Research map / Label-efficient and robust learning / Noisy labels
A Supervised Multi-task Framework for Joint cryo-ET Restoration Enabled by Generative Physical Simulation
Xinsheng Wang, Zhidong Yang, Xiaohua Wan, Renmin Han, Shuai Tang, Hao Dong, Fa Zhang, Bin Hu
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 missing-wedge reconstruction loss L_recon to fidelity loss (ablation) | cryoDeRec · EMPIAR-10643 | CNR: 2.174 -> 2.434 | Improved |
| add missing-wedge reconstruction loss L_recon to fidelity loss (ablation) | cryoDeRec · EMPIAR-10499 | CNR: 1.318 -> 1.637 | Improved |
| add missing-wedge reconstruction loss L_recon to fidelity loss (ablation) | cryoDeRec · EMPIAR-10643 | ENL: 171.225 -> 193.694 | Improved |
| add missing-wedge reconstruction loss L_recon to fidelity loss (ablation) | cryoDeRec · EMPIAR-10499 | ENL: 145.883 -> 162.764 | Improved |
| train supervised on simulated data jointly for denoising and missing wedge instead of IsoNet | cryoDeRec · EMPIAR-10643 | CNR: 1.874 -> 2.434 | Improved |
| cryoDeRec versus IsoNet on EMPIAR-10499 low-contrast data | cryoDeRec · EMPIAR-10499 | CNR: 0.336 -> 1.637 | Improved |
| cryoDeRec versus IsoNet on simulated Tomo-5lzf | cryoDeRec · Simulated Tomo-5lzf (SNR=0.5, phi=60) | PSNR: 7.15 -> 10.84 | Improved |
| lower SNR from 0.5 to 0.1 in simulated data | cryoDeRec · Simulated Tomo-5lzf (phi=60) | PSNR: 10.84 -> 10.82 | No clear effect |
| reduce tilt range from phi=60 to phi=50 | cryoDeRec · Simulated Tomo-5lzf | PSNR: 10.84 -> 10.37 | Hurt |
| reduce tilt range from phi=60 to phi=50 | cryoDeRec · Simulated Tomo-1qvr | PSNR: 10.80 -> 10.81 | No clear effect |
| increase tilt increment from 3 to 5 degrees (fewer tilts) (ablation) | cryoDeRec · EMPIAR-10499 | qualitative restoration | No clear effect |
Closest papers by meaning
- Vox-UDA: Voxel-wise Unsupervised Domain Adaptation for Cryo-Electron Subtomogram Segmentation with Denoised Pseudo-Labeling (AAAI 2025 · similarity 0.89)
- DRACO: A Denoising-Reconstruction Autoencoder for Cryo-EM (NeurIPS 2024 · similarity 0.89)
- CryoSPIN: Improving Ab-Initio Cryo-EM Reconstruction with Semi-Amortized Pose Inference (NeurIPS 2024 · similarity 0.89)
- cryoSPHERE: Single-Particle HEterogeneous REconstruction from cryo EM (ICLR 2025 · similarity 0.87)
- CryoBench: Diverse and challenging datasets for the heterogeneity problem in cryo-EM (NeurIPS 2024 · similarity 0.87)
- CryoGEM: Physics-Informed Generative Cryo-Electron Microscopy (NeurIPS 2024 · similarity 0.87)
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.
