feat: Improve color handling, cluster analysis and add error resilience

[souvik03-136]

• Enhanced color space handling in capturer.py with explicit controls
• Improved cluster analysis with silhouette scoring for better face matching
• Added caching for NSFW model to improve performance
• Implemented proper error handling and logging across all modules
• Added configurable settings system in globals.py
• Created comprehensive unit tests
• Added detailed documentation with docstrings

Summary by Sourcery

Enhance system robustness and performance by improving color handling, cluster analysis, error resilience, and adding a configurable settings system

New Features:

• Added configurable settings system with JSON config file
• Implemented caching for NSFW model
• Added comprehensive error handling and logging across modules

Enhancements:

• Improved color space handling with explicit conversion controls
• Enhanced cluster analysis with silhouette scoring for better face matching
• Added more robust error handling in video and frame processing

Chores:

• Refactored code to improve type hints and logging
• Added default configuration settings