Machine Vision System for Quality Control of Stents Used in Angioplasty

Traditional visual inspection – compared to automated inspection – is not very effective, but for various reasons it is still used in many industries, including medical one. This paper discusses an example of visual inspection in the manufacture of stents used in angioplasty. The use of such inspection is economically justified here due to the low probability of non-conformity, even with the associated high risk of error, which can result in loss of patient health. This paper presents the concept of automating stent quality assessment through the implementation of a Machine Vision System. To support operator decision making in the proposed system, a decision model using artificial neural networks has been developed. The paper demonstrates the potential of such an implementation to minimize multiple sources of weakness in the visual inspection used so far. The paper presents the verification results and evaluation of the feasibility of using a vision-based system for quality control of angioplasty stents.