Metraipolis › archive › dc_db933ae5bc25
Joint-Embedding Prediction of Masked Point Tubes for Self-Supervised Learning on 4D Point Cloud Videos — arXiv:2608.24093
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.24093
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-18T06:36:42.513Z
- SHA-256 of the archived text
d416551efab799f40b6b6c1ac3abb3e8ee086e6eecc478a648b8df8c90cf17c1- Stable URL of this page
https://metraipolis.com/documento/dc_db933ae5bc25
How to cite
Joint-Embedding Prediction of Masked Point Tubes for Self-Supervised Learning on 4D Point Cloud Videos — arXiv:2608.24093 — arXiv (en), https://arxiv.org/abs/2608.24093 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_db933ae5bc25 (Metraipolis archive dc_db933ae5bc25, sha256 d416551efab799f40b6b6c1ac3abb3e8ee086e6eecc478a648b8df8c90cf17c1)
BibTeX
@misc{dcdb933ae5bc25,
title = {Joint-Embedding Prediction of Masked Point Tubes for Self-Supervised Learning on 4D Point Cloud Videos — arXiv:2608.24093},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.24093},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_db933ae5bc25; sha256 d416551efab799f40b6b6c1ac3abb3e8ee086e6eecc478a648b8df8c90cf17c1},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.24093. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Self-supervised representation learning for 4D point cloud videos is challenging because annotations are costly and reconstruction-based pretraining can overemphasize low-level geometric details. We propose a JEPA-style framework that learns from unlabeled spatiotemporal point clouds through latent point-tube prediction. Instead of reconstructing raw coordinates, the model masks spatiotemporal regions and predicts their target representations from visible context representations in feature space. To stabilize latent prediction, we incorporate Sketched Isotropic Gaussian Regularization, which encourages non-collapsed embeddings without relying on explicit reconstruction targets. This formulation aims to capture both spatial structure and temporal dynamics while keeping the pretraining objective aligned with downstream semantic recognition. Experiments on action and gesture recognition benchmarks show that the learned representations improve downstream fine-tuning, limited-label learning, and cross-dataset transfer. These results suggest that JEPA-style latent prediction is a promising alternative to reconstruction-centered pretraining for 4D point cloud videos. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-26 in cs.LG). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2608.24093 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-18T06:28:52.337Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-08-26?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).