Скачать или смотреть PCB Defect Detection with Deep Learning in LabVIEW

Unsupervised visual anomaly detection demo.
The model is trained only on normal PCB images and automatically identifies missing components, bent pins, extra parts, and other defects with high precision.

Key Features:
1. Unsupervised Learning – No labeled data required.
2. Few-Shot Training – Only a few hundred good samples needed.
3. Fast Adaptation – Train in under a minute for new PCB designs.
4. Real-Time Inference – ~20 ms on RTX 2080 Ti.
5. High Accuracy – Over 95% accuracy across 100+ datasets.
6. Robust – Reliable even with simple webcams and non-ideal lighting.
7. Versatile – Applicable across industries: automotive, pharma, textiles, electronics, food production, and more.

Learn more:
DeepLTK: https://www.ngene.co/deep-learning-to...
UVAD: https://www.ngene.co/uvad-add-on-for-...
CuLab: https://www.ngene.co/gpu-toolkit-for-...

#labview #deeplearning #anomalydetection #unsupervisedlearning #gpu #deepltk #culab