Simulation of microstructures and machine learning
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_ | 1866910805874704384 |
|---|---|
| author | Schladitz, Katja Redenbach, Claudia Barisin, Tin Jung, Christian Jeziorski, Natascha Bosnar, Lovro Fulir, Juraj Gospodnetić, Petra |
| author_facet | Schladitz, Katja Redenbach, Claudia Barisin, Tin Jung, Christian Jeziorski, Natascha Bosnar, Lovro Fulir, Juraj Gospodnetić, Petra |
| contents | Machine learning offers attractive solutions to challenging image processing tasks. Tedious development and parametrization of algorithmic solutions can be replaced by training a convolutional neural network or a random forest with a high potential to generalize. However, machine learning methods rely on huge amounts of representative image data along with a ground truth, usually obtained by manual annotation. Thus, limited availability of training data is a critical bottleneck. We discuss two use cases: optical quality control in industrial production and segmenting crack structures in 3D images of concrete. For optical quality control, all defect types have to be trained but are typically not evenly represented in the training data. Additionally, manual annotation is costly and often inconsistent. It is nearly impossible in the second case: segmentation of crack systems in 3D images of concrete. Synthetic images, generated based on realizations of stochastic geometry models, offer an elegant way out. A wide variety of structure types can be generated. The within structure variation is naturally captured by the stochastic nature of the models and the ground truth is for free. Many new questions arise. In particular, which characteristics of the real image data have to be met to which degree of fidelity. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2501_18313 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Simulation of microstructures and machine learning Schladitz, Katja Redenbach, Claudia Barisin, Tin Jung, Christian Jeziorski, Natascha Bosnar, Lovro Fulir, Juraj Gospodnetić, Petra Computer Vision and Pattern Recognition 60D05 Machine learning offers attractive solutions to challenging image processing tasks. Tedious development and parametrization of algorithmic solutions can be replaced by training a convolutional neural network or a random forest with a high potential to generalize. However, machine learning methods rely on huge amounts of representative image data along with a ground truth, usually obtained by manual annotation. Thus, limited availability of training data is a critical bottleneck. We discuss two use cases: optical quality control in industrial production and segmenting crack structures in 3D images of concrete. For optical quality control, all defect types have to be trained but are typically not evenly represented in the training data. Additionally, manual annotation is costly and often inconsistent. It is nearly impossible in the second case: segmentation of crack systems in 3D images of concrete. Synthetic images, generated based on realizations of stochastic geometry models, offer an elegant way out. A wide variety of structure types can be generated. The within structure variation is naturally captured by the stochastic nature of the models and the ground truth is for free. Many new questions arise. In particular, which characteristics of the real image data have to be met to which degree of fidelity. |
| title | Simulation of microstructures and machine learning |
| topic | Computer Vision and Pattern Recognition 60D05 |
| url | https://arxiv.org/abs/2501.18313 |