← Back to all news
arXiv Computer Vision By Feixing Chen, Hao Lu, Lin Luo, Yan Xu

Semantic-Aware Subgraph State Space Model for WSI Classification in Histopathology

Read the original on arXiv Computer Vision →

The paper introduces the Semantic-Aware Subgraph State Space Model (SASG-SSM) for classifying whole slide images (WSIs) in histopathology. It groups spatially connected patches into semantic-aware subgraphs that preserve internal spatial organization, then uses a graph neural network encoder combined with a Mamba-based state space encoder to capture both local and global contextual information. Experiments on four WSI subtyping datasets show consistent performance gains over state‑of‑the‑art methods, with additional robustness in small‑cohort and few‑shot scenarios.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computer Vision.

Related stories

Hugging Face Trending Papers
Sep 3

Semantic-Aware Subgraph State Space Model for WSI Classification in Histopathology

The paper introduces the Semantic-Aware Subgraph State Space Model (SASG-SSM) for classifying whole slide images in histopathology. It groups spatially connected patches into semantic subgraphs that preserve internal spatial organization, then uses a graph neural network encoder combined with a Mamba-based state space encoder to integrate local and global contextual information. Experiments on four WSI subtyping datasets show consistent performance gains over state‑of‑the‑art methods, with additional robustness in small‑cohort and few‑shot scenarios.

benchmarks
More like this →
arXiv AI
Sep 10

Semantic Context-aware mOdality fUsion Transformer (SCOUT): A Context-Aware Multimodal Transformer for Concept-Grounded Pathology Report Generation

SCOUT is a concept‑grounded multimodal transformer that generates whole‑slide pathology reports by integrating local histological patterns, whole‑slide context, and expert‑curated diagnostic concepts. It uses evolving visual representations and recursively updated slide‑ and concept‑conditioned representations, with separate attention pathways during decoding that are fused adaptively for each token. Evaluated on TCGA‑BRCA, HistAI, and REG‑2025, SCOUT outperformed existing methods, improving BLEU, METEOR, and ROUGE‑L scores and raising the Clinical Report Quality Score on REG‑2025.

By Suryakant Singh, Saarthak Kapse, Joel Saltz, Prateek Prasanna
llmsmultimodal
More like this →