Progressive and corrective feedback for latent fingerprint enhancement

Name: Kittipol Horapong

LCSH: Kasetsart University -- Dissertations. D.Eng. (Electrical Engineering) 2021

Classification :.LCCS: TA1637

LCSH: Kasetsart University. -- Department of Electrical Engineering -- Dissertations

LCSH: Image processing

LCSH: Fingerprints

LCSH: Pattern recognition systems

Abstract: The progressive mechanism has shown a remarkable performance to enhance friction ridges on the corrupted areas in the Adaptive Boosted Spectral Filtering (ABSF) technique. In this technique, the fingerprint pattern propagates itself sequentially from a clear to the corrupted ridge areas. Unfortunately, there might be a propagated fingerprint pattern error in its enhanced results relating to the long propagation sequence and the incorrect starting location performed in the propagation. This study proposes a progressive and corrective feedback (PCF) mechanism for latent fingerprint enhancement. We carry out the ABSF technique to perform the progressive mechanism for input latent fingerprint. Then, we apply a spectral autoencoder to detect anomalous fingerprint patterns in the enhanced image and enable its refinement by repeating the ABSF technique in such potential anomalous areas. This method can manipulate a corrective feedback mechanism to reduce fingerprint patterns’ cumulative error over the propagation sequence. Secondly, we propose a multiple initial block localization at the first step of propagation to mitigate fingerprint patterns’ initial bias. This method can start the fingerprint pattern that propagates in several locations at once, rather than a single location. In the experiment, we conducted two performance tests: minutiae detection test and latent fingerprint identification test. The experimental results showed that the proposed algorithm could improve the accuracy of detected minutiae compared to a set of true minutiae marked by latent experts. Also, the proposed algorithm achieved high accuracy and excellent reliability in latent fingerprint identification compared to the state-of-the-art latent fingerprint enhancement algorithms.

Kasetsart University. Office of the University Library

Address: Bangkok

Email: tdckulib@ku.ac.th

Name: Vutipong Areekul

Role: Thesis Advisor

Created: 2021

Modified: 2025-07-14

Issued: 2025-07-14

วิทยานิพนธ์/Thesis

application/pdf

URL: https://www.lib.ku.ac.th/KUthesis/2564/kittipol-hor-all.pdf

CallNumber: TA1637 .K58

eng

DegreeName: Doctor of Engineering

Level: Doctoral Degraee

Descipline: Electrical Engineering

Grantor: Kasetsart University

©copyrights Kasetsart University

ลำดับที่.	ชื่อแฟ้มข้อมูล	ขนาดแฟ้มข้อมูล	จำนวนเข้าถึง	วัน-เวลาเข้าถึงล่าสุด
1	kittipol-hor-all.pdf	4.85 MB

ใช้เวลา

0.018092 วินาที