Research map / Image generation / Autoregressive and latent generation
Cell-Type Prototype-Informed Neural Network for Gene Expression Estimation from Pathology Images
Kazuya Nishimura, Ryoma Bise, Shinnosuke Matsuo, Haruka Hirose, Yasuhiro Kojima
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 prototype initialization+modality correction+update (full CPNN) vs prototype from scratch (ablation) | CPNN · TCGA-BRCA | SCC: 0.305 -> 0.338 | Improved |
| use fixed single-cell prototype without modality correction vs scratch (ablation) | CPNN · TCGA-BRCA | SCC: 0.305 -> 0.174 | Hurt |
| add modality correction (alpha, beta) to fixed prototype (ablation) | CPNN · TCGA-BRCA | SCC: 0.174 -> 0.248 | Improved |
| add prototype update (ablation) | CPNN · TCGA-BRCA | SCC: 0.248 -> 0.336 | Improved |
| add deconvolution regularization loss (ablation) | CPNN · TCGA-BRCA | SCC: 0.336 -> 0.338 | No clear effect |
| use 49 fine cell-type labels instead of 8/9 coarse (ablation) | CPNN · TCGA-BRCA | SCC: 0.317 -> 0.338 | Improved |
| use medium (29) vs fine (49) labels (ablation) | CPNN · TCGA-BRCA | SCC: 0.336 vs 0.338 | No clear effect |
| integrate CPNN into STNet/TRIPLEX | STNet, TRIPLEX · CSCC, Her2st, STNet | SCC: e.g. +0.0506 (STNet CSCC), +0.0582 (TRIPLEX CSCC) | Improved |
| integrate CPNN into STNet/TRIPLEX | STNet, TRIPLEX · CSCC, Her2st, STNet | PCC: e.g. +0.0152 CSCC, -0.0083 STNet (TRIPLEX) | Mixed |
| CPNN vs baselines on LUAD PCC | CPNN · TCGA-LUAD | PCC: 0.271 (third; MambaMIL 0.280) | Mixed |
Closest papers by meaning
- Towards Scalable Context-Aware Single-Cell Spatial Transcriptomics Prediction from Histology Images (NeurIPS 2026 · similarity 0.89)
- Learning Relative Gene Expression Trends from Pathology Images in Spatial Transcriptomics (NeurIPS 2025 · similarity 0.87)
- Diffusion Generative Modeling for Spatially Resolved Gene Expression Inference from Histology Images (ICLR 2025 · similarity 0.87)
- Beyond ViT Tokens: Masked-Diffusion Pretrained Convolutional Pathology Foundation Model for Cell-Level Dense Prediction (NeurIPS 2026 · similarity 0.86)
- STimage-1K4M: A histopathology image-gene expression dataset for spatial transcriptomics (NeurIPS 2024 · similarity 0.86)
- GeneFlow: Translation of Single-cell Gene Expression to Histopathological Images via Rectified Flow (NeurIPS 2025 · similarity 0.86)
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.
