Remote sensing semantic segmentation requires models that can jointly capture fine spatial details and high-level semantic context across complex scenes. While classical encoder-decoder architectures such as U-Net remain strong baselines, they often struggle to fully exploit global semantics and structured feature interactions. In this work, we propose HQF-Net, a hybrid quantum-classical multi-scale fusion network for remote sensing image segmentation. HQF-Net integrates multi-scale semantic guidance from a frozen DINOv3 ViT-L/16 backbone with a customized U-Net architecture through a Deformable Multiscale Cross-Attention Fusion (DMCAF) module. To enhance feature refinement, the framework further introduces quantum-enhanced skip connections (QSkip) and a Quantum bottleneck with Mixture-of-Experts (QMoE), which combines complementary local, global, and directional quantum circuits within an adaptive routing mechanism. Experiments on three remote sensing benchmarks show consistent improvements with the proposed design. HQF-Net achieves 0.8568 mIoU and 96.87% overall accuracy on LandCover.ai, 71.82% mIoU on OpenEarthMap, and 55.28% mIoU with 99.37% overall accuracy on SeasoNet. An architectural ablation study further confirms the contribution of each major component. These results show that structured hybrid quantum-classical feature processing is a promising direction for improving remote sensing semantic segmentation under near-term quantum constraints.
Figures & tables
Figure 1: Overview of the proposed HQF-Net architecture. The model integrates a frozen DINOv3 ViT-L/16 encoder with Deformable Multi-Scale Cross-Attention Fusion (DMCAF) module. The main hybrid components are Quantum Skip (QSkip) refinement blocks and a bottleneck with Quantum Mixture-of-Experts (QMoE) block that adaptively combines local, global, and diagonal quantum experts. CT2D denotes transposed convolution.
Figure 2: Overview of Quantum Circuits used in HQF-Net - a) Enrichment Multiscale Circuit, b) Local Circuit, c) Global Circuit, and d) Diagonal Circuit
Source
Model
mIoU
OA (%)
Fan et al. [ 11 ]
FQCNN
0.2000
53.26
Fan et al. [ 11 ]
MQCNN
0.1520
39.03
Kumar et al. [ 25 ]
CNN
0.1500
45.87
Kumar et al. [ 25 ]
COQCNN
0.1280
36.65
Ronneberger et al. [ 39 ]
UNet
0.6451
82.43
Zhou et al. [ 55 ]
UNet++
0.6553
70.89
Table 1: Performance comparison on the LandCover.ai test set.
Model
Backbone
mIoU
UNet [ 39 ]
—
60.4
DANet [ 12 ]
ResNet101
60.1
ClipSeg [ 30 ]
CLIP
58.6
BoTNet [ 44 ]
ResNet50
61.5
MANet [ 27 ]
ResNet101
64.0
SegFormer [ 53 ]
MiT
66.0
Table 2: Comparison on OpenEarthMap with State-of-the-Art Methods (mIoU %)
Model
Backbone
Acc.
mIoU
DeepLabv3 [ 24 ]
DenseNet121
–
47.53
DeepLabv3 PT [ 24 ]
DenseNet121
–
48.69
DeepLabv3 [ 5 ]
ResNet-50
83.52
50.79
DeepLabv3 [ 5 ]
ConvNeXt-Small
81.36
46.39
DeepLabv3 [ 5 ]
Swin-Tiny
82.75
50.81
UperNet [ 52 ]
ResNet-50
83.20
49.59
Table 3: Performance Comparison on SeasoNet (Accuracy and mIoU %)
Figure 3: Qualitative segmentation on the LandCover.ai dataset, showing input images, ground-truth masks, and predictions by HQF-Net.
Model Variant
DINOv3
DMCAF
Q-Skip
Q-MoE
mIoU
OA (%)
Baseline Fusion Methods
U-Net + DINOv3 (Multiplication)
✓
0.7335
90.24
U-Net + DINOv3 (Addition)
✓
0.7520
91.80
Proposed Hybrid Architectures
+ DMCAF (Deformable Fusion)
✓
✓
0.7815
92.10
+ DMCAF + Q-Skip
✓
✓
✓
0.8387
94.22
Table 4: Ablation study showing the impact of different fusion strategies and quantum components on the LandCover.ai dataset.
Appendix figures & tables4 assets
Supplementary material from the paper’s appendix.
