Key result
DeepLab v3 demonstrated superior pixel accuracy and distance measurements compared to U-Net and Obelisk-Net for real-time segmentation of endoscopic scenes during mitral valve repair.
Why the study?
Extracting information from live endoscopic video streams during minimally invasive mitral valve surgery is challenging, and there is no time to manually segment regions of interest.
Does a deep-learning-based approach accurately segment anatomical structures and instruments in endoscopic scenes during mitral valve repair?
Does a deep-learning-based approach accurately segment anatomical structures and instruments in endoscopic scenes during mitral valve repair?
DeepLab v3 provides superior real-time segmentation of anatomical structures and instruments during minimally invasive mitral valve repair.
Supports AI-assisted endoscopic guidance in mitral valve repair; hypothesis-generating and requires prospective validation before clinical use.
Minimally invasive surgery is increasingly utilized for mitral valve repair and replacement. The intervention is performed with an endoscopic field of view on the arrested heart. Extracting the necessary information from the live endoscopic video stream is challenging due to the moving camera position, the high variability of defects, and occlusion of structures by instruments. During such minimally invasive interventions there is no time to segment regions of interest manually. We propose a real-time-capable deep-learning-based approach to detect and segment the relevant anatomical structures and instruments. For the universal deployment of the proposed solution, we evaluate them on pixel accuracy as well as distance measurements of the detected contours. The U-Net, Google’s DeepLab v3, and the Obelisk-Net models are cross-validated, with DeepLab showing superior results in pixel accuracy and distance measurements.
No takes yet. Share an insight, caveat, or question.
Ivantsits et al. (2020) studied Mitral valve repair. Deep-learning-based segmentation models (DeepLab v3, U-Net, Obelisk-Net) vs. Cross-validation between models was evaluated on Pixel accuracy and distance measurements of detected contours. DeepLab v3 demonstrated superior pixel accuracy and distance measurements compared to U-Net and Obelisk-Net for real-time segmentation of endoscopic scenes during mitral valve repair.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: