Drexel University

College of Computing and Informatics


Automated sports card condition grading using computer vision and machine learning.
The collectible trading card market has increased exponentially since the pandemic. The small number of industry-recognized grading firms use only a handful of people to manually grade each card. The system is prone to human error and turnaround time can reach close to 1 year. Recently some card graders have stopped accepting submissions due to their excessive backlog. Our team of CCI students aim to use computer vision and machine learning to grade trading cards based on centering, surface condition, edge crispness, corner wear and other card conditions. Once cards are graded, a web page will be automatically generated showing high resolution images of both the front and back of the card, scoring details for each category, and the ability to overlay measurements and grading calculations on top of the clean card image. 

Team Members