Stroke Patient Monitoring System

The Biometrics Technologies laboratory at the University of Calgary has worked with the Foothills hospital to collect data that could help medical professionals better identify patients who are more at risk of having a stroke. One of the main identifying signs is physical impairments and abnormal movements or sudden numbness of the body. Our monitoring system provides a base for data collection through video recordings that can be correlated with numerous other detection systems to provide a more complete profile of a patient’s risk level.

Mobility:

One of our main areas of focus was to ensure that the control interface for the cameras was both easy to use and portable. We were required to visualize a high-paced hospital environment where every minute can become crucial to saving a person’s life. Our tablet design neglects the complexity and confusion of a professional surveillance system and focuses instead on basic controls and functions that a nurse or doctor would need all at their fingertips. The 4 hour battery life allows the staff to control the cameras as they move about the unit or between patients with ease.

Error Checking:

The user program is designed so that it regularly runs diagnostics on camera connectivity. Medical staff should not have to take time out of their busy schedules to investigate when a camera goes down. If a camera has a malfunction (due to power failure, full storage etc.) An error message will appear on the screen notifying the user which camera has lost connection and possible fixes. Once all checks are made, the software will attempt to re-establish the connection automatically.

The design for our control interface was focused on simplicity and user friendliness. We wanted to make sure that someone who uses the interface for the first time can tell within seconds how everything works simply by looking at the screen. Below, we detail some of the innovative features of our design that were inspired by the idea of simplicity.

• Interoperability: Our python interface was programmed in such a way that it is very simple to duplicate on other devices. Currently there are two versions of the GUI: the tablet version, and a PC version as shown in figure 3. Both can be used simultaneously, meaning if synched with the correct cameras, any status changes or commands sent from one UI in one device, will reflect in the other UI on the other device.
• Expandability: One of the main purposes of our design is for it to be a basis for incorporating multiple data collection techniques moving forward. The idea is for a doctor to collect video recordings in unison with data from a pressure mattress, thermal imagers, and gesture recognition techniques. This along with patient vitals will give medical staff a broader picture of a patient’s health and their overall risk of having a stroke. Unlike other commercial applications, our design can be easily modified to be installed on multiple devices, used with different camera types, and increase the number of cameras that can be controlled at a time.

Our design solution is effective because it showcases three major deliverables that were requested by the stroke unit representatives and sponsors themselves.

• Easy to use Interface: The colour indicators on the display as well as large buttons and labels are key in making sure that the interface display on both our tablet design as well as on a PC are sharp, easy to read, and complete the function that they are supposed to when selected.
• Storage monitoring: One of the features of our graphical user display is the storage space indicators on each of the cameras. Since all the cameras utilize local recording on their SD storage, we wanted to ensure a way for the user to keep track of how much storage space is left in each cameras. As shown for camera 3 in figure 4 below, the Free Space value has reached 0B out of 13.8GB available, notice the RECORD button is also disabled. Storage space remaining is important to keep in mind when planning to record a patient for multiple hours or days at a time. The storage updates in real time as the camera records, processes, and stores video files.
• Privacy: What differentiates our design from an ordinary security camera setup is the absence of a live video feed. Maintaining patient privacy is very important, particularly in spaces where diagnosis, medical exams, and other personal information is exchanged between patients and doctors or nurses. Instead, our interface has status indicators that clearly demonstrates when a camera is recording as well as a “hide” functionality where the user can pause the recording for a period of time and unpause it without having to create a new video file or using up any additional storage space.

From the very beginning of our project, our team devised a plan to ensure we could remain on track with progression on all deliverables that were required. This involved making use of multiple organizational strategies from team calendars, to gantt charts, to weekly update meetings. To validate our design progression and ensure that each step we took was contributing to the completion of the project, two very important actions were taken:

• Weekly Sponsor Meetings: One of the situations our team wanted to avoid is to make significant progress on our project, only to later realize that a deliverable was not met or that the design our sponsor had in mind was different than what we were delivering. Frequent communication within the team as well as with our stakeholders was crucial despite the difficulties imposed by the COVID pandemic. Utilizing virtual meeting applications, we organized weekly meetings with the team and stakeholders involved. This gave us an opportunity to discuss everything from issues encountered in the construction, to design alternatives, and testing results. Through this method, our team was able to stay on schedule with the development of our project and maintain flexibility when any changes in the design were required or issues had to be addressed.
• Testing Strategies: Due to the inability for our group to meet in person due to the pandemic, formatting a complete testing plan was quite a challenge. Ultimately, our plan was split into different stages. On the software side, tests were first done at home while the program was running on a local PC with one or two cameras. Eventually all six were tested. Parallel to this, the hardware build of our tablet was also tested for overall performance, power consumption, and responsiveness. This test was performed by some of our team members located in BC. Once both tests yielded acceptable results, both the hardware and software components were combined and tested on location at the University. Special focus was given during these final tests on following all safety measures strictly. Below is a table describing some of our team’s final test results.

• Cost: There are many video security management softwares and applications that can be downloaded or purchased however many of these involve a subscription service and require constant communication with an external server provided by the company. Our application allows for full control of the cameras and since the recordings are stored locally, require no connection to any external service, or any recurring costs. Additionally, this setup of local recording requires fairly inexpensive smart cameras compared to professional security cameras which are required when using other video management softwares like the ones described previously.
• Python Based: A significant benefit of our program is that it was coded entirely using python language. Python language is one of the most versatile yet easy to learn because its syntax is based on day-to-day english language structures. This allows tasks performed through python to usually be shorter compared to other languages such as java. Since it has grown in popularity over the past 2 decades, it is very easy to find an extensive variety of support libraries, open source programs, and even online communities of developers.