Introduction of an Algorithm Based on Convolutional Neural Networks for an Automated Online Correction of Braided Cardiovascular Implants

The expense of treating cardiovascular diseases is significant. For example, sediments on the coronary arteries’ inner walls are among the most common risks of a heart attack. One possible treatment includes cardiovascular implants or stents. Stents are manufactured by a braiding process and afterward inspected for defects by human visual inspection. To reduce production costs, an automated inspection system is, therefore, the subject of this work. First, we propose a formalized problem description for camera-based automated visual inspection. Next, a machine learning based divide-and-conquer algorithm is presented. The CNN based algorithm can be used both to supervise the braiding process and to correct braiding errors.