Research map / Label-efficient and robust learning / OOD and anomaly detection
Building Robust Vision Encoders for Cross-Dataset Evaluation in Immunofluorescent Microscopy
Umar Marikkar, Syed Sameed Husain, Muhammad Awais, Sara Atito
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 per-group convolutional stems (ablation) | ViT-S baseline · HPA-Loc | mAP-31: 0.380 -> 0.523 | Improved |
| add branched per-group encoders (fc) (ablation) | C3R ViT-S · JUMP-Ret | kNN: 0.510 -> 0.529 | Improved |
| add Masked Context Distillation (MCD) (ablation) | C3R ViT-S · HPA-Loc | mAP-31: 0.520 -> 0.535 | Improved |
| add MCD (ablation) | C3R ViT-S · JUMP-Ret | kNN: 0.529 -> 0.518 | No clear effect |
| add MCD (ablation) | C3R ViT-B · CHAMMI-FE | CPS: 0.338 -> 0.543 | Improved |
| add MCD (ablation) | C3R ViT-B · JUMP-TRex | compound/MoA: 0.338 -> 0.345 (U2OS compound) | Improved |
| flip context/concept group assignment at evaluation (ablation) | C3R · JUMP-Ret | mAP / kNN | Hurt |
| flip context/concept assignment for WTC-11 (ablation) | C3R · CHAMMI-FE WTC-11 | average F1: 0.551 -> 0.536 | Hurt |
| drop context channels on student only in MCD (ablation) | C3R ViT-S · HPA-Loc | mAP-31: 0.519 -> 0.536 | Improved |
| also mask teacher context channels (S + T) (ablation) | C3R ViT-S · JUMP-Ret | kNN: 0.518 -> 0.504 | Hurt |
| random context sampling c in {1,2,3} instead of fixed c (ablation) | C3R ViT-S · HPA-Loc | mAP-31: 0.531 -> 0.536 | Improved |
| keep only 1 context channel in MCD (ablation) | C3R ViT-S · JUMP-Ret | kNN: 0.518 -> 0.499 | Hurt |
Closest papers by meaning
- ChAda-ViT : Channel Adaptive Attention for Joint Representation Learning of Heterogeneous Microscopy Images (CVPR 2024 · similarity 0.88)
- Masked Autoencoders for Microscopy are Scalable Learners of Cellular Biology (CVPR 2024 · similarity 0.88)
- ViTally Consistent: Scaling Biological Representation Learning for Cell Microscopy (ICML 2025 · similarity 0.87)
- CellCLIP - Learning Perturbation Effects in Cell Painting via Text-Guided Contrastive Learning (NeurIPS 2025 · similarity 0.86)
- Micro-Bench: A Microscopy Benchmark for Vision-Language Understanding (NeurIPS 2024 · similarity 0.86)
- A Cross Modal Knowledge Distillation & Data Augmentation Recipe for Improving Transcriptomics Representations through Morphological Features (ICML 2025 · similarity 0.85)
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.
