Computer vision applied to analysis of learning interactions

Professor Yu Xiao’s group in Aalto ELEC received seed funding from HIIT to initiate multi-disciplinary research collaboration with professor Kristiina Kumpulainen’s group at the Faculty of Educational Sciences of the University of Helsinki. The objective is to develop tools for analyzing student-student and student-supervisor interactions on videos captured in classrooms. An approach based on using a deep neural network for pose recognition was presented in the CICERO workshop on digitalization and artificial intelligence.

The convolutional neural network (CNN) can extract skeleton key-points from a video and this information can be used to analyze learning interactions.