基于多尺度特征融合的由粗到精点云形状补全
Coarse-to-Fine Point Cloud Shape Complementation Based on Multi-Scale Feature Fusion
张德军 1王杨 1谭雪峰 1吴亦奇 1陈壹林 2何发智3
作者信息
- 1. 中国地质大学(武汉)计算机学院 武汉 430078
- 2. 智能机器人湖北省重点实验室(武汉工程大学)武汉 430205
- 3. 武汉大学计算机学院 武汉 430072
- 折叠
摘要
为了以由粗到精的方式实现点云形状补全,提出一个端到端的两阶段多尺度特征融合网络,其中的每个阶段都是由一个编码器-解码器构成.第1阶段中,首先利用点集抽取模块提取残缺点云的全局特征,在获取不同分辨率点特征的同时能关注更多的局部邻域特征,然后使用多层感知机作为解码器生成粗糙的点云骨架;第2阶段中,利用点云骨架和残缺点云提取多尺度局部特征,并通过注意力机制与第1阶段中的多尺度全局特征相互融合,使得每个点都包含全局和局部几何信息;最后将第2阶段中的全局特征和多尺度局部特征逐步进行上采样,并通过多层感知机生成精细的完整点云.采用倒角距离作为评价标准,在ShapeNet,MVP和Completion3D数据集上进行点云补全实验的结果表明,误差分别比基准网络降低17.1％,3.9％和13.9％,验证了所提网络的有效性.
Abstract
To implement the point cloud shape completion in a coarse-to-fine manner,an end-to-end two-stage multi-scale feature fusion network is proposed,in which each stage consists of an encoder-decoder.In the first stage,the set abstraction module extracts the global features of the incomplete point cloud,which can focus on more local neighborhood features while acquiring point features of different resolutions.A decoder built from multilayer perceptrons generates a coarse skeleton.In the second stage,the coarse skeleton and incomplete point cloud are used to learn multi-scale local features.The multi-scale local features are fused with the multi-scale global features of the first stage via an attention mechanism so that each point contains global and local geometric information.Finally,the global features and multi-scale local features are progressively upsampled,and a fine-grained complete point cloud is generated via multilayer perceptrons.The point cloud completion experi-ments on ShapeNet,MVP,and Completion3D datasets show that the chamfer distance is reduced by 17.1％,3.9％,and 13.9％compared with the baseline,respectively,demonstrating the effectiveness of the proposed method.
关键词
点云补全/多尺度特征融合/由粗到精/编码器-解码器Key words
point cloud completion/multi-scale feature fusion/coarse-to-fine/encoder-decoder引用本文复制引用
基金项目
国家自然科学基金(61802355)
智能机器人湖北省重点实验室开放基金(HBIR 202105)
出版年
2024
NETL
NSTL
万方数据