Cross-Domain Consistent Fingerprint Denoising

Abstract

Performance of state-of-the-art fingerprint denoising model on poor quality fingerprints degrades due to crossdomain shift observed between training and testing domains. To address this limitation, we present a cross-domain consistent fingerprint denoising model, which ensures that the output of two fingerprint images with the same ridge structure, however varying contrast and ridge-valley clarity should be similar. Results indicate that the proposed CDC-GAN outperforms state-of-the-art fingerprint denoising algorithms on challenging publicly available poor quality fingerprint databases.

Context-Aware Restoration of Noisy Fingerprints

“…CA-GAN is implemented using PyTorch, v1.11.0 and exploits Adam optimizer with a learning rate of 0.0002. 4.90 DeConvNet [15] 4.09 DU-GAN [19] 3.01 MU-GAN [21] 1.48 CDC-GAN [3] 2.38 CA-GAN 2.03…”

“…Indu Joshi 1,2 , Tushar Prakash 3 , B. S. Jaiswal 4 , Rohit Kumar 3 , Antitza Dantcheva 1 , Sumantra Dutta Roy 2 and Prem Kumar Kalra 2 1 Inria Sophia Antipolis, France 2 IIT Delhi, India 3 Delhi Technological University, India 4 Delhi Police, India…”

“…To the best of our knowledge, CA-GAN is the introductory research that uses the principle of learning spatial context through solving jigsaw puzzles to boost the generalization ability of a fingerprint restoration model. 2.05 DeconvNet [15] 1.95 Cycle-GAN [16] 1.76 MU-GAN [21] 1.33 FP-E-GAN [18] 1.31 CDC-GAN [3] 1.30 DU-GAN [19] 1.26 CA-GAN 1.29…”

Deep learning with fetal ECG recognition

“…There is a difference between the data distribution in the training set (source domain) and the test set (target domain) (Joshi et al 2022). In addition to the baseline loss, a cross-domain consistency loss ℓ cdc (•) is used to reduce the gap between the source and target domains (Wang and Zheng 2022).…”