Krezy Detection
A Python emotion-detection library for images and webcam streams with reusable app entry points.
- Status
- Completed
- Date
- November 2024
- Category
- Intelligence
Summary
A Python emotion-detection library that analyzes images and webcam streams with DeepFace and exposes reusable web, desktop, and mobile integration entry points.
Description
Krezy is a packaged Python library for facial emotion analysis across multiple app surfaces. It solves the problem of wiring emotion inference separately for each platform by centralizing detection in one EmotionDetector API.
Architecture
- Face detection
- Emotion inference
- Structured output
- Cross-platform integration
Technology
- Python
- DeepFace
- OpenCV
- NumPy
- Flask
- Kivy
- CustomTkinter
Features
- Real-time emotion detection from webcam streams with live visual overlays and dominant-emotion highlighting
- Image emotion analysis for file paths or in-memory image arrays with per-emotion percentage breakdowns
- Configurable face-detection backends (OpenCV, SSD, MTCNN, RetinaFace) to balance speed and detection accuracy
- Structured inference output with success status, dominant emotion, full emotion scores, error details, and ISO timestamps
- Cross-platform integration layer with ready-to-use web, desktop, and mobile app entry points from one core detector API
- Flexible Python packaging with optional platform extras for targeted installs (core, web, mobile, desktop, or all)
Links
Images
Machine-readable: /projects.json · /llms.txt · /llms-full.txt