Revisiting Handcrafted Minutiae Detection: A Simple and Effective Open Source Baseline for Modern Fingerprint Workflows

Raffaele Cappelli

cs.CVOct 8, 2026

Revisiting Handcrafted Minutiae Detection: A Simple and Effective Open Source Baseline for Modern Fingerprint Workflows

Authors: Raffaele Cappelli

Organizations: Department of Computer Science and Engineering, University of Bologna, Cesena, Italy

Abstract

Handcrafted minutiae detection algorithms remain fundamental to biometric science and forensic practice due to their full auditability, adherence to international standards, and operational independence from training datasets or GPU hardware. However, current open-source traditional baselines are severely outdated, relying almost exclusively on legacy C/C++ codebases that lack seamless integration with modern scientific software ecosystems. To bridge this gap, the present work introduces SBMEX (Skeleton-Based Minutiae EXtraction), a fast and deterministic minutiae detection method integrated into the open source \texttt{pyfing} package. SBMEX achieves high computational throughput by employing a dual Look-Up Table architecture that replaces runtime neighborhood scanning during Crossing Number computation and skeleton tracking. Additionally, it incorporates a continuous quality scoring framework driven by tracking path length, dual ridge-valley skeleton fusion, and spatial density decay. Rigorous evaluation on NIST SD302 datasets demonstrates that SBMEX delivers feature extraction accuracy comparable to or outperforming traditional open-source baselines without fine-tuning, while achieving a drastic reduction in minutiae detection latency relative to classical Crossing Number Python implementations.

Figures & tables

Figure 1: Classical minutiae extraction pipeline. (a) Portion of a fingerprint image displaying dark gray ridges separated by lighter valleys. (b) Enhanced and binarized image, where ridges are mapped to white pixels and valleys to black pixels. (c) Topological skeletonization of the ridges (1-pixel wide white paths). (d) Extracted minutiae features, highlighting a ridge ending near the top (blue circle) and a bifurcation located lower towards the center (red circle). The line segment originating from the center of each circle indicates the direction of the corresponding minutia, defined by an angle in $[0,2\pi)$ .

Symbol	Description	Default	Unit
$t_{b}$	Global binarization threshold	$0$	–
$d_{\text{bg}}$	Minimum distance to mask border	$10$	px
$d_{\text{min}}$	Minimum tracking distance	$5$	px
$d_{\text{max}}$	Maximum tracking distance	$20$	px
$d_{\text{pair}}$	Max pairing distance	$6$	px
$\theta_{\text{pair}}$	Max pairing angle difference	$\pi/6$	rad

Table 1: Summary of SBMEX parameters and default values at

500

DPI.

Pattern	$E(x,y)$	$\text{LUT}_{CN}$	$\text{LUT}_{ND}(d_{\text{prev}}\to[\mathcal{D}_{\text{next}}^{1},\mathcal{D}_{\text{next}}^{2},...])$
	0	0	—
	1	1	$4\to[\,]$
	1	1	$8\to[0]$
…	…	…	…
	32	1	$1\to[\,]$
	32	1	$8\to[5]$

Table 2: Representative

3\times 3

neighborhood patterns, their encoded state

E(x,y)

CN

values from

\text{LUT}_{CN}

, and valid successor directions

\mathcal{D}_{\text{next}}

from

\text{LUT}_{ND}

for reachable prior directions

d_{\text{prev}}\in\{0,\dots,7\}

and start state (

d_{\text{prev}}=8

Figure 2: Representative samples from the NIST SD302 datasets: (a) high-quality optical exemplar impression from SD302g, and (b) challenging latent impression from SD302h showing partial ridge patterns and severe background noise.

