Related Experiment Video
Updated: Jun 1, 2025
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
Deep learning in surgical process modeling: A systematic review of workflow recognition
Zhenzhong Liu1, Kelong Chen1, Shuai Wang1
1Tianjin Key Laboratory for Advanced Mechatronic System Design and Intelligent Control, School of Mechanical Engineering, Tianjin University of Technology, Tianjin 300384, China; National Demonstration Center for Experimental Mechanical and Electrical Engineering Education (Tianjin University of Technology), China.
Deep learning models, including neural networks and transformers, are key to analyzing surgical workflows in minimally invasive surgery. Challenges remain in surgical data annotation and dataset availability, driving the use of advanced learning techniques.
Area of Science:
- Artificial Intelligence in Healthcare
- Surgical Process Modeling
- Deep Learning Applications
Background:
- Artificial intelligence (AI) is driving innovation in healthcare, particularly in surgical process modeling (SPM).
- Interest is growing in leveraging AI for analyzing complex surgical procedures.
Purpose of the Study:
- Investigate deep learning's role in recognizing surgical workflows.
- Extract reliable patterns from minimally invasive surgery datasets.
- Advance context-aware intelligent systems for endoscopic surgeries.
Main Methods:
- Comprehensive literature search (2018-2024) across major scientific databases.
- Focused on surgical videos with annotations for surgical process modeling.
- Examined specific methods and research outcomes of selected studies.
Main Results:
- Identified 59 relevant articles from an initial 2937.
- Highlighted the use of neural networks and transformers for surgical workflow analysis (SWA).
- Noted a focus on minimally invasive surgeries but a lack of detail in surgical annotation processes.
Conclusions:
- Temporal and spatial sequences are crucial for surgical phase identification.
- Recurrent Neural Networks (RNN), Temporal Convolutional Networks (TCN), and transformers excel at capturing long-range temporal relationships.
- Multimodal data enhances analysis, but challenges persist with clinical knowledge in public datasets and annotation costs, leading to the adoption of various advanced learning methods.
