Researchers develop system to improve recognition of latent fingerprints

Reservoir computer system based on a photo-synapse and an array of memristor devices. a Traditional fingerprint recognition system data processing mode, realized by independent optical sensor, memory chip and CPU. b Schematic of the human visual recognition system including the retina, optic neurons and visual cortex in the human brain. c Proposed RC system with optical synapses as the reservoir input layer and the memristor device array as the readout array. The inset in the dotted box is the abstract photoelectronic RC system. The original optical information is passed into the photoelectron reservoir, where the inputs are nonlinearly mapped into feature outputs based on the PPC effect. And then the memristor network receives the outputs from the reservoir and implements the learning to read. Credit: Nature Communication (2022). DOI: 10.1038/s41467-022-34230-8

Recently, a research group led by Professor Long Shibing of the University of Science and Technology of China (USTC) of the Chinese Academy of Sciences, in collaboration with Professor Liu Qi of Fudan University, developed a system Reservoir computing in the sensor for latent fingerprint recognition with deep ultraviolet photo-synapses and a memristor array. This study was published in Nature Communication.

Deep ultraviolet (DUV) photodetectors play a central role in deep space exploration, environmental monitoring, and bioinformation identification. However, conventional ex-situ DUV fingerprint recognition systems use a separate sensor, memory and processor, which greatly increases the latency in decision-making and thus the overall computing power.

Inspired by the human visual perception system, the research group built a DUV sensor RC system with optical synapses as the tank input layer and the memristor device network as the readout network, which can detect and process in parallel to ensure high efficiency. and low power consumption.

The research team used Ga-rich component design and developed amorphous GaOx (a-GaOx) photosynapses with enhanced persistent photoconductivity (PPC) effects. A nonlinear mapping relationship for the RC system in the DUV sensor was constructed by inputting 4-bit equivalent light pulses for simulation so that image pixel sequence information could be sampled for feature values.

Ultimately, the formation of the tank outlets was achieved through the stable polymorphic modulation properties of the memristor device array, enabling small-scale DUV fingerprint recognition. The excellent recognition accuracy of DUV fingerprint images when using a dual functionality strategy and this hardware system is almost identical to the simulated results.

The system achieves 100% recognition accuracy after 100 training epochs and maintains 90% accuracy even in the presence of 15% background noise, in accordance with the anti-noise characteristics of DUV light.

This all-hardware DUV sensor-integrated RC system provides a good reference prototype for efficient recognition and secure applications of latent fingerprints. It is also an essential reference for the development of smart optoelectronic devices in the DUV frequency band.

“This prototype system…will provide more insight into emerging computing in sensor reservoirs. Overall, the subject of this work is really interesting,” said a reviewer from Nature Communication.

Provided by University of Science and Technology of China