Quantum Enhanced Hybrid Face Recognition Using MultiGate-Quantum Convolutional Neural Network (MG-QCNN)

Abstract

Face recognition has become an essential biometric technology for applications such as security monitoring, user authentication, and intelligent surveillance systems. Conventional approaches based on deep neural networks, particularly convolution-based architectures, provide high recognition performance but often demand considerable computational resources and extended training durations as network complexity increases. To overcome these challenges, Quantum Machine Learning (QML) utilizes quantum computing concepts, including quantum superposition and entangled states, to achieve efficient feature representation and processing.This paper presents a Hybrid Quantum-Classical Face Recognition System built on a Multi-Gate Quantum Convolutional Neural Network (MG-QCNN). The proposed framework employs a 4-qubit variational quantum circuit that incorporates RY state encoding together with RX, RZ, and CNOT quantum operations to extract discriminative facial representations. In addition, a structural ablation analysis is conducted to investigate the influence of individual quantum components on the overall recognition capability of the model. To enhance spatial feature learning and classification performance, several classical neural network layers, namely convolution, batch normalization, max-pooling, dropout, and fully connected layers, are integrated after the quantum feature extraction stage. Within the hybrid framework, the quantum circuit is responsible for generating meaningful feature representations, whereas the classical neural network performs feature refinement and final identity classification.The framework is further combined with the DeepFace verification approach to perform facial similarity analysis and identity verification. Moreover, an Open-Set Face Recognition mechanism is incorporated to determine if an input facial image corresponds to a registered identity or represents an unseen individual. This capability increases system reliability and supports accurate unknown-person identification. The effectiveness of the proposed framework is validated using the ORL and Yale face datasets. The obtained experimental findings indicate that the complete MG-QCNN architecture achieves superior performance compared with reduced quantum configurations, emphasizing the significance of multigate quantum operations and entanglement strategies. The developed hybrid model attains an accuracy of 99.17% on the ORL dataset and 97.80% on the Yale dataset.Furthermore, the application of data augmentation techniques enhances the model\'s capability to capture diverse facial characteristics and minimizes overfitting, particularly when training data are limited. Overall, the developed hybrid quantum-classical framework offers accurate face recognition, efficient feature extraction, and dependable unknown-person identification, making it a promising solution for next-generation secure biometric systems.

Introduction

This paper proposes a Hybrid Quantum-Classical Face Recognition System that combines a Multi-Gate Quantum Convolutional Neural Network (MG-QCNN) with classical deep learning, DeepFace verification, and Open-Set Face Recognition to improve facial recognition accuracy, verification reliability, and unknown-person detection.

Background

Face recognition is widely used in surveillance, banking, attendance systems, and biometric authentication. Traditional methods such as PCA, LDA, Eigenfaces, and Fisherfaces rely on handcrafted features and perform poorly under variations in lighting, facial expression, pose, and occlusion. Deep learning approaches, particularly CNNs, have significantly improved recognition performance through automatic feature extraction. Advanced models such as FaceNet, VGGFace, ArcFace, and DeepFace achieve high accuracy but require large datasets, high computational power, and long training times.

Quantum Machine Learning

Quantum computing introduces qubits, which exploit superposition and entanglement to process information more efficiently than classical binary systems. Quantum Machine Learning (QML) combines quantum computing with machine learning to improve computational efficiency and feature representation. A key development is the Quantum Convolutional Neural Network (QCNN), which performs feature extraction using parameterized quantum circuits.

The proposed MG-QCNN extends conventional QCNNs by using:

• Multiple quantum gate operations (RY, RX, and RZ rotations)
• CNOT-based entanglement
• Richer quantum feature extraction capable of capturing complex facial characteristics

Hybrid Quantum-Classical Framework

To overcome the limitations of both quantum and classical methods, the proposed system integrates:

• MG-QCNN for quantum feature extraction
• Classical CNN for feature refinement and classification
• DeepFace for facial verification using similarity scores
• Open-Set Recognition for detecting previously unseen individuals

This hybrid architecture combines the computational advantages of quantum computing with the robustness and scalability of classical deep learning.

Related Work

Previous research has demonstrated:

• MG-QCNNs outperform simplified quantum models by using multiple quantum gates and entanglement.
• CNN-based architectures such as FaceNet, ArcFace, ResNet, and DenseNet achieve excellent face recognition performance.
• Hybrid quantum-classical models improve feature representation while reducing computational complexity.
• Existing face recognition systems mainly address closed-set recognition, with limited research integrating quantum learning, DeepFace verification, and open-set recognition into a unified framework.

Proposed Methodology

The proposed system consists of eight stages:

1. Dataset Loading
   • Uses the ORL and Yale face datasets.
2. Image Preprocessing
   • Resize images to 48×48 pixels.
   • Normalize pixel values.
   • Apply data augmentation through rotation and horizontal shifting.
3. Dataset Splitting
   • 70% training and 30% testing.
4. MG-QCNN Feature Extraction
   • Uses a four-qubit variational circuit.
   • RY state encoding.
   • RX and RZ rotation gates.
   • CNOT entanglement for learning facial relationships.
5. Hybrid CNN Classification
   • Classical CNN with Conv2D, Batch Normalization, Max Pooling, Dropout, and Dense layers refines quantum features and performs classification.
6. Model Training and Evaluation
   • Evaluates recognition accuracy and loss.
   • Compares full MG-QCNN with reduced quantum architectures.
7. DeepFace Verification
   • Confirms identity through facial similarity matching.
8. Open-Set Recognition
   • Uses similarity thresholds to classify faces as:
     • Known Person
     • Unknown Person

The final decision combines outputs from quantum feature extraction, CNN classification, DeepFace verification, and open-set recognition.

