Near-Sensor Neuromorphic Computing System Based on a Thermopile Infrared Detector and a Memristor for Encrypted Visual Information Transmission.

Near-sensor neuromorphic computing systems that utilize photodetectors and memristors exhibit significant promise in the domains of visual information processing, transmission, and noise reduction recognition. In comparison to conventional photodetectors operating within the visible-light spectrum, thermopile infrared detectors offer distinct advantages in terms of concealment and security. This study proposes an integrated near-sensor computing system that combines a thermoelectric infrared detector with a memristor, which demonstrates a broad detection range (100-310 °C), rapid response time for sensing infrared signals, and excellent neuromorphic computing characteristics for information processing. Besides high-accuracy recognition of handwritten digits, near-infrared visual information recognition and voice recognition for double information encryption were demonstrated in the system. This neuromorphic computing system holds considerable potential for applications in the propagation, encryption, and recognition of security information within the infrared spectrum.