Abstract
Multimodal biometrics are able to improve the accuracy and security of authentication by integrating more than one biometric characteristic, minimizing errors, and maximizing the resistance to attacks. The primary drawback of multimodal biometric verification is the complexity of the systems that are introduced by multiple sensors, more computing, and fusion issues. Multimodal feature extraction methods are inadequate in traditional feature extraction methods as they generate modality-specific, handcrafted representations which are not robust, compatible and discriminative enough to support effective feature-level fusion. Deep learning feature extractors produce robust, discriminative, and fusion-friendly representations which are very important in multimodal biometric authentication systems to enhance accuracy and reliability. Trust and confidence are crucial in multimodal biometric authentication systems utilizing deep learning, as the models operate as black boxes, handle irreversible biometric data, and make high-impact security decisions. This motivates the development of a secure, explainable, multimodal biometric authentication framework. The proposed system is a privacy-preserving and explainable multimodal biometric solution that combines deep learning, trust-adaptive fusion, and encrypted domain matching. It utilizes MobileNet for extracting discriminative features. A Trust Adaptive Fusion (TAF) Strategy adjusts the contribution of each modality based on its quality or confidence, enhancing the robustness against the noisy inputs. The fused features are secured using the Cheon-Kim-Kim-Song (CKKS) homomorphic encryption, without revealing the raw biometric data. Transparency is enhanced with the help of the Grad-CAM, which provides interpretability of the model’s decision. The proposed system is evaluated on the CASIA-FaceV5 and CASIA-FingerprintV5 datasets, demonstrates the low error rate of 0.0038 on fused feature representation.
Data availability
The datasets used in this study are publicly available biometric datasets. The CASIA-Face V5 dataset is publicly available via Figshare at https://doi.org/10.6084/m9.figshare.26509591. The CASIA-Fingerprint V5 dataset is publicly available from the Institute of Automation, Chinese Academy of Sciences at http://english.ia.cas.cn/rs/sd/201611/t20161123_170932.html.
References
Gawande, U. & Golhar, Y. Biometric security system: A rigorous review of unimodal and multimodal biometrics techniques. Int. J. Biom. 10, 142–175 (2018).
Wild, P., Radu, P., Chen, L. & Ferryman, J. Robust multimodal face and fingerprint fusion in the presence of spoofing attacks. Pattern Recognit. 50, 17–25 (2016).
Trigueros, D. S., Meng, L. & Hartnett, M. Face recognition: From traditional to deep learning methods. arXiv preprint arXiv:1811.00116 (2018).
Kumar, T. A. & Ilango, S. Robust forgery detection via ensemble methods using intuitionistic fuzzy LBP and sift features. Int. J. Comput. Appl. https://doi.org/10.1080/1206212X.2025.2469909 (2025).
Ametefe, D. S. et al. Enhancing fingerprint authentication: A systematic review of liveness detection methods against presentation attacks. J. Inst. Eng. (India): Ser. B 105, 1451–1467 (2024).
Adjabi, I., Ouahabi, A., Benzaoui, A. & Taleb-Ahmed, A. Past, present, and future of face recognition: A review. Electronics 9, 1188 (2020).
Sundararajan, K. & Woodard, D. L. Deep learning for biometrics: A survey. ACM Comput. Surv. (CSUR) 51, 1–34 (2018).
Zhao, Z. & Kumar, A. Improving periocular recognition by explicit attention to critical regions in deep neural network. IEEE Trans. Inf. Forensics Secur. 13, 2937–2952 (2018).
Acar, A., Aksu, H., Uluagac, A. S. & Conti, M. A survey on homomorphic encryption schemes: Theory and implementation. ACM Comput. Surv. (Csur) 51, 1–35 (2018).
Rivest, R. L. Cryptography and machine learning. In International Conference on the Theory and Application of Cryptology, 427–439 (Springer, 1991).
