FVeinSyn: Synthetic Finger Vein Image Generator

Yifan Wang; Jie Gui; Adams Wai Kin Kong; Baosheng Yu; Changsheng Chen; Qi Li; Zhenan Sun; James Tin-Yau Kwok

cs.CVAug 27, 2026

FVeinSyn: Synthetic Finger Vein Image Generator

Authors: Yifan Wang, Jie Gui, Adams Wai Kin Kong, Baosheng Yu, Changsheng Chen, Qi Li, Zhenan Sun, James Tin-Yau Kwok, +1 more

Organizations: School of Cyber Science and Engineering, Southeast University, Nanjing 210000, China · Engineering Research Center of Blockchain Application, Supervision and Management (Southeast University), Ministry of Education; Purple Mountain Laboratories, Nanjing 210000, China · School of Computer Science and Engineering, Nanyang Technological University, Singapore 639798 · Lee Kong Chian School of Medicine, Nanyang Technological University, Singapore 639798 · Faculty of Engineering, Shenzhen MSU-BIT University, Shenzhen 518000, China · New Laboratory of Pattern Recognition and the State Key Laboratory of Multimodal Artificial Intelligence Systems, Institute of Automation, Chinese Academy of Sciences, Beijing 100190, China · Department of Computer Science and Engineering, The Hong Kong University of Science and Technology, Hong Kong 999077, China · VinUniversity, Hanoi, Vietnam 100000

Abstract

A major challenge in finger vein recognition is the lack of large-scale public datasets. Existing datasets contain few identities and limited samples per finger, restricting the advancement of deep learning-based methods. To address this, we propose FVeinSyn, a large-scale controllable synthetic data generation framework for finger vein. It explicitly decouples synthesis of vascular topology and imaging appearance to mitigate the limitations caused by insufficient training samples, such as inadequate identity diversity and restricted realism. Specifically: first, a finger vein identity generator models vascular topology under physiological and geometric constraints using stochastic L-systems, producing anatomically valid and identity-distinctive vascular patterns. Then, a cascaded region-aware GAN renders the topological maps into realistic near-infrared images. Finally, an intra-class diversity generator introduces geometric and optical perturbations to simulate realistic intra-class variations. Using FVeinSyn, we generated 500,000 images (10,000 vein identities, 50 samples per identity) and conducted extensive evaluations. Results show that FVeinSyn holds significant advantages in realism, identity diversity, vascular pattern consistency, and intra-class diversity. Models trained with FVeinSyn outperform real-data-only baselines a cross eight public datasets, achieving an average accuracy improvement of 27.43%. The code is available at: https://github.com/EvanWang98/Synthetic-Finger-Vein-Generator.

Figures & tables

Fig. 1: Comparison of publicly available datasets representing face [ 11 , 12 , 13 , 14 , 15 , 16 , 5 ] , fingerprint [ 17 , 18 , 19 , 20 , 21 ] , and finger vein [ 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 ] biometric modalities. FVeinSyn-5M (50,000 identities/5,000,000 samples) [ 32 ] and FVeinSyn (10,000 identities/500,000 samples) are synthetic finger vein datasets proposed in this work.

Panel of Figure 2 — Fig. 2: Traditional modeling (left): vessels are unsmooth and rigid, with unrealistic infrared appearance. Learning-based (right): noticeable artifacts are observed.

Fig. 3: Bifurcation parameters: diameters ( $d_{\mathrm{parent}}$ , $d_{\mathrm{left}}$ , $d_{\mathrm{right}}$ ), lengths ( $l_{\mathrm{parent}}$ , $l_{\mathrm{left}}$ , $l_{\mathrm{right}}$ ), volumetric flows ( $Q_{\mathrm{parent}},Q_{\mathrm{left}},Q_{\mathrm{right}}$ ) and bifurcation angles ( $\theta_{\mathrm{left}}$ , $\theta_{\mathrm{right}}$ ).

Fig. 4: Overview of the FVeinSyn framework for large-scale synthetic finger vein image generation. 1) Finger Vein Identity Generator, where unique vascular topologies are created. 2) Grayscale Finger Vein Image Render, performed by a Cascaded Region-Aware GAN (CascadedRA-GAN) to produce realistic images. 3) Intra-Class Diversity Generation, which applies geometric and optical perturbations to enhance variability within each identity.

