Deep Learning Approaches to Image Texture Analysis in Material Processing
Texture analysis is key to better understanding of the relationships between the microstructures of the materials and their properties, as well as the use of models in process systems using raw signals or images as input. Recently, new methods based on transfer learning with deep neural networks hav...
| Main Authors: | , |
|---|---|
| Format: | Journal Article |
| Published: |
MDPI AG
2022
|
| Online Access: | http://purl.org/au-research/grants/arc/CE200100009 http://hdl.handle.net/20.500.11937/87865 |
| _version_ | 1848764942370471936 |
|---|---|
| author | Liu, Xiu Aldrich, Chris |
| author_facet | Liu, Xiu Aldrich, Chris |
| author_sort | Liu, Xiu |
| building | Curtin Institutional Repository |
| collection | Online Access |
| description | Texture analysis is key to better understanding of the relationships between the microstructures of the materials and their properties, as well as the use of models in process systems using raw signals or images as input. Recently, new methods based on transfer learning with deep neural networks have become established as highly competitive approaches to classical texture analysis. In this study, three traditional approaches, based on the use of grey level co-occurrence matrices, local binary patterns and textons are compared with five transfer learning approaches, based on the use of AlexNet, VGG19, ResNet50, GoogLeNet and MobileNetV2. This is done based on two simulated and one real-world case study. In the simulated case studies, material microstructures were simulated with Voronoi graphic representations and in the real-world case study, the appearance of ultrahigh carbon steel is cast as a textural pattern recognition pattern. The ability of random forest models, as well as the convolutional neural networks themselves, to discriminate between different textures with the image features as input was used as the basis for comparison. The texton algorithm performed better than the LBP and GLCM algorithms and similar to the deep learning approaches when these were used directly, without any retraining. Partial or full retraining of the convolutional neural networks yielded considerably better results, with GoogLeNet and MobileNetV2 yielding the best results. |
| first_indexed | 2025-11-14T11:27:22Z |
| format | Journal Article |
| id | curtin-20.500.11937-87865 |
| institution | Curtin University Malaysia |
| institution_category | Local University |
| last_indexed | 2025-11-14T11:27:22Z |
| publishDate | 2022 |
| publisher | MDPI AG |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | curtin-20.500.11937-878652022-03-04T02:16:12Z Deep Learning Approaches to Image Texture Analysis in Material Processing Liu, Xiu Aldrich, Chris Texture analysis is key to better understanding of the relationships between the microstructures of the materials and their properties, as well as the use of models in process systems using raw signals or images as input. Recently, new methods based on transfer learning with deep neural networks have become established as highly competitive approaches to classical texture analysis. In this study, three traditional approaches, based on the use of grey level co-occurrence matrices, local binary patterns and textons are compared with five transfer learning approaches, based on the use of AlexNet, VGG19, ResNet50, GoogLeNet and MobileNetV2. This is done based on two simulated and one real-world case study. In the simulated case studies, material microstructures were simulated with Voronoi graphic representations and in the real-world case study, the appearance of ultrahigh carbon steel is cast as a textural pattern recognition pattern. The ability of random forest models, as well as the convolutional neural networks themselves, to discriminate between different textures with the image features as input was used as the basis for comparison. The texton algorithm performed better than the LBP and GLCM algorithms and similar to the deep learning approaches when these were used directly, without any retraining. Partial or full retraining of the convolutional neural networks yielded considerably better results, with GoogLeNet and MobileNetV2 yielding the best results. 2022 Journal Article http://hdl.handle.net/20.500.11937/87865 http://purl.org/au-research/grants/arc/CE200100009 http://creativecommons.org/licenses/by/4.0/ MDPI AG fulltext |
| spellingShingle | Liu, Xiu Aldrich, Chris Deep Learning Approaches to Image Texture Analysis in Material Processing |
| title | Deep Learning Approaches to Image Texture Analysis in Material Processing |
| title_full | Deep Learning Approaches to Image Texture Analysis in Material Processing |
| title_fullStr | Deep Learning Approaches to Image Texture Analysis in Material Processing |
| title_full_unstemmed | Deep Learning Approaches to Image Texture Analysis in Material Processing |
| title_short | Deep Learning Approaches to Image Texture Analysis in Material Processing |
| title_sort | deep learning approaches to image texture analysis in material processing |
| url | http://purl.org/au-research/grants/arc/CE200100009 http://hdl.handle.net/20.500.11937/87865 |