Tourky, D., ElKawkagy, M. & Keshk, A. Homomorphic encryption the “holy grail” of cryptography. In 2016 2nd IEEE International Conference on Computer and Communications (ICCC), 196–201 (IEEE, 2016).
Chander, B., John, C., Warrier, L. & Gopalakrishnan, K. Toward trustworthy artificial intelligence (tai) in the context of explainability and robustness. ACM Comput. Surv. 57, 1–49 (2025).
Barni, M., Droandi, G., Lazzeretti, R. & Pignata, T. Semba: Secure multi-biometric authentication. IET Biom. 8, 411–421 (2019).
Walia, G. S., Singh, T., Singh, K. & Verma, N. Robust multimodal biometric system based on optimal score level fusion model. Expert. Syst. Appl. 116, 364–376 (2019).
Dwivedi, R. & Dey, S. Score-level fusion for cancelable multi-biometric verification. Pattern Recognit. Lett. 126, 58–67 (2019).
Rathgeb, C., Gomez-Barrero, M., Busch, C., Galbally, J. & Fierrez, J. Towards cancelable multi-biometrics based on bloom filters: A case study on feature level fusion of face and iris. In 3rd international workshop on biometrics and forensics (IWBF 2015), 1–6 (IEEE, 2015).
Aleem, S., Yang, P., Masood, S., Li, P. & Sheng, B. An accurate multi-modal biometric identification system for person identification via fusion of face and finger print. World Wide Web 23, 1299–1317 (2020).
Vallabhadas, D. K., Sandhya, M., Reddy, S. D., Satwika, D. & Prashanth, G. L. Biometric template protection based on a cancelable convolutional neural network over iris and fingerprint. Biomed. Signal Process. Control. 91, 106006 (2024).
Li, Y. et al. A cancelable multi-biometric system based on the feature-level fusion of fingerprint and finger vein. Multimed. Tools Appl. 84, 24765–24787 (2025).
Sasikala, T. Multimodal secure biometrics using attention efficient-net hash compression framework. Digit. Signal Process. 160, 105018 (2025).
Dang, T. M. et al. Avet: A novel transform function to improve cancellable biometrics security. IEEE Trans. Inf. Forensics Secur. 18, 758–772 (2022).
Purohit, H. & Ajmera, P. K. Optimal feature level fusion for secured human authentication in multimodal biometric system. Mach. Vis. Appl. https://doi.org/10.1007/s00138-020-01146-6 (2021).
Vijay, M. & Indumathi, G. Deep belief network-based hybrid model for multimodal biometric system for futuristic security applications. J. Inf. Secur. Appl. 58, 102707. https://doi.org/10.1016/j.jisa.2020.102707 (2021).
mehdi Cherrat, E., Alaoui, R. & Bouzahir, H. Convolutional neural networks approach for multimodal biometric identification system using the fusion of fingerprint, finger-vein and face images. PeerJ Comput. Sci. 6, e248 (2020).
Mwaura, G. W., Mwangi, W. & Otieno, C. Multimodal biometric system: Fusion of face and fingerprint biometrics at match score fusion level. Int. J. Sci. Technol. Res. (2017).
Kazi, M. et al. Face, fingerprint, and signature based multimodal biometric system using score level and decision level fusion approaches. IETE J. Res. 70, 3703–3722. https://doi.org/10.1080/03772063.2023.2217784 (2024).
Batouche, A., Meshoul, S., Shaiba, H. & Batouche, M. A novel approach to enhanced cancelable multi-biometrics personal identification based on incremental deep learning. Comput. Mater. Continua 83 (2025).
Zhou, Z., Liu, Y., Zhu, X., Zhang, S. & Liu, Z. Privacy-preserving cancelable multi-biometrics for identity information management. Inf. Process. Manag. 62, 103869 (2025).
Zhao, G., Jiang, Q., Wang, D., Ma, X. & Li, X. Deep hashing based cancelable multi-biometric template protection. IEEE Trans. Dependable Secur. Comput. 21, 3751–3767 (2023).