Operator	Params	Range / distribution	Default	Notes
Translation	$t_{x}$	$U(-\tau_{w}W,\tau_{w}W)$	$\tau_{w}=0.05$	horizontal displacement
Translation	$t_{y}$	$U(-\tau_{h}H,\tau_{h}H)$	$\tau_{h}=0.05$	vertical displacement
Rotation	$\phi$	$U(-\phi_{\max},\phi_{\max})$	$\phi_{\max}=15^{\circ}$	rotation angle
Rolling	$\psi$	$U(-\psi_{\max},\psi_{\max})$	$\psi_{\max}=10^{\circ}$	rolling angle
Scaling	$s$	$U(1\pm s_{max})$	$s_{m}=0.1$	scale factor
Exposure	$v_{0}$	$U(-v_{max},v_{max})$	$v_{max}=2$	exposure intensity

TABLE I: Parameterization for ICDG operators.

\mathcal{U}(a,b)

denotes the uniform distribution over the interval

[a,b]

W

H

are width and height,

\tau

is the translation factor.

Fig. 5: Comparison between real and synthetic finger vein images generated by FVeinSyn. From left to right: SDUMLA-FV, UTFVP, MMCBNU_6000, PLUS-FV3, FV-USM, HKPU-FV, SCUT-FV3, THU-FV3, and FVeinSyn generate synthetic finger vein images (two columns).

Fig. 6: Vein pattern extraction results on real and synthetic finger vein images using classical methods, including RLT, MC, Gabor filter, and IUWT. From top to bottom: FVeinSyn-generated synthetic finger vein images (two rows), SDUMLA-FV, UTFVP, MMCBNU_6000, PLUS-FV3, FV-USM, HKPU-FV, SCUT-FV3, and THU-FV3.

Metrics	Real Datasets									Synthetic Datasets
Metrics	SDUMLA-FV	UTFVP	MMCBNU_6000	PLUS-FV3	FV-USM	HKPU-FV	SCUT-FV	THU-FV3	Avg.	Ou et al. [ 38 ]	BTUSyn [ 37 ]	FVeinSyn(Our)
GCF	$2.02$	$2.50$	$2.43$	$2.26$	$1.35$	$2.73$	$1.99$	$2.33$	$2.16$	$1.64$	$1.95$	$2.40$
Entropy	$6.62$	$7.16$	$6.84$	$6.33$	$6.08$	$6.88$	$6.44$	$6.58$	$6.64$	$6.45$	$6.67$	$7.06$
Tenengrad	$18.96$	$26.31$	$11.79$	$26.92$	$7.96$	$19.83$	$11.87$	$25.26$	$18.73$	$15.61$	$13.09$	$24.12$
NIQE	$30.28$	$31.18$	$32.74$	$24.84$	$31.86$	$31.53$	$31.53$	$33.27$	$31.16$	$27.02$	$29.90$	$28.59$
BRISQUE	$53.11$	$46.09$	$67.52$	$57.56$	$61.46$	$67.08$	$69.00$	$35.89$	$56.94$	$58.94$	$88.37$	$44.72$

TABLE II: Comparison of image quality metrics across different finger vein datasets.

Fig. 7: Global contrast factor, Entropy, and Tenengrad gradient distributions of image quality metrics for real and synthetic finger vein datasets.

Method	Region	Avg. Realism Rating	Std. Dev.
BTUSyn [ 37 ]	ROI	$3.74$	$0.60$
Ou et al. [ 38 ]	ROI	$4.61$	$0.78$
FVeinSyn (Our)	ROI	$5.65$	$0.53$
FVeinSyn (Our)	Full Finger	$7.11$	$0.56$

TABLE III: Average realism ratings of different synthesis methods (9 = Most Realistic, 1 = Least Realistic)

Method	Region	Average Rating (%)	Std. Dev. (%)
BTUSyn [ 37 ]	ROI	$20.50$	$14.80$
Ou et al. [ 38 ]	ROI	$40.84$	$16.23$
FVeinSyn (Our)	ROI	$82.97$	$10.98$
FVeinSyn (Our)	Full Finger	$90.54$	$8.03$

