Changelog

FaceGuard release notes

v1.1.1 - 2025-09-27

Upstream Data Forwarding: Configurable forwarding of data from upstream filters
Face Detection Metadata: Rich face coordinates, confidence scores, and detection details in frame data
Environment Variable Configuration: Complete environment variable support for all parameters
Dynamic Output Naming: Automatic generation of descriptive output video filenames
Debug Logging: Optional debug mode with detailed face detection logging
Comprehensive Test Suite: Integration and smoke tests for all functionality
Enhanced Documentation: Detailed README and overview documentation with usage examples

Configuration Cleanup: Removed unused confidence_threshold parameter, kept only detection_confidence_threshold
String-to-Type Conversion: Improved configuration validation with proper type conversion
Usage Script: Simplified filter_usage.py with environment variable support
Test Coverage: Updated all tests to use cleaned configuration parameters

Configuration Validation: Fixed string-to-type conversion issues in normalize_config
Test Mocking: Resolved import path issues in test suite
Environment Variable Loading: Implemented proper environment variable handling
Documentation: Updated all documentation to reflect current configuration options

Initial Release: new filter for detecting and blurring faces in video frames using a pluggable face detection and blur backend.
Face Detection
- Uses a model (e.g., face_detection_yunet_2023mar.onnx) from a configurable remote URL
Customizable Detection and Blur Backends
- Detector name (DETECTOR_NAME) and blurring method (BLURRER_NAME) can be set via environment variables
Plug-and-Play Processing
- Processes the main frame and replaces it with a version where detected faces are blurred
Environment Variable Integration
- Model location and runtime behavior are controlled through .env or environment vars
Lightweight Configuration
- No required configuration fields; minimal setup needed for integration