Research map / Image generation / Autoregressive and latent generation
Unsupervised Multi-Scale Segmentation of 3D Subcellular World with Stable Diffusion Foundation Model
Mostofa Rafid Uddin, HM Shadman Tabib, Thanh-Huy Nguyen, Kashish Gandhi, Min Xu
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 unsupervised Stable Diffusion pipeline instead of manual-label supervised UNet for membranes | UNet membrane segmentation · S. Pombe VPP (overall) | Dice: 0.324 -> 0.309 | No clear effect |
| use unsupervised pseudo labels instead of expert labels, evaluated on the training tomogram | UNet membrane segmentation · S. Pombe TS_0008 | Dice: 0.665 -> 0.508 | Hurt |
| use the tailored Stable Diffusion pipeline instead of SAM, FreeSOLO or CutLer | membrane segmentation · S. Pombe VPP (overall) | Dice: 0.048 -> 0.309 | Improved |
| train DeepETPicker on 88 unsupervised pseudo coordinates instead of 100 ground truth coordinates | DeepETPicker · S. Pombe VPP (overall) | F1: 0.35 -> 0.43 | Improved |
| use unsupervised pseudo coordinates instead of 500 ground truth coordinates | DeepETPicker · S. Pombe VPP (overall) | F1: 0.57 -> 0.43 | Hurt |
| use unsupervised pseudo coordinates instead of CrYOLO with 500 coordinates | particle picking · S. Pombe VPP (overall) | F1: 0.31 -> 0.43 | Improved |
| use Stable Diffusion features instead of other vision foundation models (ablation) | unsupervised feature extraction · cryo-ET slab (qualitative) | feature quality | Improved |
Closest papers by meaning
- Vox-UDA: Voxel-wise Unsupervised Domain Adaptation for Cryo-Electron Subtomogram Segmentation with Denoised Pseudo-Labeling (AAAI 2025 · similarity 0.91)
- cryoSPHERE: Single-Particle HEterogeneous REconstruction from cryo EM (ICLR 2025 · similarity 0.88)
- CryoFM: A Flow-based Foundation Model for Cryo-EM Densities (ICLR 2025 · similarity 0.88)
- Reconstructing Heterogeneous Biomolecules via Hierarchical Gaussian Mixtures and Part Discovery (NeurIPS 2025 · similarity 0.87)
- CryoBench: Diverse and challenging datasets for the heterogeneity problem in cryo-EM (NeurIPS 2024 · similarity 0.86)
- BOE-ViT: Boosting Orientation Estimation with Equivariance in Self-Supervised 3D Subtomogram Alignment (CVPR 2025 · similarity 0.86)
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.
