Intelligent Inventory Management
The client receives large shipments of spare parts for its workshop that repairs big machinery. Reading and capturing the serial number of the newly received spare parts in IT systems and comparing them with the invoices is a long and cumbersome process.
Development of a model that can read the serial numbers out of photos of the spare parts demands allocation of huge amount of storage and computation power. It is important to consider that even with the allocation of the resources, it is not guaranteed that the model can achive acceptable accuracy.
KPMG Lighthouse has proposed to use the Cloud Vision API from the Google Cloud to capture the text from the photos of the spare parts. Furthermore, using the Cloud's AutoML, one can easily train a model on top of the Cloud Vision API to improve the accuracy. The resources required for the AutoML are considerably lighter and cheaper compared to an inhouse developed model. Moreover, KPMG Lighthouse has developed a mobile application that enables the inventory workers to capture the photo, edit and verify the captured serial number, and send it to the inventory management systems.