Jul 2026
On the Efficacy of Self-Supervised Point Cloud Encoders for Efficient 3D Large Language Models
This work systematically investigates whether low-cost self-supervised point cloud encoders, specifically PCP-MAE and Point-MAE, can serve as effective alternatives for cost-effective 3D-LLM design and reveals interaction patterns between self-supervised objectives and encoder architectures.
Yao Zheng, Tian Zhang
· arXiv.org · 0 citations