Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

IEEE International Geoscience and Remote Sensing Symposium (IGARSS)

Quantum Mamba for Superpixel Segmentation

Jhao-Ting Lin, Chia-Hsiang Lin, Jian-Kai Huang, Xiang-Yu You

Published Aug 1, 2026• 0 citations
Publisher

Abstract

For the first time, quantum deep network (QUEEN) is introduced for remote sensing image superpixel segmentation, where we employ a dual quantum encoding technique to mitigate the issue of imperfect gate fidelity. To facilitate quantum global understanding, which is critical for superpixel computing, we further employ the Mamba network to encode the spatial information retained in the remote sensing image. Specifically, we design a Mamba-CNN feature extractor that incorporates an interleaving strategy to efficiently capture long-range dependencies for enhancing local features. Building upon the enriched features, a quantum classification module is employed for the final superpixel assignment. Distinctively, our QUEEN adopts a spatially-aware measurement strategy. By explicitly mapping the spatial geometry of adjacent superpixels to distinct quantum observables, this strategy maintains geometric concepts consistent with the loss function. Experimental results demonstrate that the proposed method improves boundary adherence and achieves precise segmentation.

DOI
10.1109/igarss60577.2026.11731321

Related papers

Development of quantum superpixel entanglement model for remote sensing image classification

Guoqing Zhou, Mengxuan Zhang, Jianrong Xiao +3 more

The existing models based on quantum pixels can mine spectral information and entanglement features, but are time-consuming in practice due to its characteristics on pixel-level entanglement. For this reason, an innovative image classification model, named ‘Quantum Superpixel Entanglement Classification (QSEC) model’, is proposed in this study. By treating superpixels in remote sensing images (RSIs) as quantum particles, we extend the paradigm of image classification from the pixel level to the superpixel level. This design fully leverages the spectral information, textural features, and inter-superpixel quantum entanglement characteristics inherent in RSIs, enabling the effective capture and processing of complex image information via quantum entanglement-driven computational frameworks. Furthermore, the framework quantizes RSI superpixels to map their greyscale and texture features onto corresponding quantum states, constructs a quantum superpixel entanglement pattern, extracts the entanglement relationships between superpixels using entanglement coefficients, and ultimately develops a classification method applicable to multispectral images. The results from the four groups of the experiments demonstrated that: (1) The QSEC model achieves an average classification accuracy of and Kappa coefficient (KC) of 92.02% and 0.86, respectively, with an average running time of 29.25 seconds; (2) the running time of the QSEC model was decreased by up to 400 seconds, the classification speed was improved by 75.29% when compared with the neural network classification (NNC), support vector machine (SVM) and deep learning (DL) model; and (3) the QSEC model improves classification accuracy by 8.76% and increases KC by 0.12 when compared with unsupervised classification models such as ISODATA and K-means.

International Journal of Remote Sensing•Mar 3, 2026•0 citations

Research on quantum many-body entanglement image classification

Mengxuan Zhang, Guoqing Zhou, Jianrong Xiao +1 more

Remote sensing image (RSI) classification is key for extracting value from remote sensing data, aiding decision-making and resource management. Recent quantum pixel-based models excel at handling complex spectral and entanglement features but face high computational costs and lengthy processing due to quantum pixel entanglement, limiting large-scale and real-time applications. To address this, this study proposes a novel Quantum Many-Body Entanglement Classification model. This model first performs superpixel segmentation on RSIs, and then converts the segmented images into quantum states with quantum characteristics. Each superpixel corresponds to one quantum state, thereby enhancing the classification granularity from the pixel level to the superpixel level. Subsequently, the model conducts cluster analysis based on the entanglement coefficient. This paper takes Dongying, Shandong Province as the research area and uses Landsat 8 image data to conduct a comparative experiment between the proposed classification method and the traditional remote sensing classification method. The results show that the classification accuracy of this method meets the requirements of practical applications.

