LightFusionNet: Lightweight Dual-Stream Network with Predictive Context Attention for Efficient Medical Image Fusion
Sagar, A.
Show abstract
Multimodal image fusion aims to integrate complementary information from multiple imaging modalities into a single, informative representation, which is crucial for applications in medical imaging and microscopy. Existing methods often face trade-offs between structural fidelity, edge preservation, and computational efficiency. In this work, we propose Light-FusionNet, a lightweight dual-stream network designed to efficiently fuse multimodal images while retaining key structural, textural, and intensity features. The network leverages depthwise separable convolutions to reduce model complexity and incorporates a Predictive Context Attention (PCA) mechanism to selectively emphasize informative regions in the feature maps. Extensive experiments on benchmark medical imaging datasets, including PET-MRI, SPECT-MRI, and CT-MRI, demonstrate that our approach achieves comparable qualitative and quantitative performance compared to state-of-the-art fusion methods, while maintaining low computational cost. The proposed method provides an effective and efficient solution for multimodal image fusion, suitable for both clinical and research applications.
Matching journals
The top 8 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- STAMP: Simultaneous Training and Model Pruning for Low Data Regimes in Medical Image Segmentation 95%
- Comparison of domain adaptation techniques for white matter hyperintensity segmentation in brain MR images 94%
- MISPEL: A supervised deep learning harmonizationmethod for multi-scanner neuroimaging data 94%
Similar papers in this journal
- Probabilistic Brain MR Image Transformation Using Generative Models 96%
- Dual Adversarial Deconfounding Autoencoder for joint batch-effects removal from multi-center and multi-scanner radiomics data 95%
- Effective Deep Learning Approaches for Predicting COVID-19 Outcomes from Chest Computed Tomography Volumes 93%
Similar papers in this journal
Similar papers in this journal
- GLAPAL-H: Global, Local, And Parts Aware Learner for Hydrocephalus Infection Diagnosis in Low-Field MRI 96%
- Maximum Classifier Discrepancy Generative Adversarial Network for Jointly Harmonizing Scanner Effects and Improving Reproducibility of Downstream Tasks 95%
- Saak Transform-Based Machine Learning for Light-Sheet Imaging of Cardiac Trabeculation 92%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.