Normalization Pipeline

• Automatic resolution standardization
• Color space normalization (RGB, BGR, grayscale)
• Frame extraction from video with configurable FPS
• Aspect ratio preservation & padding strategies
• Batch processing with parallel execution

ROI & Privacy Processing

• Region-of-interest (ROI) cropping per scene
• Automated face & license plate blurring
• Sensitive-area masking with configurable zones
• GDPR & PDP compliance preprocessing
• Anonymization audit trail

Scene-Aware Augmentation

• Weather simulation (rain, fog, snow, glare)
• Day/night lighting transformations
• Perspective & geometric augmentations
• Mosaic & cutout augmentation strategies
• Class-imbalance compensation via targeted augmentation

Dataset Health Analytics

• Class distribution analysis & rebalancing
• Schema validation against annotation standards
• Outlier & corrupt data detection
• Label quality scoring & anomaly flagging
• Dataset readiness report with actionable insights