What is it?
The Smart Person Identification (SPI) is a device capable of capturing specific biometric characteristics of a non-cooperative person in uncontrolled conditions. These characteristics represent a digital “identity card” through which to identify any subject.
Architecture
The SPI device is designed based on a multi-layered architecture that allows processing, in a parallel and simultaneous manner, data from different sensors. The algorithms implemented within the SPI device allow capturing the morphology,
the geometry, and the metrics of the face even if it moves or is positioned at angles that hinder the feature extraction process. Additionally, complementary sensors, such as high-performance microphones, allow enriching (e.g., voice) the digital “identity card” of the subject. Finally, the quality of the captured features allows, in post-processing, inferring potentially interesting information, such as somatic traits and distinctive signs.
Sensor Technology
The basic components of SPI include the use of four high-performance sensors. The first of these, and the core of the system, is the High-Speed and High-Resolution RGB Camera (H2SR-C), a high-resolution (spatial and temporal) sensor capable of accurately detecting the details of a person’s face. The technical specifications of the sensor allow for accurate face detection and meticulous extraction of its distinctive features (Face Recognition). The second sensor is the High Dynamic Range Imaging Camera (HDRI-C), a spatial depth sensor capable of estimating basic biometric characteristics of a person (e.g., height, skeletal structure, volume). Furthermore, the use of this sensor allows classifying visible behaviors (Behavior Analysis) and gestures (Gesture Recognition) of an individual, which can be used as additional distinctive features. The third sensor is the High Dynamic Thermal Imaging Camera (HDTI-C), a thermal sensor that helps in the processes of Detection, Analysis, and Recognition in extreme environmental conditions (e.g., night vision, compromised visibility). Finally, the fourth basic sensor is represented by the Directional Microphone Array (DMA), a sound wave sensor capable of acquiring and analyzing an individual’s voice (Voice Analysis) by extracting peculiar characteristics (e.g., tone, stress level).
Main Features
- H2SR-C: Resolution Full HD (up to) 2048×2048, Frame Rate (up to) 180 FPS, Maximum Range (up to) 20 meters. Latency (about) 2.8 ms.
- HDRI-C: Depth Resolution (up to) 1024×768, Frame Rate (up to) 120 FPS, Maximum Range (up to) 10 meters. Diffused Laser and LED.
- HDTI-C: Thermal Resolution (up to) 1024×768, Frequency 9Hz/25Hz, Thermal Sensitivity (about) 100mK, Maximum Range (up to) 4km.
- DMA: Sensitivity Tollerance (about) 1dB, SNR Ratio 65 dBA, Sensitivity of -26 dBFS.
