Neural Network Pruning for Lightweight Metal Corrosion Image Segmentation Models

Metal corrosion detection is essential for ensuring structural safety and minimizing economic losses. While deep learning (DL)-based image segmentation has improved corrosion detection accuracy and efficiency, its high computational demands hinder deployment on resource-constrained edge devices. Thi...

Full description

Saved in:
Bibliographic Details
Main Authors: Vincent F. Yu, Gemilang Santiyuda, Shih-Wei Lin, Udjianna S. Pasaribu, Yuli Sri Afrianti
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10969780/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849310482911985664
author Vincent F. Yu
Gemilang Santiyuda
Shih-Wei Lin
Udjianna S. Pasaribu
Yuli Sri Afrianti
author_facet Vincent F. Yu
Gemilang Santiyuda
Shih-Wei Lin
Udjianna S. Pasaribu
Yuli Sri Afrianti
author_sort Vincent F. Yu
collection DOAJ
description Metal corrosion detection is essential for ensuring structural safety and minimizing economic losses. While deep learning (DL)-based image segmentation has improved corrosion detection accuracy and efficiency, its high computational demands hinder deployment on resource-constrained edge devices. This study investigates lightweight DL models for corrosion segmentation by applying structured pruning to reduce computational costs while maintaining accuracy. We evaluate five segmentation architectures (U-Net, U-Net++, FPN, LinkNet, and MA-Net) and three pruning strategies (linear, automated gradual pruning, and movement pruning) on two corrosion image datasets (NEA and SSCS). Detailed trade-off analysis between model size, computational cost (MAC), and segmentation performance (IoU) reveals that pruning up to 90% sparsity leads to a <inline-formula> <tex-math notation="LaTeX">$\leq 10\%$ </tex-math></inline-formula> IoU drop on SSCS and <inline-formula> <tex-math notation="LaTeX">$\leq 5\%$ </tex-math></inline-formula> on NEA, demonstrating that significant compression is possible with minimal accuracy loss. However, some architectures (e.g., LinkNet) and pruning strategies (e.g., movement pruning) show significant performance deterioration, suggesting that pruning effectiveness varies across models. These findings provide insights into optimizing corrosion segmentation models for efficient deployment on edge devices, balancing accuracy and resource constraints.
format Article
id doaj-art-1ae929ebcc2445eca286206d823b6dc7
institution Kabale University
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-1ae929ebcc2445eca286206d823b6dc72025-08-20T03:53:42ZengIEEEIEEE Access2169-35362025-01-0113716737168710.1109/ACCESS.2025.356243510969780Neural Network Pruning for Lightweight Metal Corrosion Image Segmentation ModelsVincent F. Yu0https://orcid.org/0000-0001-8975-0606Gemilang Santiyuda1https://orcid.org/0000-0002-4432-6059Shih-Wei Lin2https://orcid.org/0000-0003-1343-0838Udjianna S. Pasaribu3Yuli Sri Afrianti4Department of Industrial Management, National Taiwan University of Science and Technology, Taipei, TaiwanDepartment of Industrial Management, National Taiwan University of Science and Technology, Taipei, TaiwanDepartment of Information Management, Chang Gung University, Taoyuan, TaiwanStatistics Research Group, Faculty of Mathematics and Natural Sciences, Bandung Institute of Technology, Bandung, IndonesiaStatistics Research Group, Faculty of Mathematics and Natural Sciences, Bandung Institute of Technology, Bandung, IndonesiaMetal corrosion detection is essential for ensuring structural safety and minimizing economic losses. While deep learning (DL)-based image segmentation has improved corrosion detection accuracy and efficiency, its high computational demands hinder deployment on resource-constrained edge devices. This study investigates lightweight DL models for corrosion segmentation by applying structured pruning to reduce computational costs while maintaining accuracy. We evaluate five segmentation architectures (U-Net, U-Net++, FPN, LinkNet, and MA-Net) and three pruning strategies (linear, automated gradual pruning, and movement pruning) on two corrosion image datasets (NEA and SSCS). Detailed trade-off analysis between model size, computational cost (MAC), and segmentation performance (IoU) reveals that pruning up to 90% sparsity leads to a <inline-formula> <tex-math notation="LaTeX">$\leq 10\%$ </tex-math></inline-formula> IoU drop on SSCS and <inline-formula> <tex-math notation="LaTeX">$\leq 5\%$ </tex-math></inline-formula> on NEA, demonstrating that significant compression is possible with minimal accuracy loss. However, some architectures (e.g., LinkNet) and pruning strategies (e.g., movement pruning) show significant performance deterioration, suggesting that pruning effectiveness varies across models. These findings provide insights into optimizing corrosion segmentation models for efficient deployment on edge devices, balancing accuracy and resource constraints.https://ieeexplore.ieee.org/document/10969780/Computer visiondeep learningimage segmentationmetal corrosionpruning
spellingShingle Vincent F. Yu
Gemilang Santiyuda
Shih-Wei Lin
Udjianna S. Pasaribu
Yuli Sri Afrianti
Neural Network Pruning for Lightweight Metal Corrosion Image Segmentation Models
IEEE Access
Computer vision
deep learning
image segmentation
metal corrosion
pruning
title Neural Network Pruning for Lightweight Metal Corrosion Image Segmentation Models
title_full Neural Network Pruning for Lightweight Metal Corrosion Image Segmentation Models
title_fullStr Neural Network Pruning for Lightweight Metal Corrosion Image Segmentation Models
title_full_unstemmed Neural Network Pruning for Lightweight Metal Corrosion Image Segmentation Models
title_short Neural Network Pruning for Lightweight Metal Corrosion Image Segmentation Models
title_sort neural network pruning for lightweight metal corrosion image segmentation models
topic Computer vision
deep learning
image segmentation
metal corrosion
pruning
url https://ieeexplore.ieee.org/document/10969780/
work_keys_str_mv AT vincentfyu neuralnetworkpruningforlightweightmetalcorrosionimagesegmentationmodels
AT gemilangsantiyuda neuralnetworkpruningforlightweightmetalcorrosionimagesegmentationmodels
AT shihweilin neuralnetworkpruningforlightweightmetalcorrosionimagesegmentationmodels
AT udjiannaspasaribu neuralnetworkpruningforlightweightmetalcorrosionimagesegmentationmodels
AT yulisriafrianti neuralnetworkpruningforlightweightmetalcorrosionimagesegmentationmodels