Reproduce “Decoding the Surgical Scene: A Scoping Review of Scene Graphs in Surgery” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Angelo Henriques, Korab Hoxha, Daniel Zapp, Peter C. Issa, Nassir Navab, M. Ali Nasseri
A scoping review maps the landscape of Scene Graphs (SGs) in surgical artificial intelligence, detailing their applications, methodologies, benefits, and challenges. It positions SGs as crucial "neuro-symbolic guardrails" for future surgical foundation models to mitigate hallucinations, and identifies a significant research gap in real-world 3D/4D dynamic SG generation and validation.
Key insights
- Scene Graphs provide an explicit, relational structure that enables higher-level cognitive reasoning, moving surgical AI beyond isolated object detection to a deeper understanding of interactions.
- SGs are emerging as essential "neuro-symbolic guardrails" to ground probabilistic outputs from Surgical Foundation Models, mitigating hallucinations and improving reliability.
- The field is evolving from foundational GNNs to Transformer-based architectures, increasingly leveraging SGs to guide Generative AI and LVLMs for tasks like Visual Question Answering and surgical simulation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.