LarGAN: A Label Auto-Rescaling Generation Adversarial Network for Rare Surface Defects
Insufficient defect data significantly limits detection accuracy in continuous casting slab production. This limitation arises from the data collection in fast-paced production environments. To address this issue, we propose LarGAN, a data augmentation approach that synthesizes similar and high-qual...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-05-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/25/10/2958 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849719604388036608 |
|---|---|
| author | Guan Qin Hanxin Zhang Ke Xu Liaoting Pan Lei Huang Xuezhong Huang Yi Wei |
| author_facet | Guan Qin Hanxin Zhang Ke Xu Liaoting Pan Lei Huang Xuezhong Huang Yi Wei |
| author_sort | Guan Qin |
| collection | DOAJ |
| description | Insufficient defect data significantly limits detection accuracy in continuous casting slab production. This limitation arises from the data collection in fast-paced production environments. To address this issue, we propose LarGAN, a data augmentation approach that synthesizes similar and high-quality defect data from a single image. We utilize a progressive GAN framework to ensure a smooth and stable generation process, starting from low-resolution image synthesis and gradually increasing the network depth. We designed a Label Auto-Rescaling strategy to better adapt to defect data with annotation, enhancing both the quality and morphological diversity of the synthesized defects. To validate the generation results, we evaluate not only standard metrics, such as FID, SSIM, and LPIPS, but also performance, through the downstream detection model YOLOv8. Our experimental results demonstrate that the LarGAN model surpasses other single-image generation models in terms of image quality and diversity. Furthermore, the experiments reveal that the data generated by LarGAN effectively enhances the feature space of the original dataset, thereby improving the accuracy and generalization performance of the detection model. |
| format | Article |
| id | doaj-art-2b21c04ca44c45d7ae1111dadf0c75af |
| institution | DOAJ |
| issn | 1424-8220 |
| language | English |
| publishDate | 2025-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj-art-2b21c04ca44c45d7ae1111dadf0c75af2025-08-20T03:12:07ZengMDPI AGSensors1424-82202025-05-012510295810.3390/s25102958LarGAN: A Label Auto-Rescaling Generation Adversarial Network for Rare Surface DefectsGuan Qin0Hanxin Zhang1Ke Xu2Liaoting Pan3Lei Huang4Xuezhong Huang5Yi Wei6Collaborative Innovation Center of Steel Technology, University of Science and Technology Beijing, Beijing 100083, ChinaCollaborative Innovation Center of Steel Technology, University of Science and Technology Beijing, Beijing 100083, ChinaCollaborative Innovation Center of Steel Technology, University of Science and Technology Beijing, Beijing 100083, ChinaChina Guangxi Beigang New Materials Co., Ltd., Beihai 536000, ChinaChina Guangxi Beigang New Materials Co., Ltd., Beihai 536000, ChinaChina Guangxi Beigang New Materials Co., Ltd., Beihai 536000, ChinaInstitute of Novel Functional Materials, Guangxi Institute of Industry and Research, Nanning 530233, ChinaInsufficient defect data significantly limits detection accuracy in continuous casting slab production. This limitation arises from the data collection in fast-paced production environments. To address this issue, we propose LarGAN, a data augmentation approach that synthesizes similar and high-quality defect data from a single image. We utilize a progressive GAN framework to ensure a smooth and stable generation process, starting from low-resolution image synthesis and gradually increasing the network depth. We designed a Label Auto-Rescaling strategy to better adapt to defect data with annotation, enhancing both the quality and morphological diversity of the synthesized defects. To validate the generation results, we evaluate not only standard metrics, such as FID, SSIM, and LPIPS, but also performance, through the downstream detection model YOLOv8. Our experimental results demonstrate that the LarGAN model surpasses other single-image generation models in terms of image quality and diversity. Furthermore, the experiments reveal that the data generated by LarGAN effectively enhances the feature space of the original dataset, thereby improving the accuracy and generalization performance of the detection model.https://www.mdpi.com/1424-8220/25/10/2958generation adversarial network (GAN)image generationdata augmentationsurface defectscasting slabs |
| spellingShingle | Guan Qin Hanxin Zhang Ke Xu Liaoting Pan Lei Huang Xuezhong Huang Yi Wei LarGAN: A Label Auto-Rescaling Generation Adversarial Network for Rare Surface Defects Sensors generation adversarial network (GAN) image generation data augmentation surface defects casting slabs |
| title | LarGAN: A Label Auto-Rescaling Generation Adversarial Network for Rare Surface Defects |
| title_full | LarGAN: A Label Auto-Rescaling Generation Adversarial Network for Rare Surface Defects |
| title_fullStr | LarGAN: A Label Auto-Rescaling Generation Adversarial Network for Rare Surface Defects |
| title_full_unstemmed | LarGAN: A Label Auto-Rescaling Generation Adversarial Network for Rare Surface Defects |
| title_short | LarGAN: A Label Auto-Rescaling Generation Adversarial Network for Rare Surface Defects |
| title_sort | largan a label auto rescaling generation adversarial network for rare surface defects |
| topic | generation adversarial network (GAN) image generation data augmentation surface defects casting slabs |
| url | https://www.mdpi.com/1424-8220/25/10/2958 |
| work_keys_str_mv | AT guanqin larganalabelautorescalinggenerationadversarialnetworkforraresurfacedefects AT hanxinzhang larganalabelautorescalinggenerationadversarialnetworkforraresurfacedefects AT kexu larganalabelautorescalinggenerationadversarialnetworkforraresurfacedefects AT liaotingpan larganalabelautorescalinggenerationadversarialnetworkforraresurfacedefects AT leihuang larganalabelautorescalinggenerationadversarialnetworkforraresurfacedefects AT xuezhonghuang larganalabelautorescalinggenerationadversarialnetworkforraresurfacedefects AT yiwei larganalabelautorescalinggenerationadversarialnetworkforraresurfacedefects |