Emanuela Marasco
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
Emanuela Marasco has a total of 40 co-authors. Here are the top co-authors by number of shared publications:
- Anudeep Vurity
George Mason University
3 shared publications - Javad Khodadoust
School of Engineering and Science, Tecnologico de Monterrey, Ciudad Lpez Mateos, Mexico
3 shared publications - Raúl Monroy
Tecnológico de Monterrey
3 shared publications - Karl Ricanek
University of North Carolina Wilmington
2 shared publications - Massimiliano Albanese
George Mason University
2 shared publications - Raghavendra Ramachandra
Norwegian University of Science and Technology (NTNU)
2 shared publications
Latest Publications
- 10 June 2026
- Published by Elsevier in Image and Vision Computing
- Vol. 173
Revisiting the machine unlearning ecosystem in vision: From afterthought to a lifecycle perspective
- 10 June 2026
- Published by Elsevier in Image and Vision Computing
- Vol. 173
Revisiting the machine unlearning ecosystem in vision: From afterthought to a lifecycle perspective
- 23 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biometrics, Behavior, and Identity Science
- Vol. 8 (3) , 431-444
Identity Reflections: Hyperspectral Skin Analysis via Reflectance Distance Learning With Context-Aware Contrastive Transformers
Hyperspectral Imaging (HSI) enables real-time analysis with rich spectral detail, offering a promising path to spoof-resistant AI-based identity verification. Leveraging this spectral information, we present a novel biometric approach that extracts and analyzes skin reflectance patterns from hyperspectral fingertip images. In prior work, hyperspectral analysis has primarily been applied to facial biometrics; however, the associated lighting requirements are often impractical or inconvenient for users in real-world scenarios. From a computational perspective, while deep learning has significantly advanced recognition using conventional imaging, effectively applying these frameworks to the rich and complex signals captured by HSI remains a critical open challenge. Convolutional Neural Networks (CNNs) often have difficulty preserving the sequential structure of spectral information. Vision transformers effectively capture global context through self-attention but often face challenges in modeling fine-grained local features due to the lack of convolutional inductive biases. To address existing limitations, this work introduces a novel integration of Context-Enriched Contrastive Loss (CECL) into two architectures: SpectralFormer, a ViT-based model that employs decoupled spatial–spectral processing to handle spectral signatures extracted from carefully selected spatial regions; and the Swin Transformer, which leverages a unified 3D joint representation. For the first time, SpectralFormer is employed for identity verification, leveraging group-wise spectral embeddings, inspired by ViT’s patch-based processing, across adjacent bands, along with intermediate skip connections to enhance the analysis of complex spectral data. The proposed approach was evaluated on three biometric databases, including the Mason FHSD-2022 hyperspectral fingertip dataset collected at George Mason University with 100 subjects. The code will be publicly available on GitHub.
- 23 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biometrics, Behavior, and Identity Science
- Vol. 8 (3) , 431-444
Identity Reflections: Hyperspectral Skin Analysis via Reflectance Distance Learning With Context-Aware Contrastive Transformers
Hyperspectral Imaging (HSI) enables real-time analysis with rich spectral detail, offering a promising path to spoof-resistant AI-based identity verification. Leveraging this spectral information, we present a novel biometric approach that extracts and analyzes skin reflectance patterns from hyperspectral fingertip images. In prior work, hyperspectral analysis has primarily been applied to facial biometrics; however, the associated lighting requirements are often impractical or inconvenient for users in real-world scenarios. From a computational perspective, while deep learning has significantly advanced recognition using conventional imaging, effectively applying these frameworks to the rich and complex signals captured by HSI remains a critical open challenge. Convolutional Neural Networks (CNNs) often have difficulty preserving the sequential structure of spectral information. Vision transformers effectively capture global context through self-attention but often face challenges in modeling fine-grained local features due to the lack of convolutional inductive biases. To address existing limitations, this work introduces a novel integration of Context-Enriched Contrastive Loss (CECL) into two architectures: SpectralFormer, a ViT-based model that employs decoupled spatial–spectral processing to handle spectral signatures extracted from carefully selected spatial regions; and the Swin Transformer, which leverages a unified 3D joint representation. For the first time, SpectralFormer is employed for identity verification, leveraging group-wise spectral embeddings, inspired by ViT’s patch-based processing, across adjacent bands, along with intermediate skip connections to enhance the analysis of complex spectral data. The proposed approach was evaluated on three biometric databases, including the Mason FHSD-2022 hyperspectral fingertip dataset collected at George Mason University with 100 subjects. The code will be publicly available on GitHub.
