โ† Back to Projects
๐Ÿ›ก๏ธ Defense & SecurityLive Demo

Camouflage Effectiveness Analyzer

Adversarial camouflage testing: an AI detector hunts for you at simulated distances. Try it live.

TensorFlow.jsObject DetectionComputer VisionReact

Overview

Modern surveillance is automated, so camouflage should be tested against the actual adversary: AI detection models. Upload a photo and an object-detection model hunts for people in it at four simulated distances, producing a detection-range profile and a stealth score. The live browser demo runs a single on-device detector; the full version expands to a multi-model ensemble (YOLO, Faster R-CNN, RetinaNet) with lighting variation and AI-powered improvement suggestions.

Key Features

  • โœ“Adversarial testing with detection models as the enemy
  • โœ“Detection-range profile across simulated distances
  • โœ“Stealth scoring with per-distance confidence
  • โœ“Runs fully on-device; multi-model ensemble planned