Research map / Label-efficient and robust learning / Recognition and long-tail
HiFusion: Hierarchical Intra-Spot Alignment and Regional Context Fusion for Spatial Gene Expression Prediction from Histopathology
Ziqiao Weng, Yaoyu Fang, Jiahe Qian, Xinkun Wang, Lee A D Cooper, Weidong Cai, Bo Zhou
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 2x2 decomposition to full-spot input (ablation) | HiFusion · HER2 (2D) decomposition levels | MSE: 0.5718 -> 0.5561 | Improved |
| add 7x7 level on top of 1x1+2x2 (ablation) | HiFusion · HER2 (2D) decomposition levels | PCC: 0.4902 -> 0.4961 | Improved |
| add 4x4 level to 1x1+2x2+7x7 (ablation) | HiFusion · HER2 (2D) decomposition levels | MSE: 0.5459 -> 0.5477 | No clear effect |
| use 1x1+4x4+7x7 instead of 1x1+2x2+7x7 (ablation) | HiFusion · HER2 (2D) decomposition levels | PCC: 0.4961 -> 0.4756 | Hurt |
| add feature alignment loss (ablation) | HiFusion · HER2 (2D) feature alignment | MSE: 0.5642 -> 0.5459 | Improved |
| add feature alignment loss (ablation) | HiFusion · ST-Data (2D) feature alignment | PCC: 0.5559 -> 0.5613 | Improved |
| reduce key/value spot tokens by pooling to 2x2 (ablation) | HiFusion · HER2 (2D) | MSE | Improved |
| enlarge neighbor patch beyond 2x spot size (ablation) | HiFusion · HER2 (2D) | MSE | Hurt |
| use ASIGN-3D multi-sample 3D strategy vs single-section intra-sample training | ASIGN-3D · HER2 (3D sample-specific) | MSE: 0.2846 -> 0.4163 | Hurt |
| HiFusion vs EGN in 3D sample-specific evaluation | HiFusion · HER2 (3D sample-specific) | MSE: 0.2917 -> 0.2846 | No clear effect |
Closest papers by meaning
- Diffusion Generative Modeling for Spatially Resolved Gene Expression Inference from Histology Images (ICLR 2025 · similarity 0.90)
- Scalable Generation of Spatial Transcriptomics from Histology Images via Whole-Slide Flow Matching (ICML 2025 · similarity 0.89)
- Towards Scalable Context-Aware Single-Cell Spatial Transcriptomics Prediction from Histology Images (NeurIPS 2026 · similarity 0.89)
- MERGE: Multi-faceted Hierarchical Graph-based GNN for Gene Expression Prediction from Whole Slide Histopathology Images (CVPR 2025 · similarity 0.89)
- Preserving DEG Rankings for Gene Discovery in Histology-Based Spatial Gene Expression Prediction (NeurIPS 2026 · similarity 0.89)
- M2OST: Many-to-one Regression for Predicting Spatial Transcriptomics from Digital Pathology Images (AAAI 2025 · similarity 0.88)
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.
