This article presents a prototype of Image and Thermal Fingerprint Dual-Sensor that includes collecting optical fingerprint ridges with simultaneous analysis of the session thermal heat distribution. The prototype uses Raspberry Pi 3B+ with a Pi Camera (OV5647) to capture optical images and an MLX90640 thermal infrared array (24×32 pixels resolution) for thermal images and to establish "physiological" confirmation. The optical and thermal exposure systems capture images in two different formats (.jpg for optical and .npy for thermal matrix). Once the images are enrolled, similarity between the input images and data are compared using SSIM and MSE algorithms. Experimental verification showed the SSIM of the genuine trials were >0.55 and the thermal MSE of <4.0. The prototype was able to fully complete a full verification cycle in ~8-10 seconds in a verifiable nature.
Gupta et al. (Fri,) studied this question.