Artificial Intelligence-Based Classification of Human Smiles


Executive Summary


Smile classification and determining the relationship of smiles to other data such as another person’s reactions or an event occurring could be used to train any kind of customer-facing personnel such as front desk employees, trainers, and sales people. Currently no system exists that can both automatically detect smiles and classify individual smile episodes.


Description of Technology


This technology takes raw index outputs from the Google Joy Detection algorithm (or any similar algorithm with image processing capabilities) and produces a characterization of smiles in the image. It can determine the relationship between smiles from two or more individuals to determine the “emotional leader” and contagious effects.


Key Benefits

  • New data type This technology provides data that no other technology provides that would be useful for a wide range of applications.
  • Simple system Only indices of the measured data leave the system. Processing of these indices to produce output data occurs on the cloud.
  • Flexible use The physical box is quite small and can be used in a wide variety of scenarios.



  • Sales and marketing training
  • Front desk and receptionist training
  • Training for instructors
  • Evaluation of focus groups
  • Over-the-counter in shops and hospitality business
  • Evaluation of screen testing and acting
  • Evaluation of video resumes
  • Evaluate of advertisement campaigns
  • Monitoring elderly
  • Monitoring of kids to detect early onset of autism
  • Employee moods in meetings
  • Evaluating the effects of political speeches and polarization


IP Status 


Software copyright


Licensing Rights Available


Full rights available




Subir Biswas


Tech ID




Patent Information:


For Information, Contact:

Raymond Devito
Technology Manager
Michigan State University
Subir Biswas