Pixel-wise recognition for holistic surgical scene understanding

Nicolás Ayobi, Santiago Rodríguez, Alejandra Pérez, Isabela Hernández, Nicolás Aparicio, Eugénie Dessevres, Sebastián Peña, Jessica Santander, Juan Ignacio Caicedo, Nicolás Fernández, Pablo Arbeláez

Abstract: Highlights•The GraSP dataset has multi-granular annotations for 4 surgical recognition tasks.•GraSP models the hierarchical complementarity in surgical workflow analysis tasks.•TAPIS surpasses alternative models in all GraSP tasks with better generalization.•TAPIS consistently achieves state-of-the-art performance on alternative benchmarks.•TAPIS leverages multi-task signals to enhance per-task performance.

External IDs:doi:10.1016/j.media.2025.103726