Cloud-based ML Technologies for Visual Inspection: A case study in manufacturing

Abstract

In recent years, cloud-based Machine Learning services have received much attention for promising fast and cost-effective deployment. At the same time, manufacturing companies are beginning to evaluate and implement these new technologies in their production processes. This paper adopts the design science research approach to demonstrate the use of cloud-based Machine Learning services to implement a visual inspection system in the manufacturing industry. As a result, our developed IT artifact can correctly classify all of the given parts in a dataset consisting of 363 images, outperforming the current manual inspection. Thereby, it addresses the various challenges faced by the industry when introducing cloud-based Machine Learning technologies, evaluating return on investment (ROI), and how this can facilitate further digital transformation in production.