This repository contains code for CP-Net, a deep learning model that combines a U-Net based architecture MLN (Macro-Localization Network) and a CNN-based architecture MRN (Micro-Regression Network) for fingerprint core-point detection. The pre-trained models (single hourglass) can be downloaded from this link.
NOTE: To use the following .py files one must add all folders in their directory. The folder names have been appropriately mentioned throughout the code files.
train_mln.py - Contains code necessary to train the MLN network. It also helps generate masked images for fingerprint images.
train_mrn.py - Contains code necessary to train the MRN network. It also helps predict core point coordinates for fingerprint images.
G. Arora, A. Kumbhat, A. Bhatia and K. Tiwari, "CP-Net: Multi-Scale Core Point Localization in Fingerprints Using Hourglass Network," 2023 11th International Workshop on Biometrics and Forensics (IWBF), Barcelona, Spain, 2023, pp. 1-6, doi: 10.1109/IWBF57495.2023.10157521.