pyOpticalPUF is a python library designed to help facilitate the analysis of optical physically unclonable functions (OPUFs). This library contains several fingerprinting algorithms, useful metrics and utility functions to make analysis easier.
Currently local binary patterns (LBP), adaptive high boost (AHB), Otsu and Sauvola are the supported fingerprinting algorithms. The parameters for each of these algorithms are accessible within a companion object designed to facilitate ease of use and minimization of errors.
from pyOpticalPUF import LBP, LBPParameters
import cv2
params = LBPParameters(radius = 2, neighbours = 8, keySize = (512, 512))
lenna = cv2.imread("Examples/Lenna_(test_image).png", cv2.IMREAD_GRAYSCALE)
fingerprint = LBP.calculateFingerprint(lenna, parameters = params)
cv2.imshow("Lenna", lenna)
cv2.imshow("Leena-print", fingerprint)
cv2.waitKey()which displays
| Lenna | Lenna-print |
|---|---|
.png) |
 |
This library comes with a few built in displays to aid with visualisation and interpretation.
Display example here
...
...
...
Ella Mann-Andrews, Thomas McGrath, Blake Halliday, Robert James Young; A guide for assessing optically imaged physically unclonable functions for authentication. Appl. Phys. Rev. 1 June 2025; 12 (2): 021314. https://doi.org/10.1063/5.0253691