Progressive and Corrective Feedback for Latent Fingerprint Enhancement Using Boosted Spectral Filtering and Spectral Autoencoder

Abstract

The objective of this research is to design an efficient algorithm that can successfully enhance a targeted latent fingerprint from various complex backgrounds under an uncontrolled environment. Most algorithms in literature exploited dictionary learning schemes and deep learning architectures to capture latent fingerprints from complicated backgrounds and noise. However, an algorithm learned from other high-quality fingerprint images may not solve all possible cases within a given unseen image. We propose a new feedback framework to distinguish latent fingerprints from complex backgrounds and gradually improve friction-ridge quality using the information provided inside a given unseen image. We combine two efficient mechanisms. The first mechanism enhances high-quality areas in priority and feeds the enhanced areas back to improve the quality of latent fingerprints in the nearby area. The second mechanism is to verify that the first mechanism works correctly by detecting anomalously enhanced fingerprint patterns. The second mechanism employs a spectral autoencoder that learns from good fingerprint spectra in the frequency domain. The anomalous fingerprint area is sent back to the first mechanism for further improving the enhanced result. We benchmark the proposed algorithm against available state-of-the-art algorithms using two fingerprint matching systems (one commercial off-the-shelf and one open-source) on two public latent fingerprint databases. The experimental results show that the proposed algorithm outperforms most state-of-the-art algorithms in the literature.

Spectral Filter Predictor for Progressive Latent Fingerprint Restoration

“…Hence, we propose a new signal-to-noise ratio (SNR) concept to measure the friction ridge quality in the local area of a given latent fingerprint. The new quality measure is simpler and more efficient than the complex rejection process used in our previous work [7]. This new quality measure helps the proposed framework select and prioritize high-quality friction ridge areas.…”

“…Unstable ridge orientation and frequency estimation 45.3% FingerNet [18] (2017) Frequency Domain Utilize progressive feedback [6] with spectral predictor filtering in the frequency domain using CNN with an improved initial block selection process [7]. Singular point areas are preserved.…”

“…Additionally, they presented a simple method for automatic initial block selection. Recently, in 2021, Horapong et al [7] introduced a novel framework that combines two mechanisms of progressive and corrective feedback, called PCF, to increase the efficiency of [32] and [33]. The progressive feedback mechanism employs spectral boosting with matched filters similar to [32].…”

“…The previous progressive and corrective feedback method [7] relied on a matched filter in the frequency domain for restoring each latent partitioning block. The matched filter used in the previous method [7] passed all spectra within the predefined ridge frequencies. Unfortunately, spectra of background and noise within these ridge frequencies could be preserved in the restored fingerprints.…”

On restoration of degraded fingerprints

“…Svoboda et al [43] propose a autoencoder network which minimizes gradient and orientation between the output and target enhanced image. Horapong et al [44] propose a spectral autoencoder based feedback mechanism to identify anomalously enhanced fingerprint regions. The poorly enhanced regions are iteratively enhanced to improve quality of fingerprint ridge patterns.…”