Facial Expressions Close-Up Dataset (2,000 Images)

Dataset Features

• Expression Type: Joy, anger, sadness, disgust, fear, surprise, neutral, confusion.
• Image Type: Close-up and extreme close-up facial photos.
• Subjects: Adults and kids (6–50), mixed gender, mainly European/Mediterranean.
• Angles & Variations: Multiple head positions, lighting conditions, gaze directions.
• Quality: High-resolution images (3000–6000 px), manually curated.
• Format: JPEG files.

Distribution

• Data Volume: 2,000 images.
• Organization: Single folder, sequential filenames for easy ingestion.
• Archive Type: ZIP package containing all files.

Usage

• Application: Facial expression recognition and affective computing.
• Application: Generative AI training and diffusion model enhancement.
• Application: Emotion classification tasks for machine learning models.
• Application: Face alignment, landmark detection, identity-agnostic CV tasks.
• Application: Academic research on human-computer interaction (HCI).

Coverage

• Geographic Coverage: Global – suitable for worldwide use cases.
• Time Range: Collected between 2024 and 2025.
• Demographics: Adults and kids, mixed gender, European/Mediterranean representation.

License

Who Can Use It

• Data Scientists: For training computer vision emotion models.
• Researchers: For academic or scientific studies on facial expressions.
• Businesses: For AI development, emotion analytics, and product R&D.
• AI Companies: To improve generative model facial realism.
• Developers: For model fine-tuning, pose estimation, and ML pipelines.

• All images were captured, curated, and are fully owned by Alessandro Grandini.
• All subjects provided full authorization for dataset distribution and AI usage.