International Conference on Remote Sensing and Global Positioning Algorithm (RSGPA)•Sep 29, 2025•0 citations

QueenNet: Quantum-Enhanced Neural Network for Hyperspectral Image Classification

Qingwang Wang, Shunyuan Wang, Yinxing Zhang +3 more

Hyperspectral image (HSI) classification is a critical task in the field of remote sensing. Existing deep learning models have made significant progress in HSI classification, but often suffer from parameter redundancy and insufficient noise robustness when handling the high-dimensional nonlinear spatial-spectral information. Recently, quantum computing have garnered significant attention in modeling higher-order correlations with remarkable parameter efficiency. Inspired by this, we propose a QUantum-Enhanced nEural Network (QueenNet) for HSI classification. QueenNet comprises three primary modules, including Quantum Encoding (QE), Quantum Convolution (QC), and Quantum Measurement (QM). Specifically, QE maps HSI into an hyper high-dimensional Hilbert space, encoding it as a quantum states. QC performs HSI feature extraction and transformation using quantum circuits composed of parameterized quantum gate sequences, which effectively enhances the model's ability to capture joint spatial-spectral features while suppressing spectral noise. QM introduces nonlinear feature extraction operations and decodes the HSI features of quantum states back into classic interpretable probability distributions for classification. QueenNet effectively alleviate the limitations of existing HSI classification deep learning models by leveraging the high-dimensional feature representation advantages of quantum computing, introducing a novel paradigm to HSI classification. Extensive experiments on three benchmark datasets demonstrate that QueenNet achieves an order-of-magnitude reduction in model parameters while surpassing the classification performance of state-of-the-art methods.

IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing•Jan 1, 2026•3 citations

Hybrid Quantum-Classical Network for Hyperspectral Change Detection

Chia-Hsiang Lin, Tzu-Hsuan Lin, Ching-Yun Liang +1 more

Hyperspectral change detection (HCD) is a critical technique for monitoring changes on the Earth's surface over time. Empowered by the spatial relation (captured by the graph learning) and the quantum deep network (QUEEN), precise detection of environmental changes is achieved in this work. With the growing attention on quantum computing technology, highly innovative hybrid quantum-classical QUEENs have been developed in recent years for various hyperspectral remote sensing applications. During the feature extraction stage, quantum modules and classical modules are employed to extract features at the pixel level and superpixel level, respectively. By fusing these two fundamentally different types of feature information, the dual-branch network structure achieves accurate HCD results. Moreover, the proposed highly entangled QUEEN module has mathematically provable quantum expressibility, which ensures a lightweight quantum architecture while preserving exceptional classification performance. After quantum/classical feature extraction, the features are fed into another quantum classifier to decide whether the hyperspectral pixels are changed or not. Experiments conducted on real-world hyperspectral datasets demonstrate the superiority of the proposed quantum-graph network (QGNet), suggesting future investigators explore hybrid quantum-classical learning methods.

IEEE International Geoscience and Remote Sensing Symposium (IGARSS)•Aug 3, 2025•2 citations

Hybrid Quantum Deep Learning With Superpixel Encoding for Earth Observation Data Classification

Fan Fan, Yilei Shi, Tobias Guggemos +1 more

Earth observation (EO) has inevitably entered the Big Data era. The computational challenge associated with analyzing large EO data using sophisticated deep learning models has become a significant bottleneck. To address this challenge, there has been a growing interest in exploring quantum computing as a potential solution. However, the process of encoding EO data into quantum states for analysis potentially undermines the efficiency advantages gained from quantum computing. This article introduces a hybrid quantum deep learning model that effectively encodes and analyzes EO data for classification tasks. The proposed model uses an efficient encoding approach called superpixel encoding, which reduces the quantum resources required for large image representation by incorporating the concept of superpixels. To validate the effectiveness of our model, we conducted evaluations on multiple EO benchmarks, including Overhead-MNIST, So2Sat LCZ42, and SAT-6 datasets. In addition, we studied the impacts of different interaction gates and measurements on classification performance to guide model optimization. The experimental results suggest the validity of our model for accurate classification of EO data. Our models and code are available on https://github.com/zhu-xlab/SEQNN.

