## GFT Facial Expression Database ## The Sayette Group Formation Task (GFT) Spontaneous Facial Expression Database has two components: meta-data and video data. Because video, by its very nature, reveals person identity, a separate procedure is needed for its distribution. The video data is available from the University of Pittsburgh using the procedures described below. The meta-data is available on this OSF webpage. This meta-data includes baseline results, MATLAB functions, frame-level annotations, and a wiki that describes the formatting and recommended analysis strategies for the data. The meta-data is available to all under the CC-BY Attribution 4.0 International license. However, the video data is not hosted on this website and is not covered by the CC-BY license. ### Database Motivation Despite the important role that facial expressions play in interpersonal communication and our knowledge that interpersonal behavior is influenced by social context, no currently available facial expression database includes multiple interacting participants. The Sayette Group Formation Task (GFT) database addresses the need for well-annotated video of multiple participants during unscripted interactions. The database includes 172,800 video frames from 96 participants in 32 three-person groups. To aid in the development of automated facial expression analysis systems, GFT includes expert annotations of FACS occurrence and intensity, facial landmark tracking, and baseline results for linear SVM, deep learning, active patch learning, and personalized classification. Baseline performance is quantified and compared using identical partitioning and a variety of metrics (including means and confidence intervals). ### Accessing the Database ### The GFT database is made available to researchers. To access the database, researchers must sign and submit a document agreeing to the terms of use. These terms require researchers to respect the data-uses that each participant consented to (see the [Meta-Data][1] file). Researchers must also agree to cite the following paper in any publications and technical reports that use the database. Note that, although the meta-data on this OSF website are licensed using the CC-BY license, the database itself has its own, separate license. To access the video data, submit a request form: [https://forms.office.com/r/stRKusD5XZ][2] Criteria for access to the video data is being an academic researcher and from a university that is not subject to US export controls. Once access to the video data has been approved, data transfer via Microsoft OneDrive will be arranged. If you have any questions or concerns about this process, please email [jmgirard@ku.edu][3]. **Update for 2021:** Our university has instituted a new policy for foreign licensing entities. There is a chance that your request will be denied due to this extra screening. ### Citing the Database ### [APA] Girard, J. M., Chu, W.S., Jeni, L. A., Cohn, J. F., De La Torre, F., & Sayette, M. A. (2017). Sayette group formation task (GFT) spontaneous facial expression database. In *Proceedings of the IEEE International Conference on Automated Face & Gesture Recognition*. [BiBTeX] @inproceedings{GFTDatabase, author = {Girard, Jeffrey M and Chu, Wen-Sheng and Jeni, L{\'{a}}szl{\'{o}} A and Cohn, Jeffrey F and {De La Torre}, Fernando and Sayette, Michael A}, booktitle = {Proceedings of the IEEE International Conference on Automated Face {\&} Gesture Recognition}, title = {{Sayette group formation task (GFT) spontaneous facial expression database}}, year = {2017} } [1]: https://osf.io/mnfyb/ [2]: https://forms.office.com/r/stRKusD5XZ [3]: http://mailto:jmgirard@ku.edu