Real-Time Audience Polling Using Computer Vision


March 11, 2014


How can teachers, executives, and other public speakers make an interactive presentation to a large audience? Up until now, techniques for polling the audience relied on personal electronic devices, such as smart phones or special purpose “clickers”. To enable real-time polling of any audience, we introduce a new technique using computer vision. Each member of the audience is given a qCard: an ordinary sheet of paper with a printed barcode. The speaker asks a multiple-choice question, and audience members respond by holding their qCard in different ways. Using a laptop and digital camera, our software automatically recognizes and aggregates the responses. In this talk, we will describe our experience piloting this technology in Bangalore schools. We will also conduct a live demonstration!


Bill Thies

Bill Thies is a researcher at Microsoft Research India, where he is a member of the Technologies for Emerging Markets Group. In addition to programming languages for biology, his research focuses on inventing and applying technologies that promote socio-economic development, especially in the context of healthcare and education. Bill received all of his degrees from the Massachusetts Institute of Technology, where he finished his Ph.D. in February, 2009. His Ph.D. thesis focused on StreamIt, a programming language and compiler for mapping streaming applications to multicore architectures.