsupervised Visual Representation Learning](https://mlsanity.com/paper/1911.05722/) (2019)
- [PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation](https://mlsanity.com/paper/1612.00593/) (2016)
- [Fully Convolutional Networks for Semantic Segmentation](https://mlsanity.com/paper/1411.4038/) (2014)
- [Bootstrap your own latent: A new approach to self-supervised Learning](https://mlsanity.com/paper/2006.07733/) (2020)
- [nuScenes: A multimodal dataset for autonomous driving](https://mlsanity.com/paper/1903.11027/) (2019)
- [PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space](https://mlsanity.com/paper/1706.02413/) (2017)
- [ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenes](https://mlsanity.com/paper/1702.04405/) (2017)
- [Dynamic Graph CNN for Learning on Point Clouds](https://mlsanity.com/paper/1801.07829/) (2018)
- [Unsupervised Feature Learning via Non-Parametric Instance-level Discrimination](https://mlsanity.com/paper/1805.01978/) (2018)
- [FitNets: Hints for Thin Deep Nets](https://mlsanity.com/paper/1412.6550/) (2014)
- [Unsupervised Representation Learning by Predicting Image Rotations](https://mlsanity.com/paper/1803.07728/) (2018)
- [Context Encoders: Feature Learning by Inpainting](https://mlsanity.com/paper/1604.07379/) (2016)
- [Unsupervised Learning of Visual Representations by Solving Jigsaw Puzzles](https://mlsanity.com/paper/1603.09246/) (2016)
- [Big Self-Supervised Models are Strong Semi-Supervised Learners](https://mlsanity.com/paper/2006.10029/) (2020)
- [Contrastive Multiview Coding](https://mlsanity.com/paper/1906.05849/) (2019)
- [Learning deep representations by mutual information estimation and maximization](https://mlsanity.com/paper/1808.06670/) (2018)
- [4D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networks](https://mlsanity.com/paper/1904.08755/) (2019)
- [KPConv: Flexible and Deformable Convolution for Point Clouds](https://mlsanity.com/paper/1904.08889/) (2019)
- [Colorful Image Colorization](https://mlsanity.com/paper/1603.08511/) (2016)
- [Paying More Attention to Attention: Improving the Performance of Convolutional Neural Networks via Attention Transfer](https://mlsanity.com/paper/1612.03928/) (2016)
- [Unsupervised Visual Representation Learning by Context Prediction](https://mlsanity.com/paper/1505.05192/) (2015)
- [Relational Knowledge Distillation](https://mlsanity.com/paper/1904.05068/) (2019)
- [Self-Supervised Learning of Pretext-Invariant Representations](https://mlsanity.com/paper/1912.01991/) (2019)
- [PointCNN: Convolution On X-Transformed Points](https://mlsanity.com/paper/1801.07791/) (2018)
- [3D Semantic Segmentation with Submanifold Sparse Convolutional Networks](https://mlsanity.com/paper/1711.10275/) (2017)

