Research map / Vision-language and multimodal / Cross-modal learning
Histopathology Multi-modal Embedding for Pathology Composed Retrieval
Qifeng Zhou, Wenliang Zhong, Thao Dang, Hehuan Ma, Saiyang Na, Yuzhi Guo, Junzhou Huang
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 progressive knowledge curriculum (ablation) | HOMIE (Qwen3-VL-8B) · Videopath video to text | Recall@1: 18.4 -> 30.7 | Improved |
| add progressive knowledge curriculum (ablation) | HOMIE (Qwen3-VL-8B) · Bookset image+text to image | Recall@1: 46.7 -> 52.7 | Improved |
| add native-resolution processing (ablation) | HOMIE (Qwen3-VL-8B) · Bookset image to text | Recall@1: 25.5 -> 34.4 | Improved |
| add bootstrap data filtering of Quilt-1M (ablation) | HOMIE (Qwen3-VL-8B) · Videopath video to text | Recall@1: 23.4 -> 30.7 | Improved |
| add stain normalization and augmentation (ablation) | HOMIE (Qwen3-VL-8B) · Bookset multi-image to text | Recall@1: 77.7 -> 79.8 | Improved |
| use EOL 'in one word' prompt at inference instead of no prompt (ablation) | HOMIE (Qwen3-VL-8B) · Bookset multi-image to text | Recall@1: 78.0 -> 79.8 | Improved |
| use EOL prompt instead of dropping 'one word' (ablation) | HOMIE (Qwen3-VL-8B) · Bookset multi-image to text | Recall@1: 78.5 -> 79.8 | Improved |
| add progressive curriculum (ablation) | HOMIE (Qwen3-VL-8B) · MMUpubmed image to text | Recall@1: 13.1 -> 13.1 | No clear effect |
| scale backbone from Qwen3-VL-2B to Qwen3-VL-8B (ablation) | HOMIE · Bookset multi-image to text | Recall@1: 75.2 -> 79.8 | Improved |
| scale backbone from Qwen3-VL-2B to Qwen3-VL-8B (ablation) | HOMIE · Quilt-VQA-Red image+text to text | Recall@1: 64.3 -> 57.5 | Hurt |
| swap backbone from Qwen2.5-VL-7B to PathoR1-7B (ablation) | HOMIE · Bookset image+text to image | Recall@1: 49.2 -> 54.1 | Improved |
| apply HOMIE adaptation to PathoR1-7B (ablation) | PathoR1-7B · Bookset multi-image to text | Recall@1: 14.7 -> 78.5 | Improved |
Closest papers by meaning
- PathGen-1.6M: 1.6 Million Pathology Image-text Pairs Generation through Multi-agent Collaboration (ICLR 2025 · similarity 0.86)
- Quilt-LLaVA: Visual Instruction Tuning by Extracting Localized Narratives from Open-Source Histopathology Videos (CVPR 2024 · similarity 0.86)
- Multi-Resolution Pathology-Language Pre-training Model with Text-Guided Visual Representation (CVPR 2025 · similarity 0.86)
- CPath-Omni: A Unified Multimodal Foundation Model for Patch and Whole Slide Image Analysis in Computational Pathology (CVPR 2025 · similarity 0.85)
- PathView-Bench: Can Multimodal Large Language Models Achieve Fine-grained Multiscale Understanding of Pathology Images? (NeurIPS 2026 · similarity 0.85)
- MLLM-HWSI: A Multimodal Large Language Model for Hierarchical Whole Slide Image Understanding (CVPR 2026 · 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.
