Usono Wearable Ultrasound

Framing the opportunity

How might we design wearable ultrasound system that enables consistent, real-time imaging dynamically and help translate complex data into actionable insights?

Solution

We developed Patella Pro with adjustable hinge, a wearable solution designed specifically for patella-related injuries with silicone inner piece. This device stabilizes ultrasound probe placement while allowing natural movement, enabling consistent imaging in dynamic condition.

Complementing this, UltraWise, a machine-learning-driven dashboard that translates ultrasound data into meaningful metrics and clear visual insights to support physiotherapists in understanding injury behavior and guiding treatment decisions.

Design Process

• The process began by evaluating the previous year's machine learning foundation, the "FluidModel" which focused on muscle fatigue. However we shifted strategy towards identifying relevant muscle health markers e.g. fascicle length, pennation angle, etc that physiotherapists actually use for injury recovery and training optimization.

• The team researched over 10 open-source machine learning models to accurately measure and annotate muscle health markers in real-time ultrasound video.

• To validate the final architecture, we applied Tamara Munzner’s Nested Model. This allowed the team to align four critical levels:

  1. Domain Situation: Understanding practitioner and patient needs.

  2. Data Abstraction: Dealing with ordered quantitative sequential data.

  3. Visual Encoding: Implementing interactive line and bar graphs that show trends over time (averaged per 30 frames).

  4. Algorithm: Ensuring the combination of ML models accurately produced the target metrics