Raffaele Cappelli
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Raffaele Cappelli has a total of 34 co-authors. Here are the top co-authors by number of shared publications:
Davide Maltoni
25 shared publications- Matteo Ferrara
University of Bologna, Bologna, Italy
20 shared publications D. Maltoni
Scienze dell’Informazione - Università di Bologna, via Sacchi 3, 47023 Cesena, Italy, DEIS – Viale Risorgimento, 2 – 40126 Bologna
11 shared publicationsD. Maio
Univ. di Bologna, Cesena
9 shared publicationsDario Maio
Biometric System Laboratory-DEIS, University of Bologna, Cesena (FC), Italy
9 shared publications- Davide Maltoni
Department of Computer Science and Engineering, University of Bologna, Italy
5 shared publications
Latest Publications
- Open Access
LEADER: Lightweight End-to-End Attention-Gated Dual Autoencoder for Robust Minutiae Extraction
Minutiae extraction, a fundamental stage in fingerprint recognition, is increasingly shifting toward deep learning. However, truly end-to-end methods that eliminate separate preprocessing and postprocessing steps remain scarce. This paper introduces LEADER (Lightweight End-to-end Attention-gated Dual autoencodER), a neural network that maps raw fingerprint images to minutiae descriptors, including location, direction, and type. The proposed architecture integrates non-maximum suppression and angular decoding to enable complete end-to-end inference using only 0.9M parameters. It employs a novel "Castle-Moat-Rampart" ground-truth encoding and a dual-autoencoder structure, interconnected through an attention-gating mechanism. Experimental evaluations demonstrate state-of-the-art accuracy on plain fingerprints and robust cross-domain generalization to latent impressions. Specifically, LEADER attains a 34% higher F1-score on the NIST SD27 dataset compared to specialized latent minutiae extractors. Sample-level analysis on this challenging benchmark reveals an average rank of 2.07 among all compared methods, with LEADER securing the first-place position in 47% of the samples-more than doubling the frequency of the second-best extractor. The internal representations learned by the model align with established fingerprint domain features, such as segmentation masks, orientation fields, frequency maps, and skeletons. Inference requires 15ms on GPU and 322ms on CPU, outperforming leading commercial software in computational efficiency. The source code and pre-trained weights are publicly released to facilitate reproducibility.
Open AccessLEADER: Lightweight End-to-End Attention-Gated Dual Autoencoder for Robust Minutiae Extraction
Minutiae extraction, a fundamental stage in fingerprint recognition, is increasingly shifting toward deep learning. However, truly end-to-end methods that eliminate separate preprocessing and postprocessing steps remain scarce. This paper introduces LEADER (Lightweight End-to-end Attention-gated Dual autoencodER), a neural network that maps raw fingerprint images to minutiae descriptors, including location, direction, and type. The proposed architecture integrates non-maximum suppression and angular decoding to enable complete end-to-end inference using only 0.9M parameters. It employs a novel "Castle-Moat-Rampart" ground-truth encoding and a dual-autoencoder structure, interconnected through an attention-gating mechanism. Experimental evaluations demonstrate state-of-the-art accuracy on plain fingerprints and robust cross-domain generalization to latent impressions. Specifically, LEADER attains a 34% higher F1-score on the NIST SD27 dataset compared to specialized latent minutiae extractors. Sample-level analysis on this challenging benchmark reveals an average rank of 2.07 among all compared methods, with LEADER securing the first-place position in 47% of the samples-more than doubling the frequency of the second-best extractor. The internal representations learned by the model align with established fingerprint domain features, such as segmentation masks, orientation fields, frequency maps, and skeletons. Inference requires 15ms on GPU and 322ms on CPU, outperforming leading commercial software in computational efficiency. The source code and pre-trained weights are publicly released to facilitate reproducibility.
- Open Access
- 16 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 153605-153617
No Feature Left Behind: Filling the Gap in Fingerprint Frequency Estimation
Fingerprint recognition remains a cornerstone of biometric identification systems. However, the critical task of ridge-line frequency estimation has been surprisingly underexplored. To address this research gap, I introduce the first publicly available benchmark for fingerprint frequency estimation, a pioneering resource that includes both high- and low-quality fingerprints with meticulously labeled ground truth features: segmentation masks, orientation fields, and frequency maps. Furthermore, two novel frequency estimation methods are proposed: one that significantly enhances a well-known frequency estimation method based on traditional image processing techniques and another that, for the first time, applies deep learning to this context. Experimental results on the new benchmark demonstrate that both methods surpass existing state-of-the-art techniques, particularly in challenging low-quality fingerprint scenarios. By providing an open-source implementation and a comprehensive benchmark, this work sets a new standard for the evaluation of frequency estimation methods, fostering further research and development in this crucial area of fingerprint recognition.
