Ground-Based Remote Sensing Cloud Image Segmentation Using Convolution-MLP Network

Recently, multilayer perceptron (MLPs)-based methods in computer vision have attracted much attention due to the ability of learning long-range dependencies. However, MLPs-based methods usually treat all the tokens equally, which is difficult to segment challenging cloud regions. In this article, we...

Full description

Saved in:
Bibliographic Details
Main Authors: Shuang Liu, Jiafeng Zhang, Zhong Zhang, Shuzhen Hu, Baihua Xiao
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11098898/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849762555144175616
author Shuang Liu
Jiafeng Zhang
Zhong Zhang
Shuzhen Hu
Baihua Xiao
author_facet Shuang Liu
Jiafeng Zhang
Zhong Zhang
Shuzhen Hu
Baihua Xiao
author_sort Shuang Liu
collection DOAJ
description Recently, multilayer perceptron (MLPs)-based methods in computer vision have attracted much attention due to the ability of learning long-range dependencies. However, MLPs-based methods usually treat all the tokens equally, which is difficult to segment challenging cloud regions. In this article, we propose a novel network named convolution-MLP network (Con-MLPNet) for ground-based remote sensing cloud image segmentation, which could effectively learn long-range dependencies via the combination of MLPs and the attention mechanism. To this end, we propose the attention-guided MLPs module to highlight salient features and suppress irrelevant features from the spatial and channel aspects. Meanwhile, different from existing MLPs methods where the long-range dependencies are learned from one single scale, we propose the dilated MLPs (DMLPs) to learn long-range dependencies at different scales by sampling different channels of tokens. Furthermore, we design the parallel dilated MLPs module to integrate multiple DMLPs with different parameters in order to extract multiscale information. We conduct a series of experiments on three public ground-based cloud image segmentation datasets, i.e., TLCDD, SWIMSEG, and TCDD, and the results demonstrate that the proposed Con-MLPNet achieves state-of-the-art performance. Specifically, on the TLCDD dataset, our method surpasses the competing method across all five evaluation metrics, with the improvements of 3.3% in precision, 2.48% in recall, 3.74% in F-score, 1.76% in accuracy, and 4.0% in IoU over the second-best results.
format Article
id doaj-art-7feaba50c3b642f4b7ee0476dcefde75
institution DOAJ
issn 1939-1404
2151-1535
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
spelling doaj-art-7feaba50c3b642f4b7ee0476dcefde752025-08-20T03:05:42ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-0118195631957610.1109/JSTARS.2025.359347311098898Ground-Based Remote Sensing Cloud Image Segmentation Using Convolution-MLP NetworkShuang Liu0https://orcid.org/0000-0002-9027-0690Jiafeng Zhang1https://orcid.org/0009-0004-9363-8800Zhong Zhang2https://orcid.org/0000-0002-2993-8612Shuzhen Hu3Baihua Xiao4https://orcid.org/0000-0003-3941-1141Tianjin Key Laboratory of Wireless Mobile Communications and Power Transmission, Tianjin Normal University, Tianjin, ChinaTianjin Key Laboratory of Wireless Mobile Communications and Power Transmission, Tianjin Normal University, Tianjin, ChinaTianjin Key Laboratory of Wireless Mobile Communications and Power Transmission, Tianjin Normal University, Tianjin, ChinaMeteorological Observation Center, China Meteorological Administration, Beijing, ChinaState Key Laboratory for Management and Control of Complex Systems, Institute of Automation, Chinese Academy of Sciences, Beijing, ChinaRecently, multilayer perceptron (MLPs)-based methods in computer vision have attracted much attention due to the ability of learning long-range dependencies. However, MLPs-based methods usually treat all the tokens equally, which is difficult to segment challenging cloud regions. In this article, we propose a novel network named convolution-MLP network (Con-MLPNet) for ground-based remote sensing cloud image segmentation, which could effectively learn long-range dependencies via the combination of MLPs and the attention mechanism. To this end, we propose the attention-guided MLPs module to highlight salient features and suppress irrelevant features from the spatial and channel aspects. Meanwhile, different from existing MLPs methods where the long-range dependencies are learned from one single scale, we propose the dilated MLPs (DMLPs) to learn long-range dependencies at different scales by sampling different channels of tokens. Furthermore, we design the parallel dilated MLPs module to integrate multiple DMLPs with different parameters in order to extract multiscale information. We conduct a series of experiments on three public ground-based cloud image segmentation datasets, i.e., TLCDD, SWIMSEG, and TCDD, and the results demonstrate that the proposed Con-MLPNet achieves state-of-the-art performance. Specifically, on the TLCDD dataset, our method surpasses the competing method across all five evaluation metrics, with the improvements of 3.3% in precision, 2.48% in recall, 3.74% in F-score, 1.76% in accuracy, and 4.0% in IoU over the second-best results.https://ieeexplore.ieee.org/document/11098898/Attention-guided MLPs (AGMs)parallel dilated MLPs (PDMs)remote sensing cloud image segmentation
spellingShingle Shuang Liu
Jiafeng Zhang
Zhong Zhang
Shuzhen Hu
Baihua Xiao
Ground-Based Remote Sensing Cloud Image Segmentation Using Convolution-MLP Network
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Attention-guided MLPs (AGMs)
parallel dilated MLPs (PDMs)
remote sensing cloud image segmentation
title Ground-Based Remote Sensing Cloud Image Segmentation Using Convolution-MLP Network
title_full Ground-Based Remote Sensing Cloud Image Segmentation Using Convolution-MLP Network
title_fullStr Ground-Based Remote Sensing Cloud Image Segmentation Using Convolution-MLP Network
title_full_unstemmed Ground-Based Remote Sensing Cloud Image Segmentation Using Convolution-MLP Network
title_short Ground-Based Remote Sensing Cloud Image Segmentation Using Convolution-MLP Network
title_sort ground based remote sensing cloud image segmentation using convolution mlp network
topic Attention-guided MLPs (AGMs)
parallel dilated MLPs (PDMs)
remote sensing cloud image segmentation
url https://ieeexplore.ieee.org/document/11098898/
work_keys_str_mv AT shuangliu groundbasedremotesensingcloudimagesegmentationusingconvolutionmlpnetwork
AT jiafengzhang groundbasedremotesensingcloudimagesegmentationusingconvolutionmlpnetwork
AT zhongzhang groundbasedremotesensingcloudimagesegmentationusingconvolutionmlpnetwork
AT shuzhenhu groundbasedremotesensingcloudimagesegmentationusingconvolutionmlpnetwork
AT baihuaxiao groundbasedremotesensingcloudimagesegmentationusingconvolutionmlpnetwork