Research map / 3D vision / Depth and stereo
Amortized Inference for Heterogeneous Reconstruction in Cryo-EM
Axel Levy, Gordon Wetzstein, Julien N. P. Martel, FREDERIC P POITEVIN, Ellen D Zhong
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 amortized encoder pose prediction instead of exhaustive pose search | cryoFIRE vs cryoDRGN2 · Synthetic ribosome Large (5M) | training time: 21:37h -> 1:55h | Improved |
| use amortized encoder pose prediction instead of exhaustive pose search | cryoFIRE vs cryoDRGN2 · Synthetic ribosome Medium (500k) | training time: 5:10h -> 1:28h | Improved |
| use amortized encoder pose prediction instead of exhaustive pose search | cryoFIRE vs cryoDRGN2 · Synthetic ribosome Small (50k) | training time: 1:21h -> 1:33h | Hurt |
| use amortized encoder instead of pose search for latent variable estimation on test images | cryoFIRE vs cryoDRGN2 · Synthetic ribosome test set (10k) | test runtime: 3 min. -> 11 sec. | Improved |
| use amortized encoder instead of exhaustive pose search | cryoFIRE vs cryoDRGN2 · Synthetic ribosome Medium | median rotation error (degrees): 0.8 -> 2.7 | Hurt |
| use amortized encoder instead of exhaustive pose search | cryoFIRE vs cryoDRGN2 · Synthetic ribosome Large | confusion error: 0.002 -> 0.0002 | Improved |
| use amortized encoder instead of exhaustive pose search | cryoFIRE vs cryoDRGN2 · Synthetic ribosome Medium | resolution of major volume (pixels): 2.5 -> 2.7 | No clear effect |
| use variational encoder instead of point-estimate encoder (ablation) | cryoFIRE · synthetic datasets | performance | No clear effect |
| increase training set size from 50k to 5M images | cryoFIRE · Synthetic ribosome test set | test median rotation error (degrees): 2.6 -> 1.2 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- cryoDRGN2 — CryoDRGN2: Ab Initio Neural Reconstruction of 3D Protein Structures From Real Cryo-EM Images (ICCV 2021) · better on 5 of 11 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- CryoSPIN: Improving Ab-Initio Cryo-EM Reconstruction with Semi-Amortized Pose Inference (NeurIPS 2024 · similarity 0.94)
- CryoBench: Diverse and challenging datasets for the heterogeneity problem in cryo-EM (NeurIPS 2024 · similarity 0.92)
- Mixture of neural fields for heterogeneous reconstruction in cryo-EM (NeurIPS 2024 · similarity 0.92)
- cryoSPHERE: Single-Particle HEterogeneous REconstruction from cryo EM (ICLR 2025 · similarity 0.91)
- Reconstructing Heterogeneous Biomolecules via Hierarchical Gaussian Mixtures and Part Discovery (NeurIPS 2025 · similarity 0.91)
- Multiscale guidance of protein structure prediction with heterogeneous cryo-EM data (NeurIPS 2025 · similarity 0.91)
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.
