Toward real-time remote processing of laparoscopic video.

Area of Science:

• Medical Imaging
• Surgical Technology
• High-Performance Computing

Background:

• Laparoscopic surgery offers minimally invasive benefits but suffers from limited subsurface tissue visualization, causing navigational challenges.
• The daVinci-Si robotic surgical system generates large video data streams (approx. 360 MB/sec), straining bedside processing capabilities.
• Real-time data acquisition, processing, and visualization are crucial for complex surgeries and patient safety.

Purpose of the Study:

• To investigate the use of remote high-performance computing (HPC) clusters for real-time processing of laparoscopic surgical video data.
• To improve navigational capabilities and reduce risks in minimally invasive surgery through image-guided techniques.
• To optimize data transfer and processing times for high-definition laparoscopic video streams.

Main Methods:

• Implemented image processing algorithms on high-definition video data (1920 × 1080 pixels) from a surgical phantom.
• Utilized a message passing interface for video data transfer to remote HPC clusters.
• Measured video frame transfer time and processing rate.

Main Results:

• Achieved a total video transfer time of approximately 53 ms, corresponding to 19 frames per second (fps).
• Demonstrated the feasibility of processing high-definition laparoscopic video data using remote HPC.
• Identified areas for optimization to reach the target processing rate of 30 fps.

Conclusions:

• High-speed networks coupled with remote HPC can enable real-time medical image processing for laparoscopic surgery.
• Optimizing and parallelizing algorithms is essential to achieve the required 30 fps for real-time augmented laparoscopic data.
• This approach has the potential to significantly enhance surgical experiences and patient outcomes.