Research map / Image generation / Autoregressive and latent generation
SKM-TEA: A Dataset for Accelerated MRI Reconstruction with Dense Image Labels for Quantitative Clinical Evaluation
Arjun D Desai, Andrew M Schmidt, Elka B Rubin, Christopher Michael Sandino, Marianne Susan Black, Valentina Mazzoli, Kathryn J Stevens, Robert Boutin, Christopher Re, Garry E Gold, Brian Hargreaves, Akshay Chaudhari
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 |
|---|---|---|---|
| Use an unrolled reconstruction network instead of U-Net | MRI reconstruction on SKM-TEA · 6x and 8x acceleration, both qDESS echoes | pSNR and SSIM | Improved |
| Change how the two echoes are supplied to models of the same architecture (ablation) | MRI reconstruction on SKM-TEA · 6x and 8x acceleration | pSNR and SSIM | No clear effect |
| Use 8x rather than 6x acceleration | MRI reconstruction on SKM-TEA · qDESS echoes | pSNR and SSIM | Hurt |
| Use the E2 echo alone rather than E1, E1⊕E2, or RSS input (ablation) | DICOM segmentation models · DICOM Track, four tissues | segmentation performance | Hurt |
| Use a V-Net rather than a U-Net segmentation architecture | DICOM segmentation models · DICOM Track | DSC and ASSD | Mixed |
| Evaluate reconstruction using standard pSNR and SSIM instead of regional T2 error | MRI reconstruction models · SKM-TEA Raw Data Track | correlation with absolute T2 estimation error: |ρ| ≤ 0.52 | Hurt |
| Evaluate segmentation using DSC and ASSD instead of regional T2 error | DICOM segmentation models · SKM-TEA DICOM Track | correlation with absolute T2 error: |ρ| ≤ 0.4 | Hurt |
| Use regional qMRI T2 biomarkers as direct evaluation endpoints | MRI reconstruction and segmentation evaluation · SKM-TEA | clinical relevance of evaluation | Improved |
Closest papers by meaning
- MosaicMRI: A Diverse Dataset and Benchmark for Raw Musculoskeletal MRI (NeurIPS 2026 · similarity 0.90)
- AeSPa : Attention-guided Self-supervised Parallel Imaging for MRI Reconstruction (CVPR 2025 · similarity 0.85)
- Adaptive Sampling of k-Space in Magnetic Resonance for Rapid Pathology Prediction (ICML 2024 · similarity 0.84)
- Efficient Noise Calculation in Deep Learning-based MRI Reconstructions (ICML 2025 · similarity 0.84)
- Fast MRI for All: Bridging Access Gaps by Training without Raw Data (NeurIPS 2025 · similarity 0.83)
- A Unified Model for Compressed Sensing MRI Across Undersampling Patterns (CVPR 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.
