MIT Mood Meter

 

 

To celebrate MIT's 150th anniversary, we designed a computer vision system, MIT Mood Meter, that counted smiles of people in four different public spaces (i.e., Media Lab, Stata Center, Student Center, Infinite corridor) of MIT through cameras. By sensing data in real time; we converted the map of MIT as a heat map of smiles and refreshed it in every 2 seconds on the World Wide Web.

 

With data collected for more than 10 weeks using computer vision techniques, we were able to objectively answer questions such as whether students from one department smile more than another, or do people smile less during midterms? This was one of the pioneering examples of testing computer vision algorithms through a real world deployment, and collect data to draw new insights.

Part of the software that I have written for Mood Meter has been sub-licensed for internal Research and Development by P&G, Shell, Intuit, State Farm, Fox News, and Hisense towards objectively understanding customer behaviors for marketing and to measure customer satisfaction.

Here is the video of the four different mood meters on campus.

 



Publications:

M. E. Hoque*, J. Hernandez*, W. Drevo, R. W. Picard, Mood Meter: Counting smiles in the Wild,  To appear in the 14th International Conference on Ubiquitous Computing (Ubicomp), September 2012. (*equal contribution)

M. E. Hoque*, J. Hernandez*, R. Picard, Mood Meter: Large-Scale and Long-Term Smile Monitoring System, ACM SIGGRAPH Emerging Technologies, August 2012. (*equal contribution) (PDF: 239 KB)

Project website: http://moodmeter.media.mit.edu/