Cephalopod-inspired embodied camouflage system for simultaneous evasion of artificial and human visual perception.

CephCam, a cephalopod-inspired embodied closed-loop camouflage system on a model vehicle, significantly outperforms strong baselines by evading both human observers and modern object detectors.

• Why it matters: Visual perception differences between biology and AI make camouflage that is natural to humans and effective against detectors difficult, especially when adversarial methods are static or conspicuous.
• What they did: CephCam uses an eye-brain-skin sensing-computation-display loop with environment-aware color extraction, optimization-based texture generation balancing visual naturalness and detector evasion, and angle-aware OLED display compensation.
• The result: Across diverse digital and physical environments, it achieves simultaneous evasion of human observers and modern object detectors, suggesting shared vulnerabilities that enable robust, adaptive embodied camouflage.