Tracking and mapping in medical computer vision: A review.

Area of Science:

• Medical computer vision
• Surgical robotics
• Diagnostic imaging

Background:

• Computer vision algorithms are increasingly integrated into clinical systems for diagnostics and interventions.
• Applications range from colonoscopy and bronchoscopy to image-guided surgery and automated instrument motion.

Purpose of the Study:

• To provide an updated review of camera-based tracking and scene mapping in medical computer vision.
• To summarize the state of the art, datasets, and recent algorithmic developments in the field.

Main Methods:

• Systematic review of 515 relevant papers.
• Focus on algorithms for deformable environments, with consideration of rigid tracking and mapping techniques.
• Analysis of datasets, clinical needs, and algorithmic capabilities.

Main Results:

• Identified significant advancements in computer vision for medical applications.
• Highlighted the need for algorithms tailored to deformable environments and improved datasets for training and evaluation.
• Summarized current tracking and mapping methods and their limitations.

Conclusions:

• New or combined methods are essential for clinical applications in deformable medical environments.
• Increased focus on dataset collection for training and evaluation is crucial for future progress.
• Further research is needed to enhance quantification and clinical viability of these technologies.