Finger-UNet: A U-Net based Multi-Task Architecture for Deep Fingerprint Enhancement

TL;DR

This paper introduces Finger-UNet, a multi-task deep learning architecture based on U-Net, enhanced with wavelet transforms and domain knowledge, to improve low-quality fingerprint image enhancement for recognition tasks.

Contribution

The paper proposes a novel multi-task U-Net based architecture with wavelet attention and depthwise separable convolutions, integrating domain knowledge for superior fingerprint enhancement.

Findings

Outperforms previous methods on FVC 2002 and NIST SD302 datasets.

Effective in enhancing low-quality fingerprints with reduced memory footprint.

Multi-task learning improves fingerprint reconstruction accuracy.

Abstract

For decades, fingerprint recognition has been prevalent for security, forensics, and other biometric applications. However, the availability of good-quality fingerprints is challenging, making recognition difficult. Fingerprint images might be degraded with a poor ridge structure and noisy or less contrasting backgrounds. Hence, fingerprint enhancement plays a vital role in the early stages of the fingerprint recognition/verification pipeline. In this paper, we investigate and improvise the encoder-decoder style architecture and suggest intuitive modifications to U-Net to enhance low-quality fingerprints effectively. We investigate the use of Discrete Wavelet Transform (DWT) for fingerprint enhancement and use a wavelet attention module instead of max pooling which proves advantageous for our task. Moreover, we replace regular convolutions with depthwise separable convolutions, which significantly reduces the memory footprint of the model without degrading the performance. We also demonstrate that incorporating domain knowledge with fingerprint minutiae prediction task can improve fingerprint reconstruction through multi-task learning. Furthermore, we also integrate the orientation estimation task to propagate the knowledge of ridge orientations to enhance the performance further. We present the experimental results and evaluate our model on FVC 2002 and NIST SD302 databases to show the effectiveness of our approach compared to previous works.