An Improved U-Net Image Segmentation Method and Its Application for Metallic Grain Size Statistics
Grain size is one of the most important parameters for metallographic microstructure analysis, which can partly determine the material performance. The measurement of grain size is based on accurate image segmentation methods, which include traditional image processing methods and emerging machine-l...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2022-06-01
|
Series: | Materials |
Subjects: | |
Online Access: | https://www.mdpi.com/1996-1944/15/13/4417 |
_version_ | 1797442970533756928 |
---|---|
author | Peng Shi Mengmeng Duan Lifang Yang Wei Feng Lianhong Ding Liwu Jiang |
author_facet | Peng Shi Mengmeng Duan Lifang Yang Wei Feng Lianhong Ding Liwu Jiang |
author_sort | Peng Shi |
collection | DOAJ |
description | Grain size is one of the most important parameters for metallographic microstructure analysis, which can partly determine the material performance. The measurement of grain size is based on accurate image segmentation methods, which include traditional image processing methods and emerging machine-learning-based methods. Unfortunately, traditional image processing methods can hardly segment grains correctly from metallographic images with low contrast and blurry boundaries. Moreover, the proposed machine-learning-based methods need a large dataset to train the model and can hardly deal with the segmentation challenge of complex images with fuzzy boundaries and complex structure. In this paper, an improved U-Net model is proposed to automatically accomplish image segmentation of complex metallographic images with only a small training set. The experiments on metallographic images show the significant advantage of the method, especially for the metallographic images with low contrast, a fuzzy boundary and complex structure. Compared with other deep learning methods, the improved U-Net scored higher in ACC, MIoU, Precision, and F1 indexes, among which ACC was 0.97, MIoU was 0.752, Precision was 0.98, and F1 was 0.96. The grain size was calculated based on the segmentation according to the American Society for Testing Material (ASTM) standards, producing a satisfactory result. |
first_indexed | 2024-03-09T12:49:21Z |
format | Article |
id | doaj.art-baf1e647f4b4415b8cf6afe0f162591d |
institution | Directory Open Access Journal |
issn | 1996-1944 |
language | English |
last_indexed | 2024-03-09T12:49:21Z |
publishDate | 2022-06-01 |
publisher | MDPI AG |
record_format | Article |
series | Materials |
spelling | doaj.art-baf1e647f4b4415b8cf6afe0f162591d2023-11-30T22:08:39ZengMDPI AGMaterials1996-19442022-06-011513441710.3390/ma15134417An Improved U-Net Image Segmentation Method and Its Application for Metallic Grain Size StatisticsPeng Shi0Mengmeng Duan1Lifang Yang2Wei Feng3Lianhong Ding4Liwu Jiang5National Center for Materials Service Safety, University of Science and Technology Beijing, Beijing 100083, ChinaNational Center for Materials Service Safety, University of Science and Technology Beijing, Beijing 100083, ChinaNational Center for Materials Service Safety, University of Science and Technology Beijing, Beijing 100083, ChinaSchool of Automation, Nanjing University of Science and Technology, Nanjing 211103, ChinaSchool of Information, Beijing Wuzi University, Beijing 101149, ChinaNational Center for Materials Service Safety, University of Science and Technology Beijing, Beijing 100083, ChinaGrain size is one of the most important parameters for metallographic microstructure analysis, which can partly determine the material performance. The measurement of grain size is based on accurate image segmentation methods, which include traditional image processing methods and emerging machine-learning-based methods. Unfortunately, traditional image processing methods can hardly segment grains correctly from metallographic images with low contrast and blurry boundaries. Moreover, the proposed machine-learning-based methods need a large dataset to train the model and can hardly deal with the segmentation challenge of complex images with fuzzy boundaries and complex structure. In this paper, an improved U-Net model is proposed to automatically accomplish image segmentation of complex metallographic images with only a small training set. The experiments on metallographic images show the significant advantage of the method, especially for the metallographic images with low contrast, a fuzzy boundary and complex structure. Compared with other deep learning methods, the improved U-Net scored higher in ACC, MIoU, Precision, and F1 indexes, among which ACC was 0.97, MIoU was 0.752, Precision was 0.98, and F1 was 0.96. The grain size was calculated based on the segmentation according to the American Society for Testing Material (ASTM) standards, producing a satisfactory result.https://www.mdpi.com/1996-1944/15/13/4417complex imagegrain sizeimage segmentationimproved U-Netmetallographic microstructure analysis |
spellingShingle | Peng Shi Mengmeng Duan Lifang Yang Wei Feng Lianhong Ding Liwu Jiang An Improved U-Net Image Segmentation Method and Its Application for Metallic Grain Size Statistics Materials complex image grain size image segmentation improved U-Net metallographic microstructure analysis |
title | An Improved U-Net Image Segmentation Method and Its Application for Metallic Grain Size Statistics |
title_full | An Improved U-Net Image Segmentation Method and Its Application for Metallic Grain Size Statistics |
title_fullStr | An Improved U-Net Image Segmentation Method and Its Application for Metallic Grain Size Statistics |
title_full_unstemmed | An Improved U-Net Image Segmentation Method and Its Application for Metallic Grain Size Statistics |
title_short | An Improved U-Net Image Segmentation Method and Its Application for Metallic Grain Size Statistics |
title_sort | improved u net image segmentation method and its application for metallic grain size statistics |
topic | complex image grain size image segmentation improved U-Net metallographic microstructure analysis |
url | https://www.mdpi.com/1996-1944/15/13/4417 |
work_keys_str_mv | AT pengshi animprovedunetimagesegmentationmethodanditsapplicationformetallicgrainsizestatistics AT mengmengduan animprovedunetimagesegmentationmethodanditsapplicationformetallicgrainsizestatistics AT lifangyang animprovedunetimagesegmentationmethodanditsapplicationformetallicgrainsizestatistics AT weifeng animprovedunetimagesegmentationmethodanditsapplicationformetallicgrainsizestatistics AT lianhongding animprovedunetimagesegmentationmethodanditsapplicationformetallicgrainsizestatistics AT liwujiang animprovedunetimagesegmentationmethodanditsapplicationformetallicgrainsizestatistics AT pengshi improvedunetimagesegmentationmethodanditsapplicationformetallicgrainsizestatistics AT mengmengduan improvedunetimagesegmentationmethodanditsapplicationformetallicgrainsizestatistics AT lifangyang improvedunetimagesegmentationmethodanditsapplicationformetallicgrainsizestatistics AT weifeng improvedunetimagesegmentationmethodanditsapplicationformetallicgrainsizestatistics AT lianhongding improvedunetimagesegmentationmethodanditsapplicationformetallicgrainsizestatistics AT liwujiang improvedunetimagesegmentationmethodanditsapplicationformetallicgrainsizestatistics |