- 16 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 153605-153617
Open AccessNo Feature Left Behind: Filling the Gap in Fingerprint Frequency Estimation
Fingerprint recognition remains a cornerstone of biometric identification systems. However, the critical task of ridge-line frequency estimation has been surprisingly underexplored. To address this research gap, I introduce the first publicly available benchmark for fingerprint frequency estimation, a pioneering resource that includes both high- and low-quality fingerprints with meticulously labeled ground truth features: segmentation masks, orientation fields, and frequency maps. Furthermore, two novel frequency estimation methods are proposed: one that significantly enhances a well-known frequency estimation method based on traditional image processing techniques and another that, for the first time, applies deep learning to this context. Experimental results on the new benchmark demonstrate that both methods surpass existing state-of-the-art techniques, particularly in challenging low-quality fingerprint scenarios. By providing an open-source implementation and a comprehensive benchmark, this work sets a new standard for the evaluation of frequency estimation methods, fostering further research and development in this crucial area of fingerprint recognition.
- Open Access
- 16 April 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 55998-56018
Exploring the Power of Simplicity: A New State-of-the-Art in Fingerprint Orientation Field Estimation
Fingerprint recognition is an important tool for personal identification due to its versatility, user-friendliness, and accuracy. Fingerprint orientation field estimation, a crucial step in fingerprint feature extraction, significantly impacts recognition performance. While numerous methods have been proposed, achieving state-of-the-art accuracy often comes at the cost of increased complexity, hindering practical implementation. This work addresses this challenge by introducing two novel, simple, and efficient fingerprint orientation field estimation methods: GBFOE and SNFOE. Both methods adhere to the KISS (Keep It Simple and Straightforward) principle, achieving remarkable performance on publicly available benchmarks. GBFOE outperforms all local methods and rivals more complex approaches, while SNFOE establishes itself as a new state-of-the-art, achieving the highest accuracy on all datasets in both evaluated benchmarks. Surpassing methods designed specifically for latent fingerprints, SNFOE demonstrates exceptional performance on this challenging task within the evaluated benchmarks, highlighting its generalizability despite not being trained on such data. These results underline the potential of simple and efficient methods in fingerprint orientation field estimation, paving the way for practical and resource-efficient fingerprint recognition systems. An open-source Python implementation of both methods is available, fostering further research and development in this field.
- 16 April 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 55998-56018
Open AccessExploring the Power of Simplicity: A New State-of-the-Art in Fingerprint Orientation Field Estimation
Fingerprint recognition is an important tool for personal identification due to its versatility, user-friendliness, and accuracy. Fingerprint orientation field estimation, a crucial step in fingerprint feature extraction, significantly impacts recognition performance. While numerous methods have been proposed, achieving state-of-the-art accuracy often comes at the cost of increased complexity, hindering practical implementation. This work addresses this challenge by introducing two novel, simple, and efficient fingerprint orientation field estimation methods: GBFOE and SNFOE. Both methods adhere to the KISS (Keep It Simple and Straightforward) principle, achieving remarkable performance on publicly available benchmarks. GBFOE outperforms all local methods and rivals more complex approaches, while SNFOE establishes itself as a new state-of-the-art, achieving the highest accuracy on all datasets in both evaluated benchmarks. Surpassing methods designed specifically for latent fingerprints, SNFOE demonstrates exceptional performance on this challenging task within the evaluated benchmarks, highlighting its generalizability despite not being trained on such data. These results underline the potential of simple and efficient methods in fingerprint orientation field estimation, paving the way for practical and resource-efficient fingerprint recognition systems. An open-source Python implementation of both methods is available, fostering further research and development in this field.
- Open Access
- 21 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 144530-144544
Unveiling the Power of Simplicity: Two Remarkably Effective Methods for Fingerprint Segmentation
Accurate fingerprint segmentation is crucial for reliable fingerprint recognition systems. This paper presents two novel segmentation methods, GMFS and SUFS, inspired by the KISS (Keep It Simple and Straightforward) principle. Both methods, evaluated on a public benchmark and compared to eighteen state-of-the-art approaches, excel in terms of accuracy, while maintaining simplicity and computational efficiency. GMFS utilizes a single handcrafted feature for straightforward yet effective fingerprint segmentation, achieving superior performance compared to previously reported traditional methods. SUFS employs a simplified U-net architecture for end-to-end segmentation, demonstrating remarkable performance: it achieves an average classification error rate of 1.51% across the entire benchmark, with an improvement of over 40% compared to the previously best-performing method. Furthermore, despite being trained on a relatively small dataset, it exhibits significant generalization capabilities, effectively segmenting fingerprints from very different acquisition technologies without requiring fine-tuning. An open-source Python implementation of both methods is available, fostering further research and development in the field of fingerprint recognition.