TABLE IV: percentage classified as real per synthesis data

Method	Configs			Performance (TAR@FAR=1e-3) $\uparrow$
Method	#IDs	#per ID	#Image	SDUMLA-FV	UTFVP	MMCBNU_6000	PLUS-FV3	FV-USM	HKPU-FV	SCUT-FV3	THU-FV3	Avg.
MC	-	-	-	$0.3117$	$0.5593$	$0.7669$	$0.6895$	$0.7632$	$0.3066$	$0.6261$	$0.5016$	$0.5656$
RLT	-	-	-	$0.4006$	$0.7093$	$0.7650$	$0.5363$	$0.8406$	$0.4753$	$0.5853$	$0.5788$	$0.6114$
IUWT	-	-	-	$0.4273$	$0.8269$	$0.7951$	$0.6190$	$0.7955$	$0.5758$	$0.5775$	$0.5405$	$0.6447$
GF	-	-	-	$0.3820$	$0.7232$	$0.8640$	$0.5104$	$0.8585$	$0.4743$	$0.5721$	$0.6158$	$0.6250$
Real data	1.9k	-	18k	$0.4602$	$0.7593$	$0.8385$	$0.5385$	$0.7709$	$0.6447$	$0.5504$	$0.5799$	$0.6428$

TABLE V: Performance comparison of various enhancement or synthesis methods for finger vein recognition

Fig. 8: Comparison of receiver operating characteristic (ROC) curves and equal error rates (EERs) for FVeinSyn and other representative methods under the open-set finger vein recognition protocol.

Panel of Table 14 — TABLE VI: Cross-dataset evaluation using or not using FVeinSyn for pre-training. ✔indicates that pre-training with synthetic data is used (TAR@FAR=1e-3) $\uparrow$

Panel of Table 15 — TABLE VII: Performance comparison with increasing training samples (TAR@FAR=1e-3) $\uparrow$

Dataset		SDUMLA-FV	UTFVP	MMCBNU_6000	PLUS-FV3	FV-USM	HKPU-FV	SCUT-FV3	THU-FV3
$N=1$	Real data	0.4736	0.2815	0.6874	0.2655	0.4365	0.2359	0.1034	0.4902
	Ou et al. [ 38 ]	$0.5319$	$0.3380$	$0.7338$	0.3928	$0.4459$	$0.2649$	$0.0576$	$0.5628$
	BTUSyn [ 37 ]	0.5855	0.5333	0.8376	0.3834	0.6125	0.3849	0.1023	0.6213
	FVeinSyn (Our)	0.7220	0.7315	0.8710	0.8874	0.5650	0.8096	0.2626	0.5826
$N=2$	Real data	0.5440	0.4333	0.7815	0.3823	0.6490	0.3889	0.4070	0.5681
	Ou et al. [ 38 ]	$0.5810$	$0.4583$	$0.8184$	$0.5367$	$0.5384$	$0.3462$	$0.3030$	$0.6123$

TABLE VIII: Performance comparison of one/few-shot finger vein recognition (TAR@FAR=1e-3)

\uparrow

Fig. 9: Comparison of receiver operating characteristic (ROC) curves and equal error rates (EERs) for FVeinSyn and other representative methods under the limited finger vein sample recognition protocol.

Dataset	$U_{class}$	$C_{intra}$	$D_{intra}$
Ou et al. [ 38 ]	$24.57$	–	–
BTUSyn [ 37 ]	$19.19$	$99.98$	$3.28$
FVeinSyn (Our)	$99.83$	$100.00$	$89.67$

TABLE IX: Dataset Evaluation of Synthetic Finger Vein

Panel of Figure 19 — Fig. 10: Genuine/Impostor matching score distributions across multiple real datasets and the FVeinSyn synthetic dataset.

Fig. 11: Distribution of the number of intra-class variations in 500,000 FVeinSyn generated samples.

Intra-Class Variation	SDUMLA-FV	UTFVP	MMCBNU_60000	PLUS-FV3	FV-USM	HKPU-FV	SCUT-FV	THU-FV3	Avg.
Real data	$0.4602$	$0.7593$	$0.8385$	$0.5385$	$0.7709$	$0.6447$	$0.5504$	$0.5799$	$0.6428$
FVeinSyn (Mixed)	$0.9402$	$0.9778$	$0.9538$	$0.9132$	$0.9551$	$0.9655$	$0.7078$	$0.9233$	$0.9171$
Geometric Perturbations
Shift	$0.8994$	$0.9528$	$0.9395$	$0.8749$	$0.9009$	$0.9264$	$0.6642$	$0.8736$	0.8789
Rotation	$0.9283$	$0.9528$	$0.9209$	$0.8086$	$0.8875$	$0.8850$	$0.6965$	$0.8371$	0.8646
Rolling	$0.9107$	$0.9509$	$0.8901$	$0.8619$	$0.8726$	$0.8545$	$0.6675$	$0.7963$	$0.8506$

TABLE X: Impact of different intra-class variations on recognition performance across datasets (TAR@FAR=