Naeem, E. A. et al. Efficient cancelable authentication system based on DRPE and adaptive filter. Multimed. Tools Appl. 83, 76131–76175 (2024).
Elsheikh, A. G. et al. Application of mace filter with DRPE for cancelable biometric authentication. J. Opt. 53, 101–116 (2024).
Wang, Y., Shi, D. & Zhou, W. Convolutional neural network approach based on multimodal biometric system with fusion of face and finger vein features. Sensors 22, 6039 (2022).
Lee, M. J., Teoh, A. B. J., Uhl, A., Liang, S.-N. & Jin, Z. A tokenless cancellable scheme for multimodal biometric systems. Comput. Secur. 108, 102350 (2021).
Kim, J., Jung, Y. G. & Teoh, A. B. J. Multimodal biometric template protection based on a cancelable softmaxout fusion network. Appl. Sci. 12, 2023 (2022).
Vallabhadas, D. K. & Sandhya, M. Cancelable bimodal shell using fingerprint and iris. J. Electron. Imaging 32, 063027–063027 (2023).
Morampudi, M. K., Sandhya, M. & Dileep, M. Privacy-preserving bimodal authentication system using fan-vercauteren scheme. Optik 274, 170515 (2023).
Jha, K., Jain, A. & Srivastava, S. Multimodal biometric authentication system leveraging optimally trained ensemble classifier using feature-level fusion. Technol. Health Care 09287329251363424 (2025).
El Rahman, A. S. & Alluhaidan, A. S. Enhanced multimodal biometric recognition systems based on deep learning and traditional methods in smart environments. PLoS One 19, e0291084 (2024).
Zhang, K., Zhang, Z., Li, Z. & Qiao, Y. Joint face detection and alignment using multitask cascaded convolutional networks. IEEE Signal Process. Lett. 23, 1499–1503 (2016).
Jalali, S., Boostani, R. & Mohammadi, M. Efficient fingerprint features for gender recognition. Multidimens. Syst. Signal Process. 33, 81–97 (2022).
Ahmadyfard, A. A comprehensive survey of channel attention mechanisms in single image super-resolution. J. Electr. Syst. 20, 9571–9583 (2024).
Guo, M.-H. et al. Attention mechanisms in computer vision: A survey. Comput. Vis. Media 8, 331–368 (2022).
Institute of Automation, Chinese Academy of Sciences (CASIA). Casia-facev5 dataset. http://www.idealtest.org/ (2010).
Institute of Automation, Chinese Academy of Sciences (CASIA). Casia-fingerprintv5 dataset. http://www.idealtest.org/ (2010).
Selvaraj, A., Russel, N. S. & Seenivasan, M. Robust penta-modal biometric identification through deep learning and weighted score fusion. Iran J. Comput. Sci. 8, 553–569 (2025).
Es-Sobbahi, H., Radouane, M. & Nafil, K. Multimodal biometrics: A review of handcrafted and ai-based fusion approaches. IET Biom. 2025, 5055434 (2025).
Funding
The authors received no funding for this work.
Author information
Authors and Affiliations
Contributions
Mrs Pavani Chitrapu conducted the literature survey, developed the conceptualisation, implemented the code, and authored the manuscript. Dr Mahesh Kumar provided technical guidance throughout the project, while Dr Hemantha Kumar served as the primary supervisor for the research.
Corresponding author
Ethics declarations
Competing interests
The authors declare that they have no competing interests.
Ethical approval
This study did not involve direct experiments on human participants; therefore, ethical approval was not required. All datasets used are publicly available.
Consent for publication
All human images shown in this manuscript are taken from publicly available datasets. The respective dataset providers obtained consent for publication.
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Chitrapu, P., Morampudi, M.K. & Kalluri, H.K. A secure and explainable multimodal biometric system using trust adaptive fusion for face and fingerprint. Sci Rep (2026). https://doi.org/10.1038/s41598-026-43252-x
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-026-43252-x