- 2 December 2025
- Published by Elsevier in Engineering Applications of Artificial Intelligence
- Vol. 165
Minutiae-based palm photo recognition using deep neural networks
- 2 December 2025
- Published by Elsevier in Engineering Applications of Artificial Intelligence
- Vol. 165
Minutiae-based palm photo recognition using deep neural networks
- 6 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biometrics, Behavior, and Identity Science
- Vol. 8 (1) , 122-136
PFPINet: An End-to-End Network for Partial Finger Photo Identification
Smartphones have become indispensable tools in daily life, serving various applications such as device unlocking, account login, online transactions, e-banking, e-health services, and access control. Recently, the use of smartphone cameras for fingerprint acquisition, known as finger photos, has gained attention as an alternative to dedicated fingerprint sensors due to its convenience, cost-effectiveness, and widespread availability. However, the quality of finger photos is significantly affected by various factors, including lighting conditions, camera resolution, finger or smartphone motion, out-of-focus capture, and the angle of finger placement. These factors can result in blurred or low-quality regions where ridge patterns cannot be reliably reconstructed. When certain regions are degraded, the remaining portion of the finger photo becomes partial, posing a significant challenge for recognition systems. In this paper, we propose PFPINet, a patch-based, end-to-end, single-to-multiple framework for partial finger photo identification. PFPINet comprises multiple modules, including three novel components introduced in this work: FreqQualNet, a deep convolutional neural network (DCNN) designed to assess the quality of finger photos; EnGAN, a generative adversarial network (GAN) developed to enhance image quality; and PaINet, a DCNN that incorporates a variant of k-means clustering for identification—unlike most existing approaches that focus solely on verification. Extensive experiments conducted on three finger photo datasets and two cross-database contactless fingerprint datasets demonstrate that PFPINet consistently outperforms current state-of-the-art (SOTA) methods.
- 6 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biometrics, Behavior, and Identity Science
- Vol. 8 (1) , 122-136
PFPINet: An End-to-End Network for Partial Finger Photo Identification
Smartphones have become indispensable tools in daily life, serving various applications such as device unlocking, account login, online transactions, e-banking, e-health services, and access control. Recently, the use of smartphone cameras for fingerprint acquisition, known as finger photos, has gained attention as an alternative to dedicated fingerprint sensors due to its convenience, cost-effectiveness, and widespread availability. However, the quality of finger photos is significantly affected by various factors, including lighting conditions, camera resolution, finger or smartphone motion, out-of-focus capture, and the angle of finger placement. These factors can result in blurred or low-quality regions where ridge patterns cannot be reliably reconstructed. When certain regions are degraded, the remaining portion of the finger photo becomes partial, posing a significant challenge for recognition systems. In this paper, we propose PFPINet, a patch-based, end-to-end, single-to-multiple framework for partial finger photo identification. PFPINet comprises multiple modules, including three novel components introduced in this work: FreqQualNet, a deep convolutional neural network (DCNN) designed to assess the quality of finger photos; EnGAN, a generative adversarial network (GAN) developed to enhance image quality; and PaINet, a DCNN that incorporates a variant of k-means clustering for identification—unlike most existing approaches that focus solely on verification. Extensive experiments conducted on three finger photo datasets and two cross-database contactless fingerprint datasets demonstrate that PFPINet consistently outperforms current state-of-the-art (SOTA) methods.
- 27 September 2025
- Published by Springer Nature
- p. 276-294
“I Forgot About You”: Exploring Multi-Label Unlearning (MLU) for Responsible Facial Recognition Systems
- 27 September 2025
- Published by Springer Nature
- p. 276-294
“I Forgot About You”: Exploring Multi-Label Unlearning (MLU) for Responsible Facial Recognition Systems
- 1 August 2025
- Published by Springer Nature
- p. 209-228
Hyperspectral Biometrics: Towards Sensing and Monitoring of Biochemical Characteristics
- 1 August 2025
- Published by Springer Nature
- p. 209-228
Hyperspectral Biometrics: Towards Sensing and Monitoring of Biochemical Characteristics
- Open Access
- 25 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 132649-132658
Pigment-Aware: Understanding Skin Tone Variations in Digital Image Analysis
Optical sensors, such as the RGB cameras embedded in smartphones, often fail to accurately capture the full spectrum of skin tones. As a result, individuals with darker skin may experience reduced performance in machine vision-based security systems. Insufficient attention to human diversity, including variations in skin tone, can contribute to biased training data and, subsequently, to disparities in AI systems, especially in biometric recognition. This paper highlights the need for more consistent and objective approaches to assessing skin tone, which are often treated subjectively or applied inconsistently. We address this issue by analyzing RGB finger photo data using colorimetric techniques to support the development of more inclusive machine vision systems.
