Parallel Multi-Circuit Quantum Feature Fusion in Hybrid Quantum-Classical Convolutional Neural Networks for Breast Tumor Classification

Parallel Multi-Circuit Quantum Feature Fusion in Hybrid Quantum-Classical Convolutional Neural Networks for Breast Tumor Classification Ece Yurtseven Robert College of Istanbul Istanbul, Turkey yurece.27@robcol.k12.tr Abstract—Quantum machine learning has emerged as a promising approach to improve feature extraction and classi- fication tasks in high-dimensional data domains such as medical imaging. In this work, we present a hybrid Quantum-Classical Convolutional Neural Network (QCNN) architecture designed for the binary classification of the BreastMNIST dataset, a standardized benchmark for distinguishing between benign and malignant breast tumors. Our architecture integrates classical convolutional feature extraction with two distinct quantum cir- cuits: an amplitude-encoding variational quantum circuit (VQC) and an angle-encoding VQC circuit with circular entanglement, both implemented on four qubits. These circuits generate quan- tum feature embeddings that are fused with classical features to form a joint feature space, which is subsequently processed by a fully connected classifier. To ensure fairness, the hybrid QCNN is parameter-matched against a baseline classical CNN, allowing us to isolate the contribution of quantum layers. Both models are trained under identical conditions using the Adam optimizer and binary cross-entropy loss. Experimental evaluation in five independent runs demonstrates that the hybrid QCNN achieves statistically significant improvements in classification accuracy compared to the classical CNN, as validated by a one- sided Wilcoxon signed rank test (p = 0.03125) and supported by large effect size of Cohen’s d = 2.14. Our results indicate that hybrid QCNN architectures can leverage entanglement and quantum feature fusion to enhance medical image classification tasks. This work establishes a statistical validation framework for assessing hybrid quantum models in biomedical applications and highlights pathways for scaling to larger datasets and deployment on near-term quantum hardware. Index Terms—Quantum Computing, QCNN, Medical Image Classification, Hybrid Quantum-Classical Models, BreastMNIST Dataset I. INTRODUCTION Breast cancer has emerged as a global health concern, affecting millions of women worldwide. In 2020, more than 2.3 million women were diagnosed with breast cancer, making it the most commonly diagnosed cancer globally [1]. Breast cancer remains a leading cause of cancer-related mortality, with an estimated 670,000 deaths recorded in 2022 [2]. This burden is projected to continue rising in the following years. Early mammographic screening has been determined to play a critical role in lowering breast cancer mortality. Furthermore, AI-driven diagnostic systems have the potential to significantly improve the precision and efficiency of the screening process [3]. However, the classical computational models that power these AI systems are beginning to encounter fundamental architectural and practical limitations, particularly in the demanding domain of medical imaging. Classical Convolutional Neural Networks (CNNs) remain pivotal in modern computer vision, especially for biomedical image analysis, due to their efficacy in automatically extracting hierarchical features for diagnostic tasks like tumor detection [4]. However, the ever-increasing complexity and volume of medical data often strain computational resources, challeng- ing the scalability of classical deep learning algorithms. To address these limitations, Quantum Machine Learning (QML) offers a promising solution, leveraging quantum mechanics principles like superposition and entanglement for potential computational acceleration and robust pattern recognition [5]. Quantum computing represents a transformative approach with the potential to address the computational bottlenecks faced by classical machine learning [6]. Using quantum mechanical properties such as superposition and entangle- ment, quantum computers offer the potential for exponential speedups in computation [7]. This makes them suitable for complex computational tasks. Quantum algorithms can lead to several benefits over classical counterparts, offering expo- nential computational boosts and the ability to simultaneously process more inputs with fewer operations. In machine learn- ing, quantum algorithms provide significant speedups in data handling by efficiently managing high-dimensional vectors [8]. Building on these concepts, variational quantum cir- cuits (VQCs) have become a widely used framework for near-term quantum machine learning. VQCs employ pa- rameterized quantum circuits optimized through a hybrid quantum–classical loop, allowing models to access high- dimensional quantum representations while relying on clas- sical algorithms for training [9]. This approach has supported applications such as quantum feature mapping, classification, and generative modeling, where VQCs can express complex transformations with relatively minimal cir

This content is AI-processed based on open access ArXiv data.