Image classification is a fundamental task in computer vision. Although classical convolutional neural networks are particularly effective at extracting local spatial features, they are inherently limited in their ability to capture global frequency-domain information. To address this limitation, this paper introduces a hybrid model that couples a classical convolutional feature extraction branch with a quantum feature extraction branch built upon the quantum Fourier transform. The quantum branch maps the input into a low-dimensional representation and processes it through a sequence of quantum encoding, the quantum Fourier transform, a variational quantum circuit, and the inverse quantum Fourier transform; the resulting quantum features are concatenated with the classical features and jointly used for the final classification. Ablation studies quantify the individual contribution of each quantum component, and systematic experiments on the MNIST, Fashion-MNIST, and CIFAR-10 benchmark datasets, conducted over multiple independent random seeds on a strictly held-out test set, show that the complete architecture consistently outperforms its purely classical counterparts. All results and comparative analyses are presented in the experimental section.
Image classification is a fundamental task in computer vision. Although classical convolutional neural networks are particularly effective at extracting local spatial features, they are inherently limited in their ability to capture global frequency-domain information. To address this limitation, this paper introduces a hybrid model that couples a classical convolutional feature extraction branch with a quantum feature extraction branch built upon the quantum Fourier transform. The quantum branch maps the input into a low-dimensional representation and processes it through a sequence of quantum encoding, the quantum Fourier transform, a variational quantum circuit, and the inverse quantum Fourier transform; the resulting quantum features are concatenated with the classical features and jointly used for the final classification. Ablation studies quantify the individual contribution of each quantum component, and systematic experiments on the MNIST, Fashion-MNIST, and CIFAR-10 benchmark datasets, conducted over multiple independent random seeds on a strictly held-out test set, show that the complete architecture consistently outperforms its purely classical counterparts. All results and comparative analyses are presented in the experimental section.
Abstract With the rapid advancement of quantum machine learning, numerous hybrid quantum–classical convolutional neural network (HQCNN) architectures have been proposed. However, most existing HQCNNs adopt serial topologies-either classical convolution followed by quantum processing or vice versa-which suffer from low feature propagation efficiency, limited representational capacity, and poor robustness due to inherent information bottlenecks. To address these limitations, this paper proposes a novel parallel dual-branch HQCNN ( PDB-HQCNN ). The proposed architecture comprises two mutually independent branches: a classical feature branch for extracting local structural features and a quantum feature branch for capturing global high-dimensional correlations. Features from both branches are concatenated and fed into a fully connected layer to generate classification predictions based on quantum measurement outcomes. Extensive experiments on MNIST, Fashion-MNIST, and Medical MNIST datasets demonstrate that the proposed PDB-HQCNN achieves classification accuracy exceeding 96% across all benchmarks. Compared with existing serial hybrid architectures, pure classical convolutional neural networks, and pure quantum neural networks, the proposed parallel dual-branch design attains state-of-the-art classification accuracy on all test datasets, validating its effectiveness in leveraging complementary classical and quantum representational strengths.
Abstract Quantum Machine Learning (QML) has emerged as a promising paradigm that leverages the principles of quantum mechanics—such as superposition and entanglement—to overcome the limitations of classical machine learning in handling complex, high-dimensional data. However, practical challenges such as limited coherence times of qubits and the integration of quantum circuits with classical architectures have hindered the full realization of QML’s potential. To address these challenges, this paper proposes a novel hybrid architecture, termed the Classical-Quantum Hybrid Convolutional Neural Network (CQHCNN), which integrates quantum-inspired layers into conventional deep learning frameworks. These quantum layers utilize quantum information processing to extract complex features, while the classical components perform feature extraction and classification. This paper evaluates the performance of our model on several benchmark datasets, including MNIST, Fashion-MNIST, CIFAR-10 and COVID-19 Radiography, demonstrating its ability to outperform some other Quantum Convolutional Neural Networks (QCNNs) in terms of many metrics. Experimental results indicate that the hybrid model achieves significant improvements in classification accuracy and computational efficiency, highlighting the potential of quantum-inspired techniques to enhance classical machine learning models.
Abstract This paper presents a parallel quantum–classical hybrid architecture for image classification that integrates classical feature extraction with parallel quantum processing while maintaining competitive performance and improving parameter efficiency. Low-level and intermediate image features are first extracted by a shared convolutional neural network and then processed by multiple parallel quantum branches. Each branch employs a multilayer variational quantum circuit with a rotation–entanglement–rotation structure. To enhance representation diversity, two encoding-specific strategies are introduced. For amplitude encoding, deterministic spatial permutations generate distinct structural views of the same feature map across branches. For angle encoding, a layer-wise data re-uploading mechanism injects complementary feature subsets into different branches and circuit layers. Heterogeneous entanglement topologies with dynamic routing further diversify qubit interactions without increasing the number of trainable parameters. The measured expectation values from all branches are concatenated and fed into a compact linear classifier. Experiments on MNIST, CIFAR-10, and a thermal image dataset show that the proposed architecture achieves 98.35% accuracy on MNIST and competitive results on the other two datasets, while using substantially fewer parameters in the classification head than a fully classical baseline with the same feature extractor. These results demonstrate that parallel quantum processing combined with structured data embedding can serve as an effective and parameter-efficient decision stage for image classification.