## Built on by (most cited first, 30 of 33)

- [CrossPoint: Self-Supervised Cross-Modal Contrastive Learning for 3D Point Cloud Understanding](https://mlsanity.com/paper/2203.00680/) (2022)
- [CLIP2Scene: Towards Label-efficient 3D Scene Understanding by CLIP](https://mlsanity.com/paper/2301.04926/) (2023)
- [2DPASS: 2D Priors Assisted Semantic Segmentation on LiDAR Point Clouds](https://mlsanity.com/paper/2207.04397/) (2022)
- [Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders](https://mlsanity.com/paper/2212.06785/) (2022)
- [Image-to-Lidar Self-Supervised Distillation for Autonomous Driving Data](https://mlsanity.com/paper/2203.16258/) (2022)
- [CLIP-FO3D: Learning Free Open-world 3D Scene Representations from 2D Dense CLIP](https://mlsanity.com/paper/2303.04748/) (2023)
- [RangeViT: Towards Vision Transformers for 3D Semantic Segmentation in Autonomous Driving](https://mlsanity.com/paper/2301.10222/) (2023)
- [UniPAD: A Universal Pre-training Paradigm for Autonomous Driving](https://mlsanity.com/paper/2310.08370/) (2023)
- [X-Trans2Cap: Cross-Modal Knowledge Transfer using Transformer for 3D Dense Captioning](https://mlsanity.com/paper/2203.00843/) (2022)
- [ALSO: Automotive Lidar Self-supervision by Occupancy estimation](https://mlsanity.com/paper/2212.05867/) (2022)
- [SimIPU: Simple 2D Image and 3D Point Cloud Unsupervised Pre-Training for Spatial-Aware Visual Representations](https://mlsanity.com/paper/2112.04680/) (2021)
- [CLIP-Guided Vision-Language Pre-training for Question Answering in 3D Scenes](https://mlsanity.com/paper/2304.06061/) (2023)
- [Three Pillars improving Vision Foundation Model Distillation for Lidar](https://mlsanity.com/paper/2310.17504/) (2023)
- [Image2Point: 3D Point-Cloud Understanding with 2D Image Pretrained Models](https://mlsanity.com/paper/2106.04180/) (2021)
- [TiG-BEV: Multi-view BEV 3D Object Detection via Target Inner-Geometry Learning](https://mlsanity.com/paper/2212.13979/) (2022)
- [Self-Supervised Image-to-Point Distillation via Semantically Tolerant Contrastive Loss](https://mlsanity.com/paper/2301.05709/) (2023)
- [A Closer Look at Invariances in Self-supervised Pre-training for 3D Vision](https://mlsanity.com/paper/2207.04997/) (2022)
- [Let Images Give You More:Point Cloud Cross-Modal Training for Shape Analysis](https://mlsanity.com/paper/2210.04208/) (2022)
- [BEVContrast: Self-Supervision in BEV Space for Automotive Lidar Point Clouds](https://mlsanity.com/paper/2310.17281/) (2023)
- [Open-Vocabulary 3D Detection via Image-level Class and Debiased Cross-modal Contrastive Learning](https://mlsanity.com/paper/2207.01987/) (2022)
- [Multi-CLIP: Contrastive Vision-Language Pre-training for Question Answering tasks in 3D Scenes](https://mlsanity.com/paper/2306.02329/) (2023)
- [DetMatch: Two Teachers are Better Than One for Joint 2D and 3D Semi-Supervised Object Detection](https://mlsanity.com/paper/2203.09510/) (2022)
- [PointMCD: Boosting Deep Point Cloud Encoders via Multi-view Cross-modal Distillation for 3D Shape Recognition](https://mlsanity.com/paper/2207.03128/) (2022)
- [Can We Solve 3D Vision Tasks Starting from A 2D Vision Transformer?](https://mlsanity.com/paper/2209.07026/) (2022)
- [Self-Supervised Learning with Multi-View Rendering for 3D Point Cloud Analysis](https://mlsanity.com/paper/2210.15904/) (2022)
- [Geometric-aware Pretraining for Vision-centric 3D Object Detection](https://mlsanity.com/paper/2304.03105/) (2023)
- [Multi-Scale Neighborhood Occupancy Masked Autoencoder for Self-Supervised Learning in LiDAR Point Clouds](https://mlsanity.com/paper/2502.20316/) (2025)
- [Language-Assisted 3D Feature Learning for Semantic Scene Understanding](https://mlsanity.com/paper/2211.14091/) (2022)
- [Self-Supervised Generative-Contrastive Learning of Multi-Modal Euclidean Input for 3D Shape Latent Representations: A Dynamic Switching Approach](https://mlsanity.com/paper/2301.04612/) (2023)
- [Applying Plain Transformers to Real-World Point Clouds](https://mlsanity.com/paper/2303.00086/) (2023)

## Abstract (arXiv)

Most 3D neural networks are trained from scratch owing to the lack of large-scale labeled 3D datasets. In this paper, we present a novel 3D pretraining method by leveraging 2D networks learned from rich 2D datasets. We propose the contrastive pixel-to-point knowledge transfer to effectively utilize the 2D information by mapping the pixel-level and point-level features into the same embedding space. Due to the heterogeneous nature between 2D and 3D networks, we introduce the back-projection function to align the features between 2D and 3D to make the transfer possible. Additionally, we devise an upsampling feature projection layer to increase the spatial resolution of high-level 2D feature maps, which enables learning fine-grained 3D representations. With a pretrained 2D network, the proposed pretraining process requires no additional 2D or 3D labeled data, further alleviating the expensive 3D data annotation cost. To the best of our knowledge, we are the first to exploit existing 2D trained weights to pretrain 3D deep neural networks. Our intensive experiments show that the 3D models pretrained with 2D knowledge boost the performances of 3D networks across various real-world 3D downstream tasks.