		NIST SD302 Exemplar						NIST SD302 Latent
		Type-agnostic			Type-aware			Type-agnostic			Type-aware
Method	Input	$\begin{subarray}{c}16\,\text{px}\\ \pi/6\,\text{rad}\end{subarray}$	$\begin{subarray}{c}12\,\text{px}\\ \pi/8\,\text{rad}\end{subarray}$	$\begin{subarray}{c}8\,\text{px}\\ \pi/10\,\text{rad}\end{subarray}$	$\begin{subarray}{c}16\,\text{px}\\ \pi/6\,\text{rad}\end{subarray}$	$\begin{subarray}{c}12\,\text{px}\\ \pi/8\,\text{rad}\end{subarray}$	$\begin{subarray}{c}8\,\text{px}\\ \pi/10\,\text{rad}\end{subarray}$	$\begin{subarray}{c}16\,\text{px}\\ \pi/6\,\text{rad}\end{subarray}$	$\begin{subarray}{c}12\,\text{px}\\ \pi/8\,\text{rad}\end{subarray}$	$\begin{subarray}{c}8\,\text{px}\\ \pi/10\,\text{rad}\end{subarray}$	$\begin{subarray}{c}16\,\text{px}\\ \pi/6\,\text{rad}\end{subarray}$	$\begin{subarray}{c}12\,\text{px}\\ \pi/8\,\text{rad}\end{subarray}$	$\begin{subarray}{c}8\,\text{px}\\ \pi/10\,\text{rad}\end{subarray}$
SBMEX	Enh	0.68	0.66	0.63	0.50	0.49	0.47	0.39	0.36	0.32	0.25	0.24	0.21
SBMEX-R	Enh	0.67	0.65	0.61	0.49	0.48	0.46	0.36	0.34	0.30	0.24	0.22	0.20
CNBaseline	Enh	0.61	0.58	0.52	0.44	0.42	0.39	0.26	0.22	0.18	0.16	0.14	0.11
MINDTCT	Raw	0.62	0.59	0.53	0.48	0.46	0.43	0.30	0.28	0.25	0.22	0.20	0.18

Table 3: Comparative performance (

F_{1}

-score) on the two datasets across the three threshold levels under type-agnostic and type-aware regimes. Bold indicates the best performance among competing methods, while underline denotes the second best.

Figure 3: PR curves on the two datasets across two threshold levels under type-agnostic and type-aware regimes.

Figure 4: Accuracy vs. Latency trade-off, at (

16

px,

\pi/6

rad), type-agnostic, on Exemplars (blue) and Latents (red).

Explore similar work

Topological summaries of fingerprint ridge patterns carry identity information
Jun 20, 2026Chad M. Topaz, Niny Arcila-Maya, Elizabeth Munch +2Biometric Identification Persistent Homology
Williams College, Williamstown, MA 01267 · QSIDE Institute, Williamstown, MA 01267 · University of Colorado Boulder, Boulder, CO 80309 +3
AFID: A Unified Open Framework for Automated Fingermark Identification, Quality Assessment and Feature Extraction
Sep 7, 2026Tim Oblak, Rudolf Haraksim, Peter PeerBiometric Identification Image Quality Assessment
Joint Research Centre of the European Commission in Ispra, Italy
PySIFT: GPU-Resident Deterministic SIFT for Deep Learning Vision Pipelines
May 18, 2026Sivakumar K. S., Mohammad Daniyalur Rahman, Gopi Raju MattaGPU Acceleration Image Matching
1Indian Institute of Technology Madras

Jun 20, 2026, cs.CVJ/K move · Enter open · S save

Topological summaries of fingerprint ridge patterns carry identity information

Chad M. Topaz, Niny Arcila-Maya, Elizabeth Munch, Zofia Stanley+1

Williams College, Williamstown, MA 01267 · QSIDE Institute, Williamstown, MA 01267 · University of Colorado Boulder, Boulder, CO 80309+3

Biometric Identification Persistent Homology Topological Data Analysis

Open paper AskPDF arXiv

Fingerprints are the most widely deployed biometric. Verifying whether two impressions come from the same finger typically relies on minutiae, small landmarks such as skin ridge endings and bifurcations. These landmarks are extracted through a multi-stage pipeline of image enhancement, skeletonization, minutiae detection, and alignment. We investigate an alternative: using topological data analysis to represent the full pattern of skin ridges and valleys directly, bypassing minutiae detection and the downstream matching pipeline. We apply persistent homology, a topological tool that tracks how loops in the ridge pattern form and fill in across spatial scales, producing multi-scale summaries of ridge geometry. We develop and compare a range of verification methods on a standard benchmark dataset, FVC2000 DB1. Even the simplest topological summaries, with no trained parameters, substantially outperform geometry-only baselines. A trained method achieves an AUC of 0.91, while an optimal-transport method excels at the strictest false-accept thresholds, suggesting they capture different aspects of the ridge pattern. Fusing these two approaches yields the best performance at every low false-accept threshold we examine. Our results establish that these topological summaries capture substantial fingerprint identity information, far more effective for verification than raw pixel-level geometry. Because the entire pipeline is openly specified, it offers a transparent complement to minutiae-based systems, and we provide a modular framework for constructing, evaluating, and combining topological verification methods.

Abstract

Explore similar work

Topological summaries of fingerprint ridge patterns carry identity information

AFID: A Unified Open Framework for Automated Fingermark Identification, Quality Assessment and Feature Extraction

PySIFT: GPU-Resident Deterministic SIFT for Deep Learning Vision Pipelines