Papers › Point-M2AE: Multi-scale Masked Autoencoders for Hierarchical Point Cloud Pre-training

Point-M2AE: Multi-scale Masked Autoencoders for Hierarchical Point Cloud Pre-training

28 May 2022arXiv:2205.14401archive 2025-07-28

Renrui Zhang, Ziyu Guo, Rongyao Fang, Bin Zhao, Dong Wang, Yu Qiao, Hongsheng Li, Peng Gao

Masked Autoencoders (MAE) have shown great potentials in self-supervised pre-training for language and 2D image transformers. However, it still remains an open question on how to exploit masked autoencoding for learning 3D representations of irregular point clouds. In this paper, we propose Point-M2AE, a strong Multi-scale MAE pre-training framework for hierarchical self-supervised learning of 3D point clouds. Unlike the standard transformer in MAE, we modify the encoder and decoder into pyramid architectures to progressively model spatial geometries and capture both fine-grained and high-level semantics of 3D shapes. For the encoder that downsamples point tokens by stages, we design a multi-scale masking strategy to generate consistent visible regions across scales, and adopt a local spatial self-attention mechanism during fine-tuning to focus on neighboring patterns. By multi-scale token propagation, the lightweight decoder gradually upsamples point tokens with complementary skip connections from the encoder, which further promotes the reconstruction from a global-to-local perspective. Extensive experiments demonstrate the state-of-the-art performance of Point-M2AE for 3D representation learning. With a frozen encoder after pre-training, Point-M2AE achieves 92.9% accuracy for linear SVM on ModelNet40, even surpassing some fully trained methods. By fine-tuning on downstream tasks, Point-M2AE achieves 86.43% accuracy on ScanObjectNN, +3.36% to the second-best, and largely benefits the few-shot classification, part segmentation and 3D object detection with the hierarchical pre-training scheme. Code is available at https://github.com/ZrrSkywalker/Point-M2AE.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

zrrskywalker/point-m2ae officialmentioned in papermentioned on GitHubpytorchMIT report
zrrskywalker/i2p-mae mentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

3D Object Detection3D Point Cloud Classification3D Point Cloud Linear ClassificationDecoderFew-Shot 3D Point Cloud ClassificationObject DetectionOpen-Ended Question AnsweringPoint Cloud Pre-trainingRepresentation LearningSelf-Supervised Learningobject-detection

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
3D Point Cloud Classification ModelNet40 Point-M2AE Overall Accuracy 94.0 #31 of 111 Archive leaderboard report
3D Point Cloud Classification ModelNet40 Point-M2AE-SVM Overall Accuracy 92.9 #82 of 111 Archive leaderboard report
3D Point Cloud Classification ScanObjectNN Point-M2AE OBJ-BG (OA) 91.22 #49 of 77 Archive leaderboard report
3D Point Cloud Classification ScanObjectNN Point-M2AE OBJ-ONLY (OA) 88.81 #49 of 77 Archive leaderboard report
3D Point Cloud Classification ScanObjectNN Point-M2AE Overall Accuracy 86.43 #49 of 77 Archive leaderboard report
3D Point Cloud Linear Classification ModelNet40 Point-M2AE Overall Accuracy 92.9 #5 of 20 Archive leaderboard report
Few-Shot 3D Point Cloud Classification ModelNet40 10-way (10-shot) Point-M2AE Overall Accuracy 92.3 #16 of 31 Archive leaderboard report
Few-Shot 3D Point Cloud Classification ModelNet40 10-way (10-shot) Point-M2AE Standard Deviation 4.5 #16 of 31 Archive leaderboard report
Few-Shot 3D Point Cloud Classification ModelNet40 10-way (20-shot) Point-M2AE Overall Accuracy 95.0 #15 of 31 Archive leaderboard report
Few-Shot 3D Point Cloud Classification ModelNet40 10-way (20-shot) Point-M2AE Standard Deviation 3.0 #15 of 31 Archive leaderboard report
Few-Shot 3D Point Cloud Classification ModelNet40 5-way (10-shot) Point-M2AE Overall Accuracy 96.8 #13 of 30 Archive leaderboard report
Few-Shot 3D Point Cloud Classification ModelNet40 5-way (10-shot) Point-M2AE Standard Deviation 1.8 #13 of 30 Archive leaderboard report
Few-Shot 3D Point Cloud Classification ModelNet40 5-way (20-shot) Point-M2AE Overall Accuracy 98.3 #11 of 30 Archive leaderboard report
Few-Shot 3D Point Cloud Classification ModelNet40 5-way (20-shot) Point-M2AE Standard Deviation 1.4 #11 of 30 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

MAESVM

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections