This paper is available on arxiv under CC 4.0 license.
Authors:
(1) Zhihang Ren, University of California, Berkeley and these authors contributed equally to this work (Email: peter.zhren@berkeley.edu);
(2) Jefferson Ortega, University of California, Berkeley and these authors contributed equally to this work (Email: jefferson_ortega@berkeley.edu);
(3) Yifan Wang, University of California, Berkeley and these authors contributed equally to this work (Email: wyf020803@berkeley.edu);
(4) Zhimin Chen, University of California, Berkeley (Email: zhimin@berkeley.edu);
(5) Yunhui Guo, University of Texas at Dallas (Email: yunhui.guo@utdallas.edu);
(6) Stella X. Yu, University of California, Berkeley and University of Michigan, Ann Arbor (Email: stellayu@umich.edu);
(7) David Whitney, University of California, Berkeley (Email: dwhitney@berkeley.edu).
Table of Links
- Abstract and Intro
- Related Wok
- VEATIC Dataset
- Experiments
- Discussion
- Conclusion
- More About Stimuli
- Annotation Details
- Outlier Processing
- Subject Agreement Across Videos
- Familiarity and Enjoyment Ratings and References
7. More About Stimuli
All videos used in the VEATIC dataset were selected from an online video-sharing website (YouTube). The VEATIC dataset contains 124 video clips, 104 clips from Hollywood movies, 15 clips from home videos, and 5 clips from documentaries or reality TV shows. Specifically, we classify Documentary videos as any videos that show candid social interactions but have some form of video editing, while home videos refer to videos that show candid social interactions without any video editing. All Videos in the dataset had a frame rate of 25 frames per second and ranged in resolution with the lowest being 202 x 360 and the highest being 1920 x 1080.
Except for the overview of video frames in Figure 2, we show more samples in Figure 9. Moreover, unlike previously published datasets where most frames contain the main character [31, 29, 32], VEATIC not only has frames containing the selected character but also there are lots of frames containing unselected characters and pure backgrounds (Figure 10). Therefore, VEATIC is more similar to our daily life scenarios, and the algorithms trained on it will be more promising for daily applications.
This paper is available on arxiv under CC 4.0 license.