Research map / Image generation / Image restoration
CHAMMI: A benchmark for channel-adaptive models in microscopy imaging
Zitong Chen, Chau Pham, Siqi Wang, Michael Doron, Nikita Moshkov, Bryan A. Plummer, Juan C Caicedo
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 |
|---|---|---|---|
| fine-tune from ImageNet instead of training from scratch (ablation) | FixedChannels · CHAMMI overall score | F1 performance score: 0.500 to 0.612 | Improved |
| fine-tune from ImageNet instead of training from scratch (ablation) | channel-adaptive models · CHAMMI overall score | F1 performance score: 23% | Improved |
| train channels jointly with a single channel-adaptive model | TargetParam fine-tuned · CHAMMI overall score | F1 performance score: 0.614 -> 0.622 | Improved |
| add TPS augmentation (ablation) | FixedChannels fine-tuned · CHAMMI overall score | F1 performance score: 0.614 -> 0.623 | Improved |
| add TPS augmentation (ablation) | HyperNet fine-tuned · CHAMMI overall score | F1 performance score: 0.616 -> 0.630 | Improved |
| add TPS augmentation (ablation) | Template mixing · CHAMMI overall score | F1 performance score: 0.615 -> 0.616 | No clear effect |
| use SimCLR SSL loss alone instead of supervision (ablation) | FixedChannels · CHAMMI overall score | F1 performance score: 0.623 -> 0.360 | Hurt |
| add 0.2 SSL loss to supervised loss with TPS (ablation) | FixedChannels · CHAMMI overall score | F1 performance score: 0.623 -> 0.633 | Improved |
| add 0.2 SSL loss to supervised loss with TPS (ablation) | Depthwise · CHAMMI overall score | F1 performance score: 0.622 -> 0.608 | Hurt |
| add 0.2 SSL loss to supervised loss with TPS (ablation) | TargetParam · CHAMMI overall score | F1 performance score: 0.628 -> 0.618 | Hurt |
| add SWAD training (ablation) | FixedChannels · CHAMMI OOD generalization overall | F1 performance score: 0.614 -> 0.628 | Improved |
| add MIRO regularization (ablation) | FixedChannels · CHAMMI OOD generalization overall | F1 performance score: 0.614 -> 0.612 | No clear effect |
Closest papers by meaning
- ChAda-ViT : Channel Adaptive Attention for Joint Representation Learning of Heterogeneous Microscopy Images (CVPR 2024 · similarity 0.90)
- Micro-Bench: A Microscopy Benchmark for Vision-Language Understanding (NeurIPS 2024 · similarity 0.87)
- Masked Autoencoders for Microscopy are Scalable Learners of Cellular Biology (CVPR 2024 · similarity 0.85)
- ViTally Consistent: Scaling Biological Representation Learning for Cell Microscopy (ICML 2025 · similarity 0.85)
- MorphoHELM: A Comprehensive Benchmark for Evaluating Representations for Microscopy-Based Morphology Assays (NeurIPS 2026 · similarity 0.85)
- scSplit: Bringing Severity Cognizance to Image Decomposition in Fluorescence Microscopy (NeurIPS 2025 · 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.
