Research map / Image generation / Image restoration
uSplit: Image Decomposition for Fluorescence Microscopy
Ashesh Ashesh, Alexander Krull, Moises Di Sante, Francesco Pasqualini, Florian Jug
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 Regular-LC to vanilla HVAE at patch size 64 (ablation) | HVAE · PaviaATN Act vs Tub, 4 BU blocks | PSNR: 25.4 -> 27.8 | Improved |
| add Deep-LC (more hierarchy levels) to HVAE (ablation) | HVAE with Regular-LC · PaviaATN Act vs Tub | PSNR: 27.8 -> 28.6 | Improved |
| increase training patch size from 64 to 128 when LC is used (ablation) | HVAE + Regular-LC · PaviaATN Act vs Tub | PSNR: 27.8 -> 27.9 | No clear effect |
| add more hierarchy levels to vanilla HVAE at patch size 64 (ablation) | vanilla HVAE · PaviaATN Act vs Tub | PSNR | No clear effect |
| use Inner Padding instead of Outer Padding for tiled prediction (ablation) | vanilla HAE and HVAE · PaviaATN and Hagen et al. | PSNR | Improved |
| use LC with HVAE on 256x256 SinosoidalCritters | vanilla HVAE · SinosoidalCritters 256x256, Njoin=0 | PSNR: 19.4 -> 41.5 | Improved |
| use Deep-LC instead of Lean-LC (ablation) | HVAE · four microscopy tasks (average) | PSNR: +0.8 PSNR | Improved |
| use HVAE instead of HAE with LC (ablation) | LC variants · four microscopy tasks (average) | PSNR: +0.5 PSNR | Improved |
| Deep-LC HVAE (patch 64) versus Vanilla-XL HVAE (patch 512) | HVAE · PaviaATN Tub vs Nuc | PSNR: 32.5 vs 32.9 | Hurt |
| increase training patch size for vanilla U-Net (ablation) | U-Net · PaviaATN Act vs Tub | PSNR: 28.1 -> 27.9 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Double-DIP — "Double-DIP": Unsupervised Image Decomposition via Coupled Deep-Image-Priors (CVPR 2019) · better on 1 of 1 shared result
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
- scSplit: Bringing Severity Cognizance to Image Decomposition in Fluorescence Microscopy (NeurIPS 2025 · similarity 0.90)
- ViTally Consistent: Scaling Biological Representation Learning for Cell Microscopy (ICML 2025 · similarity 0.85)
- ZoomLDM: Latent Diffusion Model for Multi-scale Image Generation (CVPR 2025 · similarity 0.85)
- FluoCLIP: Stain-Aware Focus Quality Assessment in Fluorescence Microscopy (CVPR 2026 · similarity 0.84)
- Micro-Bench: A Microscopy Benchmark for Vision-Language Understanding (NeurIPS 2024 · similarity 0.84)
- Masked Autoencoders for Microscopy are Scalable Learners of Cellular Biology (CVPR 2024 · similarity 0.84)
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.
