%2520(1920%2520x%25201080%2520px).png)
Want to bring AI-driven visual inspection into your production line? Need to understand how to turn raw image data into reliable, automated quality-control systems? Join our hands-on session and learn how to apply state-of-the-art Computer Vision techniques to real industrial use cases.
Details
Building effective computer-vision pipelines for manufacturing requires much more than knowing how to train a model. It demands a solid understanding of data quality, annotation strategies, model selection, evaluation, and deployment in real-world industrial environments.
This workshop offers a practical introduction to Computer Vision for industry, guiding participants through the full workflow — from raw images to trained models — using a real Bosch-inspired bolts-and-nuts inspection scenario.
What you'll get
Over the course of this interactive session, you will:
- Explore the landscape of modern Computer Vision: classification, detection, segmentation, anomaly detection.
- Learn how to prepare high-quality datasets: collection, labeling, augmentation, and visual exploration.
- Work hands-on with pre-loaded datasets and notebooks.
- Train and compare different AI models, from custom TensorFlow CNNs to Transfer Learning architectures (ResNet) and YOLOv8.
- Understand how to evaluate metrics such as accuracy, precision, recall, F1-score, and learning curves.
- Learn the fundamentals of the data-to-deployment workflow: ETL, EDA, training, optimization, testing, and inference.
- Experiment with real industrial-quality datasets and run your own models on bolts, nuts, screw bodies, and screw heads.
- Discover good practices for data management, model versioning, and reproducible pipelines.
