First stage of the HoloMed project
A. Configuration of test environment, development and implementation of spatial recording/playback system
Due to global issues with the availability of electronic equipment (caused by significant delays in semiconductor production) and the limited ability to purchase Azure Kinect cameras, the decision was made to purchase Intel RealSense cameras instead. Intel RealSense D455 cameras, whose technical parameters align with the project’s requirements, were used to complete Stage 1. According to the hardware’s technical specifications and based on conducted tests, a depth resolution of 1280 x 720 was achieved with a field of view of 87° × 58° (compared to Azure Kinect, which offers 1024 x 1024 at 120° × 120° and 640 x 576 at 75° × 65°). Although these changes were time-consuming, they are now proving beneficial, as the end result is significantly better image quality. The work to adapt the new Intel RealSense cameras to the originally tested infrastructure has been completed. The extension of Stage 1 was primarily due to the need to familiarize with the cameras’ programming libraries and access interfaces to adapt their communication with data collection devices and software.
B. Test environment configuration, development and implementation of virtual patient record application system
At this stage, a retrospective database and access interfaces for an application running on the Hololens 2 device, which is designed to display medical data from patient records, have been prepared. The work on the user interface (UI) of the Hololens 2 application for handling patient medical data has been completed. Simultaneously, methods for accessing the system’s database interfaces are being tested. The database has been implemented using a free MySQL license and a server application embedded in the data server.
C. Sample data set selection and first iteration of segmentation algorithm
The trial data sets were selected using materials provided by partners (Smileline; PEACS BV, Netherlands; Icahn Mount Sinai School of Medicine, New York) and data from public databases such as PubMed, SICAS Medical Repository, OASIS-3 Brains, National Sleep Research Resource, AOCD database, DermNet NZ, and the University of Iowa Skin Database.
D. Configuration of the research environment and preparation of the database and server for data transfer
The purchased server has been configured and connected to a high-speed fiber optic link, enabling other project participants to access the data stored on the server. This setup allowed the collected databases to be uploaded into the system. The server consists of two SSD drives and two network cards, allowing for fast network communication. One LAN interface is used for receiving data, while the other is for sending data, with network traffic being balanced between the loads on the respective network interfaces. Additionally, the server runs two virtual machines that enable the implementation of server applications for communication systems with Microsoft Hololens 2 goggles and data transmission interfaces from Intel RealSense cameras. The entire setup is connected to a fiber optic network with a download speed of 980 Mbps and an upload speed of 60 Mbps. Access to the server’s data is available via VPN and encrypted HTTPS cloud connections. The current configuration allows all project participants to access the collected databases, as well as to edit or upload new data sets. The virtual machines can also be used for numerical calculations and more demanding data transformations.
