GLFC: Unified Global-Local Feature and Contrast Learning with Mamba-Enhanced UNet for Synthetic CT Generation from CBCT

2025-01-06

AI Summary

To address global structural distortion, local detail blurring, and insufficient multi-tissue contrast modeling in CBCT-to-synthetic CT (sCT) translation, this work proposes a Mamba-enhanced U-Net architecture: Mamba modules are embedded within skip connections to jointly capture long-range dependencies and local textures; additionally, a multi-window contrastive loss (MCL) is introduced to hierarchically enforce intensity consistency across soft-tissue and bone regions. The method synergistically integrates sequence modeling capabilities with encoder-decoder structural advantages for end-to-end supervised learning. Evaluated on the SynthRAD2023 dataset, the generated sCT achieves an SSIM of 91.50%, representing a 13.59-percentage-point improvement over raw CBCT and outperforming state-of-the-art CNN- and Transformer-based baselines. This study constitutes the first application of Mamba within skip connections for medical image synthesis and introduces a clinically motivated, window-specific contrastive loss—establishing a novel paradigm for low-dose CBCT–based precision radiotherapy.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Generating synthetic Computed Tomography (CT) images from Cone Beam Computed Tomography (CBCT) is desirable for improving the image quality of CBCT. Existing synthetic CT (sCT) generation methods using Convolutional Neural Networks (CNN) and Transformers often face difficulties in effectively capturing both global and local features and contrasts for high-quality sCT generation. In this work, we propose a Global-Local Feature and Contrast learning (GLFC) framework for sCT generation. First, a Mamba-Enhanced UNet (MEUNet) is introduced by integrating Mamba blocks into the skip connections of a high-resolution UNet for effective global and local feature learning. Second, we propose a Multiple Contrast Loss (MCL) that calculates synthetic loss at different intensity windows to improve quality for both soft tissues and bone regions. Experiments on the SynthRAD2023 dataset demonstrate that GLFC improved the SSIM of sCT from 77.91% to 91.50% compared with the original CBCT, and significantly outperformed several existing methods for sCT generation. The code is available at https://github.com/intelland/GLFC

Problem

• CBCT to CT-like image generation
• feature preservation
• contrast enhancement

Innovation

• MEUNet
• MCL loss function
• CBCT to CT image enhancement