Fully-Synthetic Training for Visual Quality Inspection in Automotive Production
Fuente:
arXiv
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916796523610112 |
|---|---|
| author | Huber, Christoph Knoll, Dino Guthe, Michael |
| author_facet | Huber, Christoph Knoll, Dino Guthe, Michael |
| contents | Visual Quality Inspection plays a crucial role in modern manufacturing environments as it ensures customer safety and satisfaction. The introduction of Computer Vision (CV) has revolutionized visual quality inspection by improving the accuracy and efficiency of defect detection. However, traditional CV models heavily rely on extensive datasets for training, which can be costly, time-consuming, and error-prone. To overcome these challenges, synthetic images have emerged as a promising alternative. They offer a cost-effective solution with automatically generated labels. In this paper, we propose a pipeline for generating synthetic images using domain randomization. We evaluate our approach in three real inspection scenarios and demonstrate that an object detection model trained solely on synthetic data can outperform models trained on real images. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_09354 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Fully-Synthetic Training for Visual Quality Inspection in Automotive Production Huber, Christoph Knoll, Dino Guthe, Michael Computer Vision and Pattern Recognition Visual Quality Inspection plays a crucial role in modern manufacturing environments as it ensures customer safety and satisfaction. The introduction of Computer Vision (CV) has revolutionized visual quality inspection by improving the accuracy and efficiency of defect detection. However, traditional CV models heavily rely on extensive datasets for training, which can be costly, time-consuming, and error-prone. To overcome these challenges, synthetic images have emerged as a promising alternative. They offer a cost-effective solution with automatically generated labels. In this paper, we propose a pipeline for generating synthetic images using domain randomization. We evaluate our approach in three real inspection scenarios and demonstrate that an object detection model trained solely on synthetic data can outperform models trained on real images. |
| title | Fully-Synthetic Training for Visual Quality Inspection in Automotive Production |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2503.09354 |