Medical image classification is a critical component of modern healthcare; however, accurate diagnosis remains challenging due to limited annotated datasets, class imbalance, and the high dimensionality of medical imaging data. To address these challenges, a hybrid quantum-classical neural network (HQCNN) is proposed, integrating classical deep learning with variational quantum learning for medical image classification. The proposed architecture combines a five-layer convolutional neural network (CNN) for hierarchical feature extraction with a lightweight 4-qubit variational quantum circuit (VQC) incorporating quantum state encoding, superposition and entanglement mechanisms, and a quantum attention-Fourier (QAF) module. This hybrid design aims to improve nonlinear feature representation and quantum parameter efficiency while maintaining a shallow quantum circuit suitable for noisy intermediate-scale quantum (NISQ)-era constraints. Experimental evaluation on six MedMNIST benchmark datasets demonstrated competitive performance across both binary and multi-class classification tasks. HQCNN achieved 98.88% accuracy on the binary subset of PathMNIST (classes 0 vs. 1), 97.61% accuracy on the multi-class OrganAMNIST dataset, and 86.29% accuracy on BreastMNIST. Comparative experiments and statistical analyses demonstrated consistent improvements over the controlled BHQNN baseline, while component-wise ablation studies showed that the QAF module, superposition and entanglement mechanisms, and expressive parameterized rotations contributed to classification performance in a complementary and dataset-dependent manner. Moreover, HQCNN reduced the number of trainable quantum parameters by approximately 55.6% compared with the baseline hybrid quantum neural network (BHQNN). Noise-aware simulations further showed that the model retained relatively stable predictive performance under moderate depolarizing noise, supporting further evaluation under near-term quantum computing conditions. Overall, the results demonstrate that HQCNN provides a parameter-efficient hybrid quantum-classical framework for medical image classification and offers a promising foundation for further investigation of quantum-enhanced medical image analysis.
Deep learning has achieved significant success in pattern recognition, with convolutional neural networks (CNNs) serving as a foundational architecture for extracting spatial features from images. Quantum computing provides an alternative computational framework, a hybrid quantum-classical convolutional neural networks (QCCNNs) leverage high-dimensional Hilbert spaces and entanglement to surpass classical CNNs in image classification accuracy under comparable architectures. Despite performance improvements, QCCNNs typically use fixed quantum layers without incorporating trainable quantum parameters. This limits their ability to capture non-linear quantum representations and separates the model from the potential advantages of expressive quantum learning. In this work, we present a hybrid quantum-classical-quantum convolutional neural network (QCQ-CNN) that incorporates a quantum convolutional filter, a shallow classical CNN, and a trainable variational quantum classifier. This architecture aims to enhance the expressivity of decision boundaries in image classification tasks by introducing tunable quantum parameters into the end-to-end learning process. Through a series of small-sample experiments on MNIST, F-MNIST, and MRI tumor datasets, QCQ-CNN demonstrates competitive accuracy and convergence behavior compared to classical and hybrid baselines. We further analyze the effect of ansatz depth and find that moderate-depth quantum circuits can improve learning stability without introducing excessive complexity. Additionally, simulations incorporating depolarizing noise and finite sampling shots suggest that QCQ-CNN maintains a certain degree of robustness under realistic quantum noise conditions. While our results are currently limited to simulations with small-scale quantum circuits, the proposed approach offers a potentially promising direction for hybrid quantum learning in near-term applications.
Image classification is a fundamental task in deep learning, and recent advances in quantum computing have generated significant interest in quantum neural networks. Traditionally, Convolutional Neural Networks (CNNs) are employed to extract image features, while Multilayer Perceptrons (MLPs) handle decision making. However, parameterized quantum circuits offer the potential to capture complex image features and define sophisticated decision boundaries. In this paper, we present a novel Hybrid Quantum–Classical Neural Network (H-QNN) for image classification, and demonstrate its effectiveness using the MNIST dataset. Our model combines quantum computing with classical supervised learning to enhance classification accuracy and computational efficiency. In this study, we detail the architecture of the H-QNN, emphasizing its capability in feature learning and image classification. Experimental results demonstrate that the proposed H-QNN model outperforms conventional deep learning methods in various training scenarios, showcasing its effectiveness in high-dimensional image classification tasks. Additionally, we explore the broader applicability of hybrid quantum–classical approaches in other domains. Our findings contribute to the growing body of work in quantum machine learning, and underscore the potential of quantum-enhanced models for image recognition and classification.