Papers › 3D-JEPA: A Joint Embedding Predictive Architecture for 3D Self-Supervised...

3D-JEPA: A Joint Embedding Predictive Architecture for 3D Self-Supervised Representation Learning

24 Sep 2024arXiv:2409.15803archive 2025-07-28

Naiwen Hu, Haozhe Cheng, Yifan Xie, Shiqi Li, Jihua Zhu

Invariance-based and generative methods have shown a conspicuous performance for 3D self-supervised representation learning (SSRL). However, the former relies on hand-crafted data augmentations that introduce bias not universally applicable to all downstream tasks, and the latter indiscriminately reconstructs masked regions, resulting in irrelevant details being saved in the representation space. To solve the problem above, we introduce 3D-JEPA, a novel non-generative 3D SSRL framework. Specifically, we propose a multi-block sampling strategy that produces a sufficiently informative context block and several representative target blocks. We present the context-aware decoder to enhance the reconstruction of the target blocks. Concretely, the context information is fed to the decoder continuously, facilitating the encoder in learning semantic modeling rather than memorizing the context information related to target blocks. Overall, 3D-JEPA predicts the representation of target blocks from a context block using the encoder and context-aware decoder architecture. Various downstream tasks on different datasets demonstrate 3D-JEPA's effectiveness and efficiency, achieving higher accuracy with fewer pretraining epochs, e.g., 88.65% accuracy on PB_T50_RS with 150 pretraining epochs.

PaperPDF

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

Code

No code repository is listed for this paper in the archive or in Syntology's graph.

Code Syntology ran Syntology

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

Tasks

3D Part Segmentation3D Point Cloud ClassificationDecoderFew-Shot 3D Point Cloud ClassificationRepresentation Learning

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
3D Part Segmentation ShapeNet-Part 3D-JEPA Class Average IoU 86.41 #56 of 67 Archive leaderboard report
3D Part Segmentation ShapeNet-Part 3D-JEPA Instance Average IoU 84.93 #56 of 67 Archive leaderboard report
3D Point Cloud Classification ModelNet40 3D-JEPA Overall Accuracy 94.0 #33 of 111 Archive leaderboard report
3D Point Cloud Classification ScanObjectNN 3D-JEPA OBJ-BG (OA) 93.63 #24 of 77 Archive leaderboard report
3D Point Cloud Classification ScanObjectNN 3D-JEPA OBJ-ONLY (OA) 94.49 #24 of 77 Archive leaderboard report
3D Point Cloud Classification ScanObjectNN 3D-JEPA Overall Accuracy 89.52 #24 of 77 Archive leaderboard report
Few-Shot 3D Point Cloud Classification ModelNet40 10-way (10-shot) 3D-JEPA Overall Accuracy 94.3 #4 of 31 Archive leaderboard report
Few-Shot 3D Point Cloud Classification ModelNet40 10-way (10-shot) 3D-JEPA Standard Deviation 3.6 #4 of 31 Archive leaderboard report
Few-Shot 3D Point Cloud Classification ModelNet40 10-way (20-shot) 3D-JEPA Overall Accuracy 96.3 #3 of 31 Archive leaderboard report
Few-Shot 3D Point Cloud Classification ModelNet40 10-way (20-shot) 3D-JEPA Standard Deviation 2.4 #3 of 31 Archive leaderboard report
Few-Shot 3D Point Cloud Classification ModelNet40 5-way (10-shot) 3D-JEPA Overall Accuracy 97.6 #3 of 30 Archive leaderboard report
Few-Shot 3D Point Cloud Classification ModelNet40 5-way (10-shot) 3D-JEPA Standard Deviation 2.0 #3 of 30 Archive leaderboard report
Few-Shot 3D Point Cloud Classification ModelNet40 5-way (20-shot) 3D-JEPA Overall Accuracy 98.8 #6 of 30 Archive leaderboard report
Few-Shot 3D Point Cloud Classification ModelNet40 5-way (20-shot) 3D-JEPA Standard Deviation 0.4 #6 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.

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