IEEE Transactions on Neural Networks and Learning Systems•Jan 1, 2025•9 citations

A multi-topology quantum convolutional neural network with qubit-measurement attention for image classification

Qingshan Wu, Wenjie Liu, Xinran Li +2 more

With the increasing scale of data and complexity of problems, some researchers have explored the integration of parameterized quantum circuits (PQCs) within convolutional neural networks (CNNs) as a means to enhance algorithmic performance. However, in most current quantum convolutional neural networks (QCNN) models, a single topological structure for quantum kernels (qkernels) is adopted and only one qubit of qkernels is measured, both of which may limit the model’s performance. To solve these problems, a novel multi-topology quantum convolutional neural networks with qubit-measurement attention for image classification is proposed. In order to enhance the capability of feature extraction, a multi-topology PQCs strategy is proposed, i.e., we adopt the different topology PQCs to construct quantum convolutional layers. In addition, a qubit-measurement attention mechanism is designed to mitigate the significant loss of entanglement information during the measurement phase. Specifically, each qubit in the qkernel is measured to generate a local feature map, and the weight of each local feature map is then calculated, resulting in the final feature map. Nine image classification experiments conducted on CIFAR-10 demonstrate that our model outperforms the state-of-the-art QCNN model, achieving an improvement of 11.7% on ten categories classification. Our model not only introduces a new approach for constructing QCNNs but also provides valuable reference for designing attention mechanisms tailored to quantum computing.

