Table2Table: Merging “Similar” Workspaces and Supporting Adaptive Telepresence Demonstration Guidance

In a typical tele-presence system, the remote participant is “trans-ported” into the local space to be shared. However, such a set-up limits the remote participant in terms of physical interaction and demonstration with the objects in the local space. One alternative is to merge the two spaces, such that the student in one space can closely watch and mimic the actions of the instructor in its space - as often occurs in science laboratories or cooking classes. As the local set-ups in the two spaces may not be exactly the same, the student has to “adapt” to the situation and identify the right objects in mimicking the instructor. We propose to semi-automatically iden-tify such object mappings, and based on the mapping, recognize the instructor's action, and replicate it in the student's space. Our preliminary user testing showed the possibility of the framework in improving the communication and process of learning.