An experimental investigation of MobileNet SSD on fabric defect detection
Fabric defect detection plays a crucial role in ensuring product quality in textile manufacturing. This study presents an investigation into improving fabric defect detection using MobileNetV2 SSD as the baseline model and exploring various existing methods to further enhance performance. The pro...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project (FYP) |
| Language: | English |
| Published: |
Nanyang Technological University
2024
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/175198 |
| _version_ | 1824454220882903040 |
|---|---|
| author | Zhou, Wei |
| author2 | Zheng Jianmin |
| author_facet | Zheng Jianmin Zhou, Wei |
| author_sort | Zhou, Wei |
| collection | NTU |
| description | Fabric defect detection plays a crucial role in ensuring product quality in textile
manufacturing. This study presents an investigation into improving fabric defect detection
using MobileNetV2 SSD as the baseline model and exploring various existing methods to
further enhance performance. The project aims to improve detection accuracy without
sacrificing speed and compares the effectiveness of focal loss, lightweight backbones, feature
fusion methods, feature pyramid networks, and transfer learning. Extensive experimentation
and evaluation were conducted using datasets from Roboflow on Google Colab and Kaggle.
Experimental results showed that while focal loss improves accuracy, it also led to the
problem of generating multiple bounding boxes for the same object and therefore not being
used. Meanwhile, the concatenation module, SE ResNeXt50, and BiFPN were the top
performers in their category. Transfer learning in smaller datasets yielded more significant
improvements compared to larger datasets. MobileNetV2 BiFPN emerged as the best model
that maintained fast speed while achieving high accuracy. The comprehensive insights
garnered from this study can provide valuable guidance for future research in surface defect
detection. Recommendations include building a custom dataset, conducting a deeper analysis
of existing methods, and integrating a module for automatically detecting the optimal scale
and aspect ratio. |
| first_indexed | 2025-02-19T03:18:51Z |
| format | Final Year Project (FYP) |
| id | ntu-10356/175198 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2025-02-19T03:18:51Z |
| publishDate | 2024 |
| publisher | Nanyang Technological University |
| record_format | dspace |
| spelling | ntu-10356/1751982024-04-26T15:41:34Z An experimental investigation of MobileNet SSD on fabric defect detection Zhou, Wei Zheng Jianmin School of Computer Science and Engineering ASJMZheng@ntu.edu.sg Computer and Information Science Defect detection Fabric defect detection plays a crucial role in ensuring product quality in textile manufacturing. This study presents an investigation into improving fabric defect detection using MobileNetV2 SSD as the baseline model and exploring various existing methods to further enhance performance. The project aims to improve detection accuracy without sacrificing speed and compares the effectiveness of focal loss, lightweight backbones, feature fusion methods, feature pyramid networks, and transfer learning. Extensive experimentation and evaluation were conducted using datasets from Roboflow on Google Colab and Kaggle. Experimental results showed that while focal loss improves accuracy, it also led to the problem of generating multiple bounding boxes for the same object and therefore not being used. Meanwhile, the concatenation module, SE ResNeXt50, and BiFPN were the top performers in their category. Transfer learning in smaller datasets yielded more significant improvements compared to larger datasets. MobileNetV2 BiFPN emerged as the best model that maintained fast speed while achieving high accuracy. The comprehensive insights garnered from this study can provide valuable guidance for future research in surface defect detection. Recommendations include building a custom dataset, conducting a deeper analysis of existing methods, and integrating a module for automatically detecting the optimal scale and aspect ratio. Bachelor's degree 2024-04-21T23:39:55Z 2024-04-21T23:39:55Z 2024 Final Year Project (FYP) Zhou, W. (2024). An experimental investigation of MobileNet SSD on fabric defect detection. Final Year Project (FYP), Nanyang Technological University, Singapore. https://hdl.handle.net/10356/175198 https://hdl.handle.net/10356/175198 en SCSE23-0049 application/pdf Nanyang Technological University |
| spellingShingle | Computer and Information Science Defect detection Zhou, Wei An experimental investigation of MobileNet SSD on fabric defect detection |
| title | An experimental investigation of MobileNet SSD on fabric defect detection |
| title_full | An experimental investigation of MobileNet SSD on fabric defect detection |
| title_fullStr | An experimental investigation of MobileNet SSD on fabric defect detection |
| title_full_unstemmed | An experimental investigation of MobileNet SSD on fabric defect detection |
| title_short | An experimental investigation of MobileNet SSD on fabric defect detection |
| title_sort | experimental investigation of mobilenet ssd on fabric defect detection |
| topic | Computer and Information Science Defect detection |
| url | https://hdl.handle.net/10356/175198 |
| work_keys_str_mv | AT zhouwei anexperimentalinvestigationofmobilenetssdonfabricdefectdetection AT zhouwei experimentalinvestigationofmobilenetssdonfabricdefectdetection |