Engineering Applications of Artificial Intelligence•Aug 5, 2025•11 citations
For the first time, quantum deep network (QUEEN) is introduced for remote sensing image superpixel segmentation, where we employ a dual quantum encoding technique to mitigate the issue of imperfect gate fidelity. To facilitate quantum global understanding, which is critical for superpixel computing, we further employ the Mamba network to encode the spatial information retained in the remote sensing image. Specifically, we design a Mamba-CNN feature extractor that incorporates an interleaving strategy to efficiently capture long-range dependencies for enhancing local features. Building upon the enriched features, a quantum classification module is employed for the final superpixel assignment. Distinctively, our QUEEN adopts a spatially-aware measurement strategy. By explicitly mapping the spatial geometry of adjacent superpixels to distinct quantum observables, this strategy maintains geometric concepts consistent with the loss function. Experimental results demonstrate that the proposed method improves boundary adherence and achieves precise segmentation.
The existing models based on quantum pixels can mine spectral information and entanglement features, but are time-consuming in practice due to its characteristics on pixel-level entanglement. For this reason, an innovative image classification model, named ‘Quantum Superpixel Entanglement Classification (QSEC) model’, is proposed in this study. By treating superpixels in remote sensing images (RSIs) as quantum particles, we extend the paradigm of image classification from the pixel level to the superpixel level. This design fully leverages the spectral information, textural features, and inter-superpixel quantum entanglement characteristics inherent in RSIs, enabling the effective capture and processing of complex image information via quantum entanglement-driven computational frameworks. Furthermore, the framework quantizes RSI superpixels to map their greyscale and texture features onto corresponding quantum states, constructs a quantum superpixel entanglement pattern, extracts the entanglement relationships between superpixels using entanglement coefficients, and ultimately develops a classification method applicable to multispectral images. The results from the four groups of the experiments demonstrated that: (1) The QSEC model achieves an average classification accuracy of and Kappa coefficient (KC) of 92.02% and 0.86, respectively, with an average running time of 29.25 seconds; (2) the running time of the QSEC model was decreased by up to 400 seconds, the classification speed was improved by 75.29% when compared with the neural network classification (NNC), support vector machine (SVM) and deep learning (DL) model; and (3) the QSEC model improves classification accuracy by 8.76% and increases KC by 0.12 when compared with unsupervised classification models such as ISODATA and K-means.
Remote sensing image (RSI) classification is key for extracting value from remote sensing data, aiding decision-making and resource management. Recent quantum pixel-based models excel at handling complex spectral and entanglement features but face high computational costs and lengthy processing due to quantum pixel entanglement, limiting large-scale and real-time applications. To address this, this study proposes a novel Quantum Many-Body Entanglement Classification model. This model first performs superpixel segmentation on RSIs, and then converts the segmented images into quantum states with quantum characteristics. Each superpixel corresponds to one quantum state, thereby enhancing the classification granularity from the pixel level to the superpixel level. Subsequently, the model conducts cluster analysis based on the entanglement coefficient. This paper takes Dongying, Shandong Province as the research area and uses Landsat 8 image data to conduct a comparative experiment between the proposed classification method and the traditional remote sensing classification method. The results show that the classification accuracy of this method meets the requirements of practical applications.
Hyperspectral image (HSI) classification is a critical task in the field of remote sensing. Existing deep learning models have made significant progress in HSI classification, but often suffer from parameter redundancy and insufficient noise robustness when handling the high-dimensional nonlinear spatial-spectral information. Recently, quantum computing have garnered significant attention in modeling higher-order correlations with remarkable parameter efficiency. Inspired by this, we propose a QUantum-Enhanced nEural Network (QueenNet) for HSI classification. QueenNet comprises three primary modules, including Quantum Encoding (QE), Quantum Convolution (QC), and Quantum Measurement (QM). Specifically, QE maps HSI into an hyper high-dimensional Hilbert space, encoding it as a quantum states. QC performs HSI feature extraction and transformation using quantum circuits composed of parameterized quantum gate sequences, which effectively enhances the model's ability to capture joint spatial-spectral features while suppressing spectral noise. QM introduces nonlinear feature extraction operations and decodes the HSI features of quantum states back into classic interpretable probability distributions for classification. QueenNet effectively alleviate the limitations of existing HSI classification deep learning models by leveraging the high-dimensional feature representation advantages of quantum computing, introducing a novel paradigm to HSI classification. Extensive experiments on three benchmark datasets demonstrate that QueenNet achieves an order-of-magnitude reduction in model parameters while surpassing the classification performance of state-of-the-art methods.
Hyperspectral change detection (HCD) is a critical technique for monitoring changes on the Earth's surface over time. Empowered by the spatial relation (captured by the graph learning) and the quantum deep network (QUEEN), precise detection of environmental changes is achieved in this work. With the growing attention on quantum computing technology, highly innovative hybrid quantum-classical QUEENs have been developed in recent years for various hyperspectral remote sensing applications. During the feature extraction stage, quantum modules and classical modules are employed to extract features at the pixel level and superpixel level, respectively. By fusing these two fundamentally different types of feature information, the dual-branch network structure achieves accurate HCD results. Moreover, the proposed highly entangled QUEEN module has mathematically provable quantum expressibility, which ensures a lightweight quantum architecture while preserving exceptional classification performance. After quantum/classical feature extraction, the features are fed into another quantum classifier to decide whether the hyperspectral pixels are changed or not. Experiments conducted on real-world hyperspectral datasets demonstrate the superiority of the proposed quantum-graph network (QGNet), suggesting future investigators explore hybrid quantum-classical learning methods.
Earth observation (EO) has inevitably entered the Big Data era. The computational challenge associated with analyzing large EO data using sophisticated deep learning models has become a significant bottleneck. To address this challenge, there has been a growing interest in exploring quantum computing as a potential solution. However, the process of encoding EO data into quantum states for analysis potentially undermines the efficiency advantages gained from quantum computing. This article introduces a hybrid quantum deep learning model that effectively encodes and analyzes EO data for classification tasks. The proposed model uses an efficient encoding approach called superpixel encoding, which reduces the quantum resources required for large image representation by incorporating the concept of superpixels. To validate the effectiveness of our model, we conducted evaluations on multiple EO benchmarks, including Overhead-MNIST, So2Sat LCZ42, and SAT-6 datasets. In addition, we studied the impacts of different interaction gates and measurements on classification performance to guide model optimization. The experimental results suggest the validity of our model for accurate classification of EO data. Our models and code are available on https://github.com/zhu-xlab/SEQNN.
With the increasing scale of data and complexity of problems, some researchers have explored the integration of parameterized quantum circuits (PQCs) within convolutional neural networks (CNNs) as a means to enhance algorithmic performance. However, in most current quantum convolutional neural networks (QCNN) models, a single topological structure for quantum kernels (qkernels) is adopted and only one qubit of qkernels is measured, both of which may limit the model’s performance. To solve these problems, a novel multi-topology quantum convolutional neural networks with qubit-measurement attention for image classification is proposed. In order to enhance the capability of feature extraction, a multi-topology PQCs strategy is proposed, i.e., we adopt the different topology PQCs to construct quantum convolutional layers. In addition, a qubit-measurement attention mechanism is designed to mitigate the significant loss of entanglement information during the measurement phase. Specifically, each qubit in the qkernel is measured to generate a local feature map, and the weight of each local feature map is then calculated, resulting in the final feature map. Nine image classification experiments conducted on CIFAR-10 demonstrate that our model outperforms the state-of-the-art QCNN model, achieving an improvement of 11.7% on ten categories classification. Our model not only introduces a new approach for constructing QCNNs but also provides valuable reference for designing attention mechanisms tailored to quantum computing.