Research map / Label-efficient and robust learning / Recognition and long-tail
OpenSRH: optimizing brain tumor surgery using intraoperative stimulated Raman histology
Cheng Jiang, Asadur Zaman Chowdury, Xinhai Hou, Akhil Kondepudi, Christian Freudiger, Kyle Stephen Conway, Sandra Camelo-Piragua, Daniel A Orringer, Honglak Lee, Todd Hollon
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 ImageNet pretraining instead of random init | ResNet50 · OpenSRH patch-level val | accuracy: 84.4 -> 86.5 | Improved |
| use ImageNet pretraining instead of random init | ViT-S · OpenSRH patch-level val | accuracy: 77.2 -> 83.7 | Improved |
| use ImageNet pretraining instead of random init | ResNet50 · OpenSRH patient-level val | accuracy: 90.0 -> 88.9 | Hurt |
| use ImageNet pretraining instead of random init | ViT-S · OpenSRH patient-level val | accuracy: 85.0 -> 88.9 | Improved |
| use ResNet50 instead of ViT-S | random init · OpenSRH patch-level val | accuracy: 77.2 -> 84.4 | Improved |
| SimCLR pretraining on OpenSRH instead of ImageNet transfer | ResNet50 linear eval · OpenSRH patch-level val | accuracy: 68.3 -> 79.1 | Improved |
| SimCLR pretraining on OpenSRH instead of ImageNet transfer | ViT-S linear eval · OpenSRH patch-level val | accuracy: 71.8 -> 76.8 | Improved |
| SupCon instead of SimCLR | ResNet50 linear eval · OpenSRH patient-level val | accuracy: 83.9 -> 90.0 | Improved |
| SupCon instead of SimCLR | ViT-S linear eval · OpenSRH patient-level val | accuracy: 80.0 -> 87.8 | Improved |
| SupCon instead of SimCLR | ViT-S fine-tuning · OpenSRH patch-level val | accuracy: 81.4 -> 81.2 | No clear effect |
| SupCon instead of ImageNet pretraining | ViT-S linear eval · OpenSRH patient-level val | accuracy: 88.3 -> 87.8 | No clear effect |
| fine-tune backbone instead of linear eval (SimCLR) | ResNet50 · OpenSRH patch-level val | accuracy: 79.1 -> 86.3 | Improved |
Closest papers by meaning
- Step-Calibrated Diffusion for Biomedical Optical Image Restoration (AAAI 2025 · similarity 0.83)
- FOCUS: Knowledge-enhanced Adaptive Visual Compression for Few-shot Whole Slide Image Classification (CVPR 2025 · similarity 0.81)
- Physics-Regularized Multi-Modal Image Assimilation for Brain Tumor Localization (NeurIPS 2024 · similarity 0.81)
- PathVQ: Reforming Computational Pathology Foundation Model for Whole Slide Image Analysis via Vector Quantization (NeurIPS 2025 · similarity 0.81)
- Are nuclear masks all you need for improved out-of-domain generalisation? A closer look at cancer classification in histopathology (NeurIPS 2024 · similarity 0.81)
- Sequential Attention-based Sampling for Histopathological Analysis (NeurIPS 2025 · similarity 0.81)
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.
