Video Emotion Recognition

a demo that recognize the conveyed emotion of the video


Emotion refers to a person’s disposition to respond to the media carrier. To this end, Video Emotion Recognition uncover the underlying attitude or opinions of viewers towards the given video.

Emotions conveyed in the video come across culture, nature, and situations. It involves film, VLog, animation, mv, and so on.
The subject also covers humans, animals, landscapes, and others.

Recognition Example

😀 Positive: 11.7

😕 Negative: 88.3

gantt title Emotion Confidence dateFormat X axisFormat %s section Amusement 5:0, 5 section Contentment 0: 0, 0 section Awe 2: 0, 2 section Excitement 2: 0, 2 section Fear 25: 0, 25 section Sadness 22: 0, 22 section Disgust 6: 0, 6 section Anger 38: 0, 38

Online Demo

We also provide an online demo that can upload any video you like. Due to the limitation of computation power of our server, if you want to process images in batch, please refer to our provided script on Github:CTEN or Github:TSL.

  • Weakly Supervised Video Emotion Detection and Prediction via Cross-Modal Temporal Erasing Network. Zhicheng Zhang, Lijuan Wang, and Jufeng Yang, IEEE CVPR, 2023.
  • Temporal Sentiment Localization: Listen and Look in Untrimmed Videos. Zhicheng Zhang and Jufeng Yang, ACM Multimedia, 2022.
  • Emotion, Theory, Research, and Experience: Theories of emotion. Kellerman Henry and Robert Plutchik, Academic Press, 1980.