Automatic visual inspection with neural networks using the example of cardiovascular implants (stents)

Environment

"Innovations for tomorrow's production, services and work" aims to improve production in medical technology. improve. The aim of the bachelor/master thesis is to a self-learning process be which automatically detects errors in the braiding pattern of stents stents and, based on this, optimizes the adaptation parameters based on this.

Task

• Image preprocessing and manipulation (defects in the braiding pattern) of stents
• Application and parameterization of suitable frameworks or algorithms for image processing
• Detection of braiding defects by Machine Learning and Deep Learning methods
• Evaluation and benchmarking investigated algorithms for the detection and classification of Stents

Requirements

• Programming skills
• Image processing
• Machine learning skills