- 21 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 144530-144544
Open AccessUnveiling the Power of Simplicity: Two Remarkably Effective Methods for Fingerprint Segmentation
Accurate fingerprint segmentation is crucial for reliable fingerprint recognition systems. This paper presents two novel segmentation methods, GMFS and SUFS, inspired by the KISS (Keep It Simple and Straightforward) principle. Both methods, evaluated on a public benchmark and compared to eighteen state-of-the-art approaches, excel in terms of accuracy, while maintaining simplicity and computational efficiency. GMFS utilizes a single handcrafted feature for straightforward yet effective fingerprint segmentation, achieving superior performance compared to previously reported traditional methods. SUFS employs a simplified U-net architecture for end-to-end segmentation, demonstrating remarkable performance: it achieves an average classification error rate of 1.51% across the entire benchmark, with an improvement of over 40% compared to the previously best-performing method. Furthermore, despite being trained on a relatively small dataset, it exhibits significant generalization capabilities, effectively segmenting fingerprints from very different acquisition technologies without requiring fine-tuning. An open-source Python implementation of both methods is available, fostering further research and development in the field of fingerprint recognition.
- Open Access
- 29 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biometrics, Behavior, and Identity Science
- Vol. 5 (4) , 476-485
Detecting Double-Identity Fingerprint Attacks
Double-identity biometrics, that is the combination of two subjects features into a single template, was demonstrated to be a serious threat against existing biometric systems. In fact, well-synthetized samples can fool state-of-the-art biometric verification systems, leading them to falsely accept both the contributing subjects. This work proposes one of the first techniques to defy existing double-identity fingerprint attacks. The proposed approach inspects the regions where the two aligned fingerprints overlap but minutiae cannot be consistently paired. If the quality of these regions is good enough to minimize the risk of false or miss minutiae detection, then the alarm score is increased. Experimental results carried out on two fingerprint databases, with two different techniques to generate double-identity fingerprints, validate the effectiveness of the proposed approach.
- 29 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biometrics, Behavior, and Identity Science
- Vol. 5 (4) , 476-485
Open AccessDetecting Double-Identity Fingerprint Attacks
Double-identity biometrics, that is the combination of two subjects features into a single template, was demonstrated to be a serious threat against existing biometric systems. In fact, well-synthetized samples can fool state-of-the-art biometric verification systems, leading them to falsely accept both the contributing subjects. This work proposes one of the first techniques to defy existing double-identity fingerprint attacks. The proposed approach inspects the regions where the two aligned fingerprints overlap but minutiae cannot be consistently paired. If the quality of these regions is good enough to minimize the risk of false or miss minutiae detection, then the alarm score is increased. Experimental results carried out on two fingerprint databases, with two different techniques to generate double-identity fingerprints, validate the effectiveness of the proposed approach.
- 24 February 2023
- Published by Springer Nature
- p. 3-15
Introduction to Presentation Attack Detection in Fingerprint Biometrics
- 24 February 2023
- Published by Springer Nature
- p. 3-15
Introduction to Presentation Attack Detection in Fingerprint Biometrics
- Open Access
Introduction to Presentation Attack Detection in Fingerprint Biometrics
This chapter provides an introduction to Presentation Attack Detection (PAD) in fingerprint biometrics, also coined anti-spoofing, describes early developments in this field, and briefly summarizes recent trends and open issues.
Open AccessIntroduction to Presentation Attack Detection in Fingerprint Biometrics
This chapter provides an introduction to Presentation Attack Detection (PAD) in fingerprint biometrics, also coined anti-spoofing, describes early developments in this field, and briefly summarizes recent trends and open issues.
- 1 March 2018
- Published by Institution of Engineering and Technology (IET)
- p. 169-191
Generating synthetic fingerprints
Synthetic fingerprint generation (SFinGe) techniques and associated tools (e.g., SFinGe) were introduced more than 15 years ago [1]. The main aim was to generate large databases for performance evaluation without allocating huge amount of resources for acquisition campaigns and, at the same time, to conform with the privacy directives that in many countries limit the exchange of biometric data. While the original scope remains central today, since the generation of very large synthetic dataset is crucial to predict accuracy on very large scenarios, new security needs (such as detecting altered fingerprints) and algorithms improvements (supervised learning approaches) are continuously renewing interest in the generation of synthetic fingerprints.
- 1 March 2018
- Published by Institution of Engineering and Technology (IET)
- p. 169-191
Generating synthetic fingerprints
Synthetic fingerprint generation (SFinGe) techniques and associated tools (e.g., SFinGe) were introduced more than 15 years ago [1]. The main aim was to generate large databases for performance evaluation without allocating huge amount of resources for acquisition campaigns and, at the same time, to conform with the privacy directives that in many countries limit the exchange of biometric data. While the original scope remains central today, since the generation of very large synthetic dataset is crucial to predict accuracy on very large scenarios, new security needs (such as detecting altered fingerprints) and algorithms improvements (supervised learning approaches) are continuously renewing interest in the generation of synthetic fingerprints.
- 1 June 2015
- Published by Elsevier in Information Sciences
- Vol. 306, 1-20
Large-scale fingerprint identification on GPU
- 1 June 2015
- Published by Elsevier in Information Sciences
- Vol. 306, 1-20
Large-scale fingerprint identification on GPU