.png){kind=icon}
返回
Hyperbolic prototype rectification for few-shot 3D point cloud classification
.png){kind=icon}
PRE
AI
DOI：10.1016/j.patcog.2024.111042.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Few-shot point cloud classification is a challenging task in 3D computer vision and has received widespread attention from researchers. Most of the works on deep learning models rely heavily on Euclidean spatial metrics. However, point cloud objects often have complex non-Euclidean geometric structures, with underlying inter/intra-class hierarchical structures, which are difficult to capture by current Euclidean-based deep learning models. Moreover, due to the lack of training samples, many few-shot learning methods often suffer from the overfitting problem. Given the Hyperbolic metric of non-Euclidean geometry offering hierarchical structural prior, as we assume to be able to assist FSL task, we propose Hyperbolic Prototype Rectification (HPR) HPR ) for few- shot point cloud classification, without requiring extra learnable parameter. Firstly, point clouds are embedded into hyperbolic space to better describe hierarchical similarity relationships in data. Secondly, the HPR utilizes hyperbolic spatial and distributional information to enhance the feature representation and improve the generalization capability, with more appropriate hyperbolic prototypes. The few-shot classification experiments and further ablation studies conducted on widely used point cloud datasets demonstrate the effectiveness of our method. On the real-world ScanObjectNN(-PB) datasets, the average classification accuracy outperforms the SOTA method by 2.08%(0.66%), respectively, indicating that the proposed HPR has great generalization capability and strong robustness against perturbed data. Our code is available at: https://github.com/JonathanUCAS/HPR.
Keyword:
Hyperbolic geometry
Few-shot learning
Point cloud classification
Prototype rectification
Feature enhancement
期刊
IF:
7.6
论文数:
1.3W
被引数:
4.5W
机构
引用论文
Meta-learning approaches for learning-to-learn in deep learning: A survey深度学习中学习到学习的元学习方法: 一项调查.png)
NEUROCOMPUTING
NEUROCOMPUTINGIF6.5
PREAI
分享收藏A Comprehensive Survey of Few-shot Learning: Evolution, Applications, Challenges, and Opportunities
OAAI
OAAI分享收藏Correlation model of deflection, vehicle load, and temperature for in‐service bridge using deep learning and structural health monitoring基于深度学习和结构健康监测的在役桥梁挠度、车辆荷载和温度关联模型OAAI
OAAI分享收藏Rotation invariant point cloud analysis: Where local geometry meets global topologyPATTERN RECOGNITION
PATTERN RECOGNITIONIF7.6
PREAI
分享收藏分享收藏Learning geometric consistency and discrepancy for category-level 6D object pose estimation from point clouds从点云学习类别级6D物体姿态估计的几何一致性和差异PATTERN RECOGNITION
PATTERN RECOGNITIONIF7.6
PREAI
分享收藏
查看更多内容