- 25 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 132649-132658
Open AccessPigment-Aware: Understanding Skin Tone Variations in Digital Image Analysis
Optical sensors, such as the RGB cameras embedded in smartphones, often fail to accurately capture the full spectrum of skin tones. As a result, individuals with darker skin may experience reduced performance in machine vision-based security systems. Insufficient attention to human diversity, including variations in skin tone, can contribute to biased training data and, subsequently, to disparities in AI systems, especially in biometric recognition. This paper highlights the need for more consistent and objective approaches to assessing skin tone, which are often treated subjectively or applied inconsistently. We address this issue by analyzing RGB finger photo data using colorimetric techniques to support the development of more inclusive machine vision systems.
- 2 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biometrics, Behavior, and Identity Science
- Vol. 7 (4) , 668-680
A Unified Framework for Segmentation, Scaling, and Indexing of Contactless Fingerprints
Contactless fingerprint identification is gaining prominence due to its convenience, accessibility, security, and hygiene benefits over traditional contact-based methods. However, achieving accurate identification and efficiently searching large-scale databases remain challenging. In this paper, we introduce CFinSegNet, an encoder-decoder network for fingerprint segmentation that operates on preprocessed fingerprints and multiple image representations from diverse color spaces rather than the original fingerprint images. Additionally, we present FreqFinScale, a no-reference regression-based model designed to handle scaling variations using frequency representations. Furthermore, we propose FreqFinIndex, a deep neural network architecture for indexing and retrieving contactless fingerprints, which compresses high-dimensional frequency domain data into a compact latent space for fast and accurate search. Our experimental evaluation on five publicly available datasets demonstrates that our method significantly outperforms state-of-the-art approaches in accuracy and efficiency.
- 2 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biometrics, Behavior, and Identity Science
- Vol. 7 (4) , 668-680
A Unified Framework for Segmentation, Scaling, and Indexing of Contactless Fingerprints
Contactless fingerprint identification is gaining prominence due to its convenience, accessibility, security, and hygiene benefits over traditional contact-based methods. However, achieving accurate identification and efficiently searching large-scale databases remain challenging. In this paper, we introduce CFinSegNet, an encoder-decoder network for fingerprint segmentation that operates on preprocessed fingerprints and multiple image representations from diverse color spaces rather than the original fingerprint images. Additionally, we present FreqFinScale, a no-reference regression-based model designed to handle scaling variations using frequency representations. Furthermore, we propose FreqFinIndex, a deep neural network architecture for indexing and retrieving contactless fingerprints, which compresses high-dimensional frequency domain data into a compact latent space for fast and accurate search. Our experimental evaluation on five publicly available datasets demonstrates that our method significantly outperforms state-of-the-art approaches in accuracy and efficiency.
- 18 March 2025
- Published by Elsevier in Image and Vision Computing
- Vol. 158
Enhancing trust in Large Language Models for streamlined decision-making in military operations
- 18 March 2025
- Published by Elsevier in Image and Vision Computing
- Vol. 158
Enhancing trust in Large Language Models for streamlined decision-making in military operations
- Open Access
ColFigPhotoAttnNet: Reliable Finger Photo Presentation Attack Detection Leveraging Window-Attention on Color Spaces
Finger photo Presentation Attack Detection (PAD) can significantly strengthen smartphone device security. However, these algorithms are trained to detect certain types of attacks. Furthermore, they are designed to operate on images acquired by specific capture devices, leading to poor generalization and a lack of robustness in handling the evolving nature of mobile hardware. The proposed investigation is the first to systematically analyze the performance degradation of existing deep learning PAD systems, convolutional and transformers, in cross-capture device settings. In this paper, we introduce the ColFigPhotoAttnNet architecture designed based on window attention on color channels, followed by the nested residual network as the predictor to achieve a reliable PAD. Extensive experiments using various capture devices, including iPhone13 Pro, GooglePixel 3, Nokia C5, and OnePlusOne, were carried out to evaluate the performance of proposed and existing methods on three publicly available databases. The findings underscore the effectiveness of our approach.
Open AccessColFigPhotoAttnNet: Reliable Finger Photo Presentation Attack Detection Leveraging Window-Attention on Color Spaces
Finger photo Presentation Attack Detection (PAD) can significantly strengthen smartphone device security. However, these algorithms are trained to detect certain types of attacks. Furthermore, they are designed to operate on images acquired by specific capture devices, leading to poor generalization and a lack of robustness in handling the evolving nature of mobile hardware. The proposed investigation is the first to systematically analyze the performance degradation of existing deep learning PAD systems, convolutional and transformers, in cross-capture device settings. In this paper, we introduce the ColFigPhotoAttnNet architecture designed based on window attention on color channels, followed by the nested residual network as the predictor to achieve a reliable PAD. Extensive experiments using various capture devices, including iPhone13 Pro, GooglePixel 3, Nokia C5, and OnePlusOne, were carried out to evaluate the performance of proposed and existing methods on three publicly available databases. The findings underscore the effectiveness of our approach.