Experimental Results

The framework was implemented in Python and evaluated using the ORL and Yale face datasets.

Yale Dataset

• Successfully handled variations in illumination and facial expressions.
• Quantum feature extraction and CNN refinement produced accurate and consistent recognition.
• Demonstrated robustness under challenging lighting conditions.

ORL Dataset

• Accurately recognized faces despite variations in expression and pose.
• Produced stable and reliable classification performance.
• Showed strong generalization capability for practical biometric applications.

Conclusion

The present study developed a Hybrid Quantum-Classical Face Recognition framework based on a Multi-Gate Quantum Convolutional Neural Network (MG-QCNN) to perform facial identification and verification tasks. The proposed approach integrates quantum-driven feature extraction with classical convolutional neural network learning to generate effective facial representations and accurate classification results. The effectiveness of the framework was examined using the ORL and Yale benchmark datasets containing variations in lighting conditions, facial expressions, and appearance changes. The experimental findings revealed that the proposed system could successfully learn discriminative facial characteristics and maintain reliable recognition performance under diverse image conditions. The hybrid architecture accurately identified enrolled individuals and remained robust even when moderate modifications were introduced into facial images. Furthermore, integrating the DeepFace verification module improved similarity evaluation and increased the reliability of the recognition process. An important aspect of this work is the implementation of an Open-Set Recognition mechanism that enables the system to distinguish between known identities and previously unseen individuals. The experimental outcomes showed that the framework accurately recognized registered users while effectively detecting faces that were not included in the training data. This capability increases the practical usefulness of the proposed approach in biometric authentication, intelligent surveillance, and security-oriented applications. In conclusion, the developed MG-QCNN-based system provides an efficient and dependable solution for advanced face recognition applications. The integration of quantum computing concepts with deep learning methodologies demonstrates the promise of hybrid quantum-classical frameworks for solving complex biometric recognition problems and offers a strong foundation for future research in quantum machine learning and intelligent recognition systems.

References

[1] Y. Zhu, A. Bouridane, M. E. Celebi, D. Konar, P. Angelov, Q. Ni, and R. Jiang, “Quantum Face Recognition Using Multi-Gate Quantum Convolutional Neural Networks,” IEEE Access, vol. 10, pp. 1–12, 2022. [2] I. Cong, S. Choi, and M. D. Lukin, “Quantum Convolutional Neural Networks,” Nature Physics, vol. 15, no. 12, pp. 1273–1278, 2019. [3] M. Schuld and F. Petruccione, Quantum Machine Learning, Springer International Publishing, Switzerland, 2018. [4] E. Henderson, S. Shakya, S. Pradhan, and T. Cook, “Quanvolutional Neural Networks: Powering Image Recognition with Quantum Circuits,” arXiv preprint arXiv:1904.04767, 2019. [5] Y. LeCun, Y. Bengio, and G. Hinton, “Deep Learning,” Nature, vol. 521, pp. 436–444, 2015. [6] F. Schroff, D. Kalenichenko, and J. Philbin, “FaceNet: A Unified Embedding for Face Recognition and Clustering,” in Proc. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Boston, MA, USA, 2015, pp. 815–823. [7] O. M. Parkhi, A. Vedaldi, and A. Zisserman, “Deep Face Recognition,” in Proc. British Machine Vision Conference (BMVC), Swansea, UK, 2015, pp. 41.1–41.12. [8] J. Deng, J. Guo, N. Xue, and S. Zafeiriou, “ArcFace: Additive Angular Margin Loss for Deep Face Recognition,” in Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2019, pp. 4690–4699. [9] A. Krizhevsky, I. Sutskever, and G. E. Hinton, “ImageNet Classification with Deep Convolutional Neural Networks,” in Proc. Advances in Neural Information Processing Systems (NeurIPS), 2012, pp. 1097–1105. [10] K. Simonyan and A. Zisserman, “Very Deep Convolutional Networks for Large-Scale Image Recognition,” arXiv preprint arXiv:1409.1556, 2014. [11] C. Szegedy, W. Liu, Y. Jia, et al., “Going Deeper with Convolutions,” in Proc. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015, pp. 1–9. [12] K. He, X. Zhang, S. Ren, and J. Sun, “Deep Residual Learning for Image Recognition,” in Proc. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 770–778. [13] G. Huang, Z. Liu, L. Van Der Maaten, and K. Q. Weinberger, “Densely Connected Convolutional Networks,” in Proc. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017, pp. 4700–4708. [14] DeepFace Contributors, “DeepFace: Lightweight Face Recognition and Facial Attribute Analysis Framework,” GitHub Repository, 2024. [15] J. Preskill, “Quantum Computing in the NISQ Era and Beyond,” Quantum, vol. 2, Art. no. 79, pp. 1–16, 2018.

Copyright

Copyright © 2026 Chamarti Venkata Sri Vyshnavi, Dr. O. Srinivas Rao . This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.