NTT, Olympus Uses IOWN APN Tech in World’s First Cloud Endoscopy System

NTT Corporation and Olympus Corporation have jointly established a cloud endoscopy system using the IOWN APN technology and demonstrated that the APN technology can solve network issues in realizing the cloud endoscopy system. 

This follows the start of their joint experiment in March of the world’s first cloud endoscope system, which processes endoscopic videos on the cloud. 

Moving forward, the two companies will collaborate for the full-scale implementation of the cloud endoscopy system and enable performance and functional improvements of endoscopy systems through the cloud, contributing to society by improving access to advanced medical care.

The companies’ demonstration experiment combined the advanced technology of Olympus’ endoscopes with NTT’s high-speed, low-latency APN technology aimed at realizing real-time endoscope video processing in the cloud. A network that connects to the endoscope device and the cloud is needed to achieve image processing on the cloud. 

Due to the nature of endoscopy procedures, the video captured by the scope needs to be processed and displayed without delay. If the video data (4K/60fps) is delayed even by a few frames, it will cause discomfort to the operator.

The experiment connected a video processing server (“server”), distanced 150km away from the device that inputs endoscopy images (“edge device”), to the APN and configured it as the cloud endoscopy system.

For the experiment, NTT provided the APN environment and evaluated the APN connectivity and network quality. Olympus provided the endoscope device and evaluated the performance of the cloud endoscopy system and the processing capability of the software.

In the experiment, the video captured by the endoscope was sent to the edge device via an endoscope processor and then transferred without any compressing to the server via the APN. The server processed the video using AI and other technology and sent the processed video back to the edge device. Finally, the monitor connected to the edge device displayed the processed video for the endoscope operator to view.

This experiment was conducted to achieve a data transfer latency within one frame (under 16 milliseconds), considering the distance of approximately 150km. As a result, the latency was 1.1 milliseconds, demonstrating that it was possible to transfer data at 1/10 of the target latency.

The experiment used a local configuration with a distance of 5m (assumed to be processed within the endoscope device) and an APN configuration with a distance of 150km, which each processed the endoscopy video. The capture cards recorded the videos displayed on the endoscopy operator’s monitor, measured the network’s data latency, and compared the two videos. 

A visual review by the operator confirmed that there was little difference in latency and fluctuation, demonstrating that latency in the APN would not be an issue in video processing.

Video processing between an edge device and a server separated by 150km was confirmed to be possible. This demonstrates the possibility of combining the video processing of hospitals in large areas to a remote server, which is a valuable insight for the implementation of a cloud endoscopy system in the future.

Through this initiative, NTT will establish a reference model for medical device networks that solves the various technical issues involved in the cloud computing of medical devices, and Olympus will contribute to the establishment of optimal networks and reference models for cloud endoscopy systems and other such systems.

Moving forward, the two companies will continue to study technical issues for social implementation of cloud endoscopy, such as advanced security measures for medical data in network transmission and real-time remote diagnosis and treatment by sharing image information among multiple hospitals