10^{-3}

)

\uparrow

Dataset	PT w/ Syn		Gender		Age			Race			Blood Type
Dataset	✗	✔	Male	Female	$\leq$ 20	21–40	$\geq$ 41	Asian	Caucasian	African	A	B	AB	O
UTFVP	$0.7593$	$0.9778$	$0.9822$	$0.9222$	$0.9815$	$0.9759$	$0.9514$	-	-	-	-	-	-	-
MMCBNU_6000	$0.8385$	$0.9538$	$0.9458$	$0.9876$	$1.0000$	$0.9494$	$0.9605$	$0.9542$	$0.9531$	$0.9741$	$0.9407$	$0.9874$	$0.9704$	$0.9663$

TABLE XI: Demographic Fairness Analysis of Recognition Performance with Synthetic Data (TAR@FAR=

10^{-3}

)

\uparrow

Step	Identity Generation	Render	Diversity Generation	Total
Time (ms)	$285.6$	$7.9$	$0.7$	$294.2$

TABLE XII: Time consumption of Finger Vein Image Sample using FVeinSyn

Explore similar work

OpenVeinNet: Robust Open-Set Finger Vein Verification with Dynamic Snake Convolution and Graph Learning
Aug 26, 2026Sushrut Patwardhan, Raghavendra RamachandraBiometric Identification Open-Set Recognition
SAFE Centre, Department of Information Security and Communication Technology (IIK), Norwegian University of Science and Technology (NTNU), Gjøvik, Norway.
MAGE-Vein: Multi-Instance Age and Gender Estimation from Finger Vein Images
Jul 23, 2026Katsuki Tanaka, Koichi Ito, Takafumi Aoki +4
Graduate School of Information Sciences, Tohoku University, Japan. · Hitachi, Ltd., Japan.
Adaptive Adversarial Augmentation for Controllable Face Synthesis
Oct 8, 2026Saransh Suri, Shivang Agarwal, Mayank Vatsa +1Synthetic Data Augmentation Data Augmentation
Manipal University Jaipur, India · Indian Institute of Technology Jodhpur, India · Birla Institute of Technology and Science Pilani Dubai Campus, Dubai, United Arab Emirates

Aug 26, 2026J/K move · Enter open · S save

OpenVeinNet: Robust Open-Set Finger Vein Verification with Dynamic Snake Convolution and Graph Learning

Sushrut Patwardhan, Raghavendra Ramachandra

SAFE Centre, Department of Information Security and Communication Technology (IIK), Norwegian University of Science and Technology (NTNU), Gjøvik, Norway.

Biometric Identification Open-Set Recognition Convolutional Neural Networks

Open paper AskPDF arXiv

Finger vein verification is a promising biometric modality for secure authentication because vascular patterns are internal, difficult to observe externally, and relatively resistant to presentation attacks. However, reliable verification remains challenging in open-set settings, where test identities are unseen during training and non-enrolled probes must be rejected at inference. This paper presents OpenVeinNet, a finger vein verification framework designed for cross-dataset and open-set evaluation. The proposed model combines Dynamic Snake Convolution with graph-based feature modelling. Dynamic Snake Convolution extracts local curvilinear and tubular vein structures using adaptive sampling, while the graph convolutional backbone models long-range topological relationships between vein regions. To improve the discriminative quality of the embedding space, we introduce a Centroid Angular Hybrid Loss, which jointly encourages intra-class compactness and inter-class angular separation for cosinesimilaritybased verification. Experiments are conducted on five public finger vein datasets: FV-300, MMCBNU, FV-USM, PolyU, and VERA. The method is evaluated using leaveonedatasetout training under both enrolmentbased unknownrejection and fullsubject verification protocols, and is compared with handcrafted and recent deep learning-based baselines. The results show that OpenVeinNet achieves strong cross-dataset generalisation, consistently low equal error rates, and competitive true accept rates at fixed false accept rate operating points. Ablation studies further confirm the individual and combined contributions of adaptive tubular feature extraction, graph-based relational modelling, and the proposed loss function. These findings indicate that explicitly modelling local vein geometry, global vascular relationships, and angularly compact embeddings is effective for openset finger vein verification.

Abstract

Explore similar work

OpenVeinNet: Robust Open-Set Finger Vein Verification with Dynamic Snake Convolution and Graph Learning

MAGE-Vein: Multi-Instance Age and Gender Estimation from Finger Vein Images

Adaptive Adversarial Augmentation for Controllable Face Synthesis