PointMamba: An Efficient State Space Model for Point Cloud Analysis
PointMamba, a state space model-based framework, achieves global modeling with linear complexity for point cloud analysis tasks.
Dense Supervision Propagation for Weakly Supervised Semantic Segmentation on 3D Point Clouds
The authors propose a novel two-stage training strategy that leverages cross-sample and intra-sample feature reallocation to densely propagate supervision signals from a small portion of labeled points to the unlabeled points, enabling weakly supervised semantic segmentation on 3D point clouds.
Efficient Meta-Episodic Learning with Dynamic Task Sampling for Improved CLIP-based Point Cloud Classification
A novel meta-episodic learning framework with dynamic task sampling is proposed to effectively encode unknown generalized class information into CLIP-based point cloud classification models, enabling improved performance on challenging and underrepresented classes.
Efficient Extraction of Geometric Information from Point Cloud Data
A kernel-based method is proposed to construct signature (defining) functions of subsets of Rd, ranging from full dimensional manifolds to point clouds. The signature function can be used to estimate the dimension, normal, and curvatures of the interpolated surface, without requiring explicit knowledge of local neighborhoods or other data structure.
© 2024 by Linnk AI