Altaravision, based in Apex, NC, has created an FDA compliant medical imaging device called NDŌʜᴅ. This portable, high definition endoscopic system uses an application running on a MacBook to display a real-time video of a swallow test on a patient.
In the past, this video was typically recorded and then manually analyzed to determine the amount of food that was actually swallowed by the patient. This painstaking process of calculating residual food particles from the swallow test was time consuming and introduced human error.
Altaravision came to Oak City Labs to automate the video analysis in order to save time and reduce mistakes.
Leveraging computer vision techniques, we set out to have the NDŌʜᴅ application process each frame of the footage, categorize pixels by color and produce a single numerical value representing the residual food material left in the throat after swallowing. Additionally, the application provides this feedback in real-time as the test is being performed.
Once the pixels are categorized by colors, the application builds an overall histogram of the colors and determines the percentage of residual food in the frame.
As such, the test administrator is able to not only visually see the results of the test in real-time, but also review the video frame by frame, using the snapshot analysis to accurately assess the patient.