Optimizing intraoperative AI: evaluation of YOLOv8 for real-time recognition of robotic and laparoscopic instruments
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.
View abstract on PubMed
Summary
This summary is machine-generated.The YOLOv8 model accurately identifies robotic and laparoscopic surgical instruments for intraoperative artificial intelligence (AI). It achieves high precision and segmentation accuracy, showing promise for real-time clinical applications.
Area Of Science
- Computer Vision
- Surgical Robotics
- Artificial Intelligence in Medicine
Background
- Accurate surgical instrument recognition is crucial for advancing intraoperative artificial intelligence (AI).
- Existing systems require robust models for real-time identification in complex surgical environments.
Purpose Of The Study
- To evaluate the YOLOv8 model's effectiveness in detecting, classifying, and segmenting robotic and laparoscopic surgical instruments.
- To assess the model's performance across various surgical instruments and contexts using a diverse dataset.
Main Methods
- Utilized a comprehensive dataset of over 7,400 frames and 17,175 annotations from multiple sources.
- Trained and tested the YOLOv8 model for instrument detection, classification, and segmentation tasks.
- Analyzed performance metrics including mean average precision (mAP), Dice score, and intersection over union (IoU).
Main Results
- YOLOv8 achieved a mAP of 0.77 for binary detection and 0.72 for multi-instrument classification.
- Optimal performance was noted with approximately 1300 instances per instrument in the training set.
- Demonstrated high segmentation accuracy with a mean Dice score of 0.91 and IoU of 0.86.
- Showed superior performance for robotic instruments compared to laparoscopic ones, attributed to dataset representation.
Conclusions
- YOLOv8 shows significant potential for precise and efficient surgical instrument recognition in robot-assisted surgeries.
- The model's rapid inference speed (1.12 ms/frame) supports its application in real-time clinical settings.
- Findings highlight the importance of dataset diversity and size for model performance in surgical AI.
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.