Panopticon Blob Detection

SurveillanceBlobs uses on-device machine learning to detect and visualize objects in real time through your camera — no internet required.

DETECTION MODES
- BOX — bounding boxes around detected objects
- LABEL — class name + confidence overlay
- LINES — connection lines between objects with distance labels
- MASK — instance segmentation with liquid mask effects
- PIX — pixelate detected regions (inside or outside)
- DIV — subdivide and animate detected regions with sine-wave motion, adjustable BPM, amplitude, and collapse effects

FEATURES
- Multiple detectable objects
- Front and rear camera support with full device rotation
- Record video with all visual effects baked in
- Import existing videos and apply detection + effects
- Per-class color coding
- Adjustable confidence threshold, line width, and tracking persistence
- Class filtering — choose which object types to detect
- All processing runs on-device using Core ML — your camera feed never leaves your phone

SETTINGS & CUSTOMIZATION
- Box, label, and grid color pickers
- Pixelation block size control
- DIV mode: grid density, collapse amount, BPM, sine amplitude, foreground/background targeting
- Detection persistence for smooth, stable tracking