Using poll sheets and computer vision as an inexpensive alternative to clickers

Classroom Response Systems, colloquially known as Clickers after the predominant hand-held input device, are widely used because they allow students to respond in class to questions posed by the lecturer. This improves active learning and interaction in large classes---students are more involved, and lecturers can assess understanding and even take remedial action. Unfortunately, Clicker systems are relatively expensive, particularly in a developing-world context. They typically cost $200--700 for a base station and $30--50 per Clicker. In this paper we present an inexpensive alternative to Clickers. Poll sheets with coloured blocks printed on a white background are held up by the students and a camera-phone is used in panoramic mode to photograph the class. This image is then processed using computer vision to count and classify the students' responses. While the 85% recognition rate we achieve is certainly not as accurate as Clickers, this approach nevertheless has many of the same benefits for active learning at a fraction of the cost: $0.20 per poll sheet, assuming a laptop and camera-phone are already available.