Fingerprint Identification System using NIR-To-NIR Upconverting Nanoparticles

Thursday, March 11, 2021: 2:10 PM - 2:30 PM

---

Description

A light emitting diode (LED) based reader system has been designed and built to collect near-infrared emissions from upconverting nanoparticles (UCNPs) used to develop fingerprints. A complete system is essential because the UCNPs emit light at 800 nm, a region that is normally filtered out by visible cameras. The reader system includes an LED that emits approximately 980 nm wavelength light. The LED is mounted on a heat sink. A parabolic reflector is mounted to the LED/heat sink to direct and collimate the LED output. A variety of collection systems have been utilized, including a commercially available digital camera with its infrared filter removed. The camera used may be one integrated into a cell phone or a more standard camera, such as a digital SLR camera. The camera input light must be filtered so that only the 800 nm UCNP emission enters the camera. This has been accomplished with a long pass filter and a short pass filter. These near-infrared emitting UCNPs have 50x greater normalized relative brightness as compared to visible light emitting UCNPs. Once the emitted light is collected by the camera, the fingerprint data must be captured, stored and compared to known fingerprint samples for identification. For the cell phone-based system, an app has been developed to identify minutiae and then to compare these minutiae to fingerprints in a database. This app is based on neural networks. With proper neural network training, minutiae identification has been shown to be quite accurate. Another aspect of system design is safely applying the UCNPs to the fingerprint. Standard dry particle methods have been shown to be highly effective, but the health and safety of workers using dry nanoparticles is of concern. As an alternative, several non-powder techniques are being evaluated. The system been tested, and shown effective, on a number of often problematic substrates including polymeric soda bottle labels and printed paper business cards.

---

Additional Info

Keywords: Please select up to 4 keywords ONLY:
Instrumentation,Neural Network