Frequently asked questions

What does this app do?


This app is designed to provide a simple tool for measuring a subject's head pose with respect to a stationary reference frame. Specifically, the app attaches a coordinate frame to the head of the subject (shown in the following diagram) and measures the angles that relate the head coordinate frame to a stationary reference coordinate frame that is aligned with the gravity vector. the following image shows an example of the output of the app. The app obscures the face of the subject to maintain their privacy. In this image you can see the three angles that specify the head pose displayed on top of the screen.




How do I operate this app?


Follow these simple steps:

  1. Put the iPhone on a tripod in portrait mode (i.e., vertically mounted)
  2. Launch the app
  3. Ask the subject to stand in front of the front camera
  4. Adjust the subject or the camera such that subject's shoulders are parallel to the iPhone
  5. Tap on the "Reset" button on screen
  6. Ask the subject to hold a relaxed stance
  7. Tap on the camera icon to capture the measurements
  8. Save the captured image




Does the app send these images to a server outside the device?


No. All data and images are processed locally on the device and no information is shared with any entity. We do not have access to any of the data that is captured using this app.




Who created this app?


  • App development: Ramin Rezaiifar, PhD
  • Idea of using the app in the context of movement disorder: Dr. Khashayar Dashtipour, PhD, MD




How does this app work?


The app uses video frames from the front camera of the iPhone together with the depth info from the True Depth camera in front of the iPhone. The RGBD image is given as input to a convolutional neural network (CNN). The outputs of the CNN are three Euler angles that transform the stationary coordinate system to the head coordinate system (as described in the first Q&A). The app then uses augmented reality (AR) to superimpose the head coordinate system on top of the camera output frames.




What is the advantage of using this app vs. estimating the angles by a human?


  • Estimating 3-dimensional angles by a human observer is error-prone: small angles are difficult to estimate by humans. The task becomes even more difficult when a human observer is asked to estimate 3-dimensional angles where perception of one angle may affect the estimate of the other two angle
  • Estimating 3-dimentional angles by human observers are not repeatable:
    • A human observer may estimate the same angle differently on different days.
    • Estimates provides by different human observes typically do not agree with each other
  • This app provides an objective measurement for the head pose angles that are consistent and repeatable





Contact

For more information/help, please email us at: