A point cloud segmentation network with hybrid convolution and differential channels

Abstract In recent years, point-based segmentation methods have made significant progress in improving segmentation accuracy. However, existing approaches still suffer from several key limitations. Traditional convolution operations make it difficult to effectively model the irregular geometry in po...

Full description

Saved in:
Bibliographic Details
Main Authors: Xiaoyan Zhang, Yantao Bu
Format: Article
Language:English
Published: Nature Portfolio 2025-04-01
Series:Scientific Reports
Subjects:
Online Access:https://doi.org/10.1038/s41598-025-95199-0
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850145422322958336
author Xiaoyan Zhang
Yantao Bu
author_facet Xiaoyan Zhang
Yantao Bu
author_sort Xiaoyan Zhang
collection DOAJ
description Abstract In recent years, point-based segmentation methods have made significant progress in improving segmentation accuracy. However, existing approaches still suffer from several key limitations. Traditional convolution operations make it difficult to effectively model the irregular geometry in point cloud data, resulting in insufficient sensitivity to spatial details. Additionally, current methods have limitations in collaboratively modeling and integrating global and local information. For this reason, we propose a 3D segmentation network based on hybrid convolution and differential channels. Specifically, we design a hybrid convolutional feature extraction (HCFE) module for processing 3D semantic information and spatial information independently, using different convolution kernels to obtain the subtle geometric structure differences between points. Then, we propose a Differential Channel Feature Interaction (DCFI) Module to enhance the local details and global channel information through Differential Convolution (DCU) and a Simplified Channel Attention Mechanism (S_ECA), respectively, and adaptively fuse the two types of information by Dynamic Interaction Mechanism (DIM), achieving their cooperative optimization. Compared to existing methods, HDC_Net has clear advantages in detail capturing and the integration of local and global information. A large number of experiments demonstrate the effectiveness and superiority of the model proposed in this study.
format Article
id doaj-art-eba3d3746e0c4e628b935d88b8667434
institution OA Journals
issn 2045-2322
language English
publishDate 2025-04-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-eba3d3746e0c4e628b935d88b86674342025-08-20T02:28:05ZengNature PortfolioScientific Reports2045-23222025-04-0115111310.1038/s41598-025-95199-0A point cloud segmentation network with hybrid convolution and differential channelsXiaoyan Zhang0Yantao Bu1Computer Science and Technology College, Xi’an University of Science and TechnologyComputer Science and Technology College, Xi’an University of Science and TechnologyAbstract In recent years, point-based segmentation methods have made significant progress in improving segmentation accuracy. However, existing approaches still suffer from several key limitations. Traditional convolution operations make it difficult to effectively model the irregular geometry in point cloud data, resulting in insufficient sensitivity to spatial details. Additionally, current methods have limitations in collaboratively modeling and integrating global and local information. For this reason, we propose a 3D segmentation network based on hybrid convolution and differential channels. Specifically, we design a hybrid convolutional feature extraction (HCFE) module for processing 3D semantic information and spatial information independently, using different convolution kernels to obtain the subtle geometric structure differences between points. Then, we propose a Differential Channel Feature Interaction (DCFI) Module to enhance the local details and global channel information through Differential Convolution (DCU) and a Simplified Channel Attention Mechanism (S_ECA), respectively, and adaptively fuse the two types of information by Dynamic Interaction Mechanism (DIM), achieving their cooperative optimization. Compared to existing methods, HDC_Net has clear advantages in detail capturing and the integration of local and global information. A large number of experiments demonstrate the effectiveness and superiority of the model proposed in this study.https://doi.org/10.1038/s41598-025-95199-0Differential convolution3D semantic segmentation3D part segmentationDeep learning
spellingShingle Xiaoyan Zhang
Yantao Bu
A point cloud segmentation network with hybrid convolution and differential channels
Scientific Reports
Differential convolution
3D semantic segmentation
3D part segmentation
Deep learning
title A point cloud segmentation network with hybrid convolution and differential channels
title_full A point cloud segmentation network with hybrid convolution and differential channels
title_fullStr A point cloud segmentation network with hybrid convolution and differential channels
title_full_unstemmed A point cloud segmentation network with hybrid convolution and differential channels
title_short A point cloud segmentation network with hybrid convolution and differential channels
title_sort point cloud segmentation network with hybrid convolution and differential channels
topic Differential convolution
3D semantic segmentation
3D part segmentation
Deep learning
url https://doi.org/10.1038/s41598-025-95199-0
work_keys_str_mv AT xiaoyanzhang apointcloudsegmentationnetworkwithhybridconvolutionanddifferentialchannels
AT yantaobu apointcloudsegmentationnetworkwithhybridconvolutionanddifferentialchannels
AT xiaoyanzhang pointcloudsegmentationnetworkwithhybridconvolutionanddifferentialchannels
AT yantaobu pointcloudsegmentationnetworkwithhybridconvolutionanddifferentialchannels