Appendix
Figure 4: Two-qubit parameterized quantum filter used as a local interaction block in HQF-Net. The unit models pairwise feature interactions and serves as a basic building block within the larger quantum circuit designs.
Figure 5: Qualitative segmentation results on the SeasoNet dataset showing original images, ground-truth masks, and comparisons with other models.
Figure 6: Qualitative segmentation results on the OpenEarthMap dataset showing original images, ground-truth masks, and comparisons with other models.
Figure 7: Qualitative segmentation results on the OpenEarthMap and SeasoNet datasets showing original images, ground-truth masks, and HQF-Net predictions.
Semantic segmentation in remote sensing is commonly addressed using classical deep learning architectures such as U-Net, which require a large number of parameters to model complex spatial relationships. Quantum machine learning (QML) provides an alternative representation paradigm by mapping classical features into quantum states, but its direct application to high-dimensional images remains challenging under near-term quantum hardware constraints. In this work, we propose HQ-UNet, a hybrid quantum-classical U-Net architecture that integrates a compact parameterized quantum circuit at the bottleneck of a classical U-Net. The proposed design uses a non-pooling quantum convolutional module to enrich highly compressed encoder features before decoding, while keeping the quantum component shallow and parameter-efficient. Experiments on the LandCover.ai dataset show that HQ-UNet achieves a mean IoU of 0.8050 and an overall accuracy of 94.76%, outperforming the classical U-Net baseline. These results suggest that compact quantum bottlenecks can enhance feature representation for remote sensing image segmentation under near-term quantum constraints. This highlights the potential of hybrid quantum-classical designs as a promising direction for parameter-efficient dense prediction in Earth observation.
Md Aminur Hossain, Ayush V. Patel, Ikshwaku Vanani +1
Space Applications Centre, Indian Space Research Organisation (ISRO), India · New-Brunswick Honors College, Rutgers University, New Jersey · Centre of Studies in Resources Engineering, Indian Institute of Technology Bombay
Reliable flood detection is critical for disaster management, yet classical deep learning models often struggle with the high-dimensional, nonlinear complexities inherent in remote sensing data. To mitigate these limitations, we introduced a novel Quantum-Enhanced Vision Transformer (ViT) that synergizes the global context-awareness of transformers with the expressive feature extraction capabilities of quantum computing. Using remote sensing imagery, we developed a hybrid architecture that processes inputs through parallel pathways, a ViT backbone and a quantum branch utilizing a 4-qubit parameterized quantum circuit for localized feature mapping. These distinct representations were fused to optimize binary classification. Results showed that the proposed hybrid model significantly outperformed a classical ViT baseline, increased overall accuracy from 84.48% to 94.47% and the F1-score from 0.841 to 0.944. Notably, the quantum integration substantially improved discriminative power in complex terrains for both class. These findings validate the potential of quantum-classical hybrid models to enhance precision in hydrological monitoring and earth observation applications.
Soumyajit Maity, Behzad Ghanbarian
iResearchE3 Lab, Department of Earth and Environmental Sciences, University of Texas at Arlington, Arlington, TX, USA
We introduce HQ-JEPA, a hybrid quantum-classical joint-embedding predictive architecture for cross-modal remote sensing representation learning. The proposed framework extends JEPA-style masked latent prediction to paired Sentinel-1 and Sentinel-2 imagery by predicting masked target representations from visible context regions while aligning heterogeneous modality features in a shared embedding space. To improve representation quality, HQ-JEPA combines four complementary objectives: latent token prediction, cross-modal token alignment, SIGReg-based Gaussian regularization in the fused latent space, and a differentiable SWAP-test-based Fidelity Quantum Similarity (FQS) loss. Unlike pixel reconstruction methods, HQ-JEPA learns semantic representations directly in latent space and uses quantum state-overlap-based similarity as an additional regularization signal. We evaluate the pretrained encoder on GeoBench classification and segmentation tasks under linear probing and fine-tuning settings. Results show that HQ-JEPA achieves competitive and often superior performance over strong self-supervised and remote sensing foundation-model baselines, demonstrating the benefit of integrating predictive self-supervision, cross-modal geometric regularization, and quantum fidelity-based representation learning for remote sensing applications.
Md Aminur Hossain, Ayush V. Patel, Sanjay K. Singh +1
Space Applications Centre, Indian Space Research Organisation, Ahmedabad, India · Centre of Studies in Resources Engineering, Indian Institute of Technology Bombay, Mumbai, India