EUVIMEDEUVIMED entdecken
Crossref

Intelligent Visual Perception in Minimally Invasive Surgical Robots: A Review

Ke Jin, Mengqiu Song, Yunkai Li, Lei Yang

Journal of Machine Learning Advances · 18.06.2026 · 1 (1) · 1-17

DOI: 10.66879/jmla.2026.21 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Machine Learning Advances, online erschienen am 18.06.2026. Verfasst von Ke Jin und 3 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Minimally invasive surgery (MIS) has revolutionized modern medicine by offering patients reduced trauma, faster recovery, and lower complication rates. However, the inherent limitations of MIS, such as restricted field of view, loss of depth perception, and limited dexterity, pose significant challenges to surgeons. The integration of surgical robotics with intelligent visual perception technologies represents a pivotal advancement in addressing these challenges, paving the way for enhanced surgical precision, autonomous assistance, and improved patient outcomes. This comprehensive review systematically explores the state-of-the-art in intelligent visual perception for robotic-assisted MIS, focusing on three core research domains: surgical scene 3D perception, surgical workflow analysis, and surgical scene segmentation. We begin by exploring 2D surgical scene segmentation, highlighting the transition from conventional convolutional networks to advanced Vision Transformers and linear-complexity State Space Models (e.g., Mamba) for precise anatomy and instrument delineation. Furthermore, we analysis 3D perception techniques, including multi-view geometry-based reconstruction and deep learning-driven monocular depth estimation, which are essential for spatial understanding and navigation. In addition, this review then delves into surgical workflow analysis, discussing methods for phase recognition and tool tracking that enable context-aware robotic assistance. Finally, we discuss prevailing clinical bottlenecks, including cross-domain generalization and edge-computing constraints, and provide insights into future research directions, including the integration of generative AI, model compression, and multi-modal data fusion. This review aims to serve as a valuable resource for researchers and practitioners in the field of surgical robotics and computer-assisted intervention.

Zitieren

APA

Jin, K., Song, M., Li, Y., & Yang, L. (2026). Intelligent Visual Perception in Minimally Invasive Surgical Robots: A Review. Journal of Machine Learning Advances, 1(1), 1-17. https://doi.org/10.66879/jmla.2026.21

Vancouver

Jin K, Song M, Li Y, Yang L. Intelligent Visual Perception in Minimally Invasive Surgical Robots: A Review. Journal of Machine Learning Advances. 2026;1(1):1-17. doi:10.66879/jmla.2026.21

BibTeX
@article{Jin2026,
  author = {Jin, Ke and Song, Mengqiu and Li, Yunkai and Yang, Lei},
  title = {Intelligent Visual Perception in Minimally Invasive Surgical Robots: A Review},
  journal = {Journal of Machine Learning Advances},
  year = {2026},
  volume = {1},
  number = {1},
  pages = {1--17},
  doi = {10.66879/jmla.2026.21}
}
Kostenlos starten →