Student Publications [Scholarly]

Facial Recognition by Using Ensemble and Network of Quantum Convolutional Neural Network

Document Type

Article

Abstract

The use of Quantum Convolutional Neural Networks (QCNNs) is a promising technique for image classification, leveraging the principles of quantum parallelism and entanglement. Network of Quantum Neural Networks (N-QCNN) is introduced as a quantum AI technique that incorporates both quantum and traditional AI techniques. In this work, a new ensemble N-QCNN model is introduced that incorporates QCNN and SVM to enhance face recognition and classification performance, particularly on small datasets. As an application of this model, 48x48 pixel grayscale facial images are processed by using the well-known Yale Face dataset and a patch-based amplitude encoding scheme. In this approach, each module utilizes a unique quantum circuit design to extract distinct image features. In this two-layer model, the first layer utilized a set of three parallel QCNNs while the second layer utilized three parallel SVMs. Hence, the quantum derived features in the first layer are fed into a set of classical SVMs for classification. The final prediction of the model relied on the majority voting mechanism among the three modules, which includes a tie breaking rule for a conclusive outcome. We tested the proposed model on a partial and full Yale Face dataset; the test accuracy of the model resulted in 95.27% accuracy on a subset of 7 subjects to test the effectiveness of the model on a partial dataset. The performance remained robust when the testing was extended onto the entire dataset, with an accuracy of 80% attainment on all 15 subjects. These results highlight the effectiveness of this newly introduced ensemble-based N-QCNN models during multiclass facial image classification even when the dataset has a limited size. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2027.

Publication Title

Communications in Computer and Information Science

Publication Date

2027

First Page

144

Last Page

154

ISSN

1865-0929

ISBN

9783032284716

DOI

10.1007/978-3-032-28472-3_11

Keywords

ensemble learning, face recognition, quantum convolutional neural networks, quantum machine learning, support vector machine

Link to Full Text

Share

COinS