Micro-Expressions Dataset

Data Collection and Processing:

• Sources: Google Image Search
• Preprocessing Steps:
  • Removal of duplicate images
  • Exclusion of images without faces
  • Cropping to focus on faces
  • Eliminating images smaller than 80×80 pixels
  • Resizing all images to 80×80 pixels
  • Manual validation to ensure quality

Applications:

• Emotion detection in AI models
• Behavioral studies
• Enhancing human-computer interaction systems

Enhanced Features:

• Metadata: Each image is annotated with metadata detailing the age group (child, adult, elder) and the specific emotion displayed.
• Diverse Demographics: Images include a balanced representation of various ethnicities and age groups to improve the robustness of models trained on this dataset.
• Quality Assurance: The dataset undergoes rigorous quality checks, including manual verification, to ensure high standards and reliability.
• Compliance: Adheres to privacy standards and ethical guidelines in data collection and usage.

This dataset is sourced from Kaggle.