Image-Based 3D Reconstruction in Laparoscopy: A Review Focusing on the Quantitative Evaluation by Applying the...

Area of Science:

• Medical Imaging
• Computer Vision
• Robotics

Background:

• Accurate 3D reconstruction is crucial for advanced laparoscopic surgery, including image-guided navigation and robot-assisted interventions.
• Various image-based 3D reconstruction techniques exist, but their comparative accuracy in intraoperative settings requires systematic evaluation.

Purpose of the Study:

• To review and compare the accuracy of different image-based 3D reconstruction techniques for laparoscopic applications.
• To identify the most promising techniques for achieving high-accuracy 3D reconstruction in surgical environments.

Main Methods:

• Systematic literature search (PubMed, Google Scholar, 2015-2023) following established review frameworks.
• Inclusion of articles with quantitative evaluation of reconstruction error (RMSE, MAE) based on Euclidean distance.
• Data generation on reconstruction error for stereo vision, Shape-from-Motion, SLAM, deep-learning, and structured light.

Main Results:

• Reconstruction errors varied widely, from sub-millimeter to over ten millimeters across different techniques.
• Deep-learning and Simultaneous Localization and Mapping (SLAM) demonstrated superior performance under intraoperative conditions.
• Significant variance in results attributed to diverse experimental conditions and lack of standardized error metrics.

Conclusions:

• Submillimeter accuracy in image-based 3D reconstruction for laparoscopy is challenging but attainable.
• Deep-learning and SLAM are promising techniques for high-accuracy intraoperative 3D reconstruction.
• Future research should focus on standardized error computation and realistic ex/in vivo organ models for validation.