Client Requirement

Client was looking for a solution to achieve automated QC on their manufacturing facility. Client has defined few processes to compare and make decision. The result will be stored in manager panel and notification will be shared to QC manager to review and decide manually about the product quality.

MTU needed a solution using Machine Learning – Computer vision to analyse and compare real part with its pre trained machine learning algorithm deployed on iPhone / iPad devices. The light weight model uses OpenCV to firstly detect which part the operators requires to detect. Once, detected it will move further on processing and once processed it will show result on panel says if the part is QC pass or failed.


  • A backend to add SKUs with their Images and videos for processing.
  • Add parts images for OpenCV image comparison.
  • Computer vision algorithm to QC.
  • Store results.
  • Pass results to QC manager for further processing.

Technologies Used

Computer vision
Fast API
aws-s3 logo