BeFaced: A Game for Crowd-sourcing Facial Expression Recognition

Brief

Machine learning algorithms for facial expression analysis systems often depend on having a set of high quality face images as training examples. To train the systems robustly, the database needs to be large and images need to have high variability in terms of facial features, pose and illumination, amongst other variables. Unfortunately, collecting such databases is costly and time consuming. Moreover the current popular databases are mainly collected in artificial lab environments with relatively small population sizes.

Using the psychological draw of popular casual games, BeFaced is a casual tablet game that enables massive crowdsourcing of facial expressions exemplars. The core mechanic is similar to Bejeweled, a very popular tile-matching puzzle game that has been downloaded over 150 million times. We created an alternative version of the tile-matching gameplay mechanic to include facial expressions as player input and aim to use the gameplay appeal to obtain a large database of natural and varied facial expressions in the wild. A major advantage is also the ability to “request” the player for any type of expressions depending on the tiles we design. Also, we can alleviate privacy concerns by providing the user with the option to just provide us with the tracked feature points instead of the raw image itself.

Members

Chek Tien Tan
Daniel Rosser
Natalie Harrold

Outputs

C. T. Tan, H. Sapkota, and D. Rosser, “BeFaced: a Casual Game to Crowdsource Facial Expressions in the Wild,” in Proc. CHI 2014 Ext. Abstracts, 2014.

C. T. Tan, H. Sapkota, D. Rosser, and Y. Pisan, “A Game to Crowdsource Data for Affective Computing,” in Proc. FDG 2014 Games, 2014.

C. T. Tan, H. Sapkota, D. Rosser, and Y. Pisan, “Initial Perceptions of a Casual Game to Crowdsource Facial Expressions in the Wild,” in Proc. FDG 2014 WIP, 2014.

C. T. Tan, D. Rosser, and N. Harrold, “Crowdsourcing facial expressions using popular gameplay,” in Proceedings of SIGGRAPH Asia 2013 Technical Briefs, 2013. [prezi

C. T. Tan, D. Rosser, and N. Harrold, “BeFaced : a game for crowdsourcing facial expressions,” in Proceedings of SIGGRAPH Asia 2013 MGIA (Demo), 2013.

Oct. 26, 2016, 3:09 p.m.