EMOPortraits: Emotion-enhanced Multimodal One-shot Head Avatars
Paper • 2404.19110 • Published • 4
Dataset Viewer
The dataset viewer is not available for this dataset.
Unexpected token '<', "<html>
<h"... is not valid JSON
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
FEED: Facial Extreme Emotions Dataset
Dataset Download Links:
Video Naming Conventions:
*_C.mp4: Video recorded from the central camera.*_R.mp4: Video recorded from the right camera.*_L.mp4: Video recorded from the left camera.
Video Content Descriptions:
blinks_*: The person performs blinks (typically both eyes together and each eye separately).eyes_*: The person performs eye movements.rotations_*: The person performs head rotations.tongue_rot_*: The person performs head rotations while being asked to demonstrate their tongue during the rotations.MAU_mild/strong_*: The person performs the Main Action Units (Surprise, Fear, Disgust, Anger, Happiness, Sadness) in mild and strong intensities without head rotations or tongue expressions.FF_*: The person performs the Main Action Units in strong intensity with head rotations and attempts to make the expressions as asymmetrical as possible.FF_NR_*: The person performs the Main Action Units in strong intensity without head rotations, focusing on making the expressions as asymmetrical as possible.tongue_em_*: Similar toFF_NR_*, but the person is asked to use their tongue in expressions where possible.random_*: The person is asked to perform random extreme expressions continuously for 2-3 minutes.
For more details, please visit the project page and refer to the EMOPortraits: Emotion-enhanced Multimodal One-shot Head Avatars paper.
- Downloads last month
- 250