Project with HMS
| Title | Project with HMS |
|---|---|
| Summary | Few-shot Learning for Quality Inspection |
| Keywords | |
| TimeFrame | Fall 2022 |
| References | |
| Prerequisites | |
| Author | |
| Supervisor | Peyman Mashhadi, Yuantao Fan |
| Level | Master |
| Status | Draft |
There is an increasing interest in intelligent applications for industrial use cases. One area where smart AI-driven applications are particularly interesting is visual inspection of products along the production line. Traditional computer-vision-based approaches rely on large amounts of data to make accurate predictions. Acquiring enough data to achieve a satisfying level of accuracy is often challenging. The aim of this thesis is to develop a tool for quality inspection based on few-shot learning. Few-shot learning refers to the practice of feeding a learning model with a small amount of training data. The goal is to detect anomalies in mounted components on circuit boards. The system should be able to differentiate between defective and functioning boards.
