Research map / Image generation / Editing and style transfer
High Tissue Contrast MRI Synthesis Using Multi-Stage Attention-GAN for Segmentation
Mohammad Hamghalam, Baiying Lei, Tianfu Wang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add attention-guided HTC synthesis | CycleGAN + segmentation · BraTS 2018, WT | Dice: 0.9231 (0.24) -> 0.9508 (0.58) | Improved |
| Add attention-guided HTC synthesis | CycleGAN + segmentation · BraTS 2018, TC | Dice: 0.9016 (0.11) -> 0.9304 (0.51) | Improved |
| Add attention-guided HTC synthesis | CycleGAN + segmentation · BraTS 2018, ET | Dice: 0.8699 (0.14) -> 0.8891 (0.13) | Improved |
| Replace T1 with synthetic FLAIR* | U-Net, Res-U-Net, and FC-DenseNet · BraTS'18 validation | Dice: U-Net EN: 0.7874 -> 0.7921; WT: 0.8913 -> 0.8998; TC: 0.8402 -> 0.8462; Res-U-Net EN: 0.7891 -> 0.7944; WT: 0.8951 -> 0. | Improved |
| Add source attention to CycleGAN (ablation) | MRI-to-HTC synthesis · ET synthesis ablation | SSIM: 0.6072 -> 0.6384 | Improved |
| Add source attention to CycleGAN (ablation) | MRI-to-HTC synthesis · ET synthesis ablation | PSNR: 17.64 -> 18.21 | Improved |
| Add target attention to CycleGAN (ablation) | MRI-to-HTC synthesis · ET synthesis ablation | SSIM: 0.6072 -> 0.6387 | Improved |
| Add target attention to CycleGAN (ablation) | MRI-to-HTC synthesis · ET synthesis ablation | PSNR: 17.64 -> 18.29 | Improved |
| Add both source and target attention blocks (ablation) | CycleGAN MRI-to-HTC synthesis · ET synthesis ablation | SSIM: 0.6384 -> 0.6647 | Improved |
| Add both source and target attention blocks (ablation) | CycleGAN MRI-to-HTC synthesis · ET synthesis ablation | PSNR: 18.21 -> 18.86 | Improved |
| Add multi-stage structure to the model with both attention blocks (ablation) | CycleGAN MRI-to-HTC synthesis · ET synthesis ablation | SSIM: 0.6647 -> 0.6849 | Improved |
| Add multi-stage structure to the model with both attention blocks (ablation) | CycleGAN MRI-to-HTC synthesis · ET synthesis ablation | PSNR: 18.86 -> 19.87 | Improved |
Closest papers by meaning
- Learn2Synth: Learning Optimal Data Synthesis Using Hypergradients for Brain Image Segmentation (ICCV 2025 · similarity 0.84)
- Translating and Segmenting Multimodal Medical Volumes With Cycle- and Shape-Consistency Generative Adversarial Network (CVPR 2018 · similarity 0.84)
- MAISI-v2: Accelerated 3D High-Resolution Medical Image Synthesis with Rectified Flow and Region-specific Contrastive Loss (AAAI 2026 · similarity 0.84)
- Noise-Consistent Siamese-Diffusion for Medical Image Synthesis and Segmentation (CVPR 2025 · similarity 0.83)
- Towards Generalizable Tumor Synthesis (CVPR 2024 · similarity 0.83)
- Boosting ViT-based MRI Reconstruction from the Perspectives of Frequency Modulation, Spatial Purification, and Scale Diversification (AAAI 2025 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
