Back to datasets
Dataset assetOpen Source CommunityComputer VisionFacial Expression Recognition

DAiSEE

The DAiSEE dataset is used for facial expression recognition and contains 300 frames extracted from each video. Each frame includes 68 facial landmark points and 52 action units.

Source
github
Created
Aug 20, 2024
Updated
Aug 21, 2024
Signals
545 views
Availability
Linked source ready
Overview

Dataset description and usage context

FER_DAiSEE_2Stream‑GCN‑CNN

Dataset Overview

  • Dataset Name: DAiSEE
  • Application Domain: Facial Expression Recognition (FER)
  • Classification: Engagement class

Data Processing

  • Video Frame Extraction: 300 frames per video
  • Facial Feature Extraction: For each frame, 68 facial landmark points and 52 types of action units are extracted
Need downstream help?

Pair the dataset with AI analysis and content workflows.

Once the source passes your review, move straight into summarization, transformation, report drafting, or presentation generation with the JuheAI toolchain.

Explore AI studio