Abstract
In co-design of several persons utilizing different materials together, capturing movement and position information of the hands as well as the speaking patterns of the designers provide answers to research questions related to social aspects of the co-design situation. Special motion-capture devices exist for precise movement tracking. They are, however, typically expensive and may restrict the movement of the designers. Recording the design sessions with a simple web camera offers a low-cost way to obtain the hand locations accurately enough but exploring the videos manually is a time-consuming and error-prone task. In this paper, we propose an inexpensive and automatic method to acquire information on the position of the hands and on the use of voice of the co-designers. We are offering our Matlab code as open source for other researchers and designers to use in their work and to amend.
Keywords
Co-design processes; Audio and video analysis; Hand tracking; Bayesian modeling; Open-source
Citation
Toivanen, M., Huotilainen, M., Chi, H., and Seitamaa-Hakkarainen, P. (2014) An automatic open-source analysis method for video and audio recordings of co-design processes, in Lim, Y., Niedderer, K., Redström, J., Stolterman, E. and Valtonen, A. (eds.), Design's Big Debates - DRS International Conference 2014, 16-19 June, Umeå, Sweden. https://dl.designresearchsociety.org/drs-conference-papers/drs2014/researchpapers/59
An automatic open-source analysis method for video and audio recordings of co-design processes
In co-design of several persons utilizing different materials together, capturing movement and position information of the hands as well as the speaking patterns of the designers provide answers to research questions related to social aspects of the co-design situation. Special motion-capture devices exist for precise movement tracking. They are, however, typically expensive and may restrict the movement of the designers. Recording the design sessions with a simple web camera offers a low-cost way to obtain the hand locations accurately enough but exploring the videos manually is a time-consuming and error-prone task. In this paper, we propose an inexpensive and automatic method to acquire information on the position of the hands and on the use of voice of the co-designers. We are offering our Matlab code as open source for other researchers and designers to use in their work and to amend.