Amsler Track

Low cost, high tech and great potential application for better health

Comment

Inspiration

In Australia, 1 of out 7 who aged over 50 may suffer from macula degeneration. Although it is not the most common ocular disease, it contributes half of the blindness in this country.

As a routine practice, eye specialists distribute a paper of grids, the Amsler Grid, to these patients and guide them to self-monitor the progression of their symptoms. For instance, assess at a reasonable reading distance and stable central fixation. However, patients who suffered from macula degeneration usually are visually impaired which is hard for them to fixate steadily and perform the test properly.

The result of the test is supposed to assist practitioners in deciding whether further management is needed. Poor fixation, unreliable test result, and lost paper grids would affect the quality and outcomes of case management.

Since a recent study revealed that over 70% of those aged over 50 own mobile devices, it would be handy if we could rebuild the Amsler Grid test on mobile devices and most of the problem we mentioned could be addressed, and promote better health.

What it does

Our application Amsler Track is a SaaS (Software as a Service), it provides a digital platform for patients and practitioners to monitor macula degeneration by allowing patients to draw and show how and where the distorted vision appear on the screen. Patients do not need to photocopy and bring a stack of paper to their eye specialists during follow up since it is paperless. Therefore, records saving is made easier for long-term monitoring.

More importantly, the test is more reliable by introducing machine learning algorithms. Face recognition and eye-tracking help to make sure patients are fixating properly while doing the test. Hence, even patients whose eyes are not good and fail to fixate voluntarily, our application assist and guide them through.

How we built it

The application interface is built on the Android platform using Java while the face recognition and eye-tracking are built by using OpenCV, an open-source library of Python.

With OpenCV, we train the program by fixating at central and peripheral gazes in return to gather a set of data including the coordinates of the face, eyes, and pupils. Through Random Forest Classifier, the machine learning algorithm, we test and predict the gazes upon different combinations of x and y coordinates.

Challenges we ran into

  • integrating the UI (Java) with backend (Python)
  • Editing videos
  • business model designing
  • functional algorithm

Accomplishments that we're proud of

  • developed a backend machine learning algorithm for eye tracking, face recognition
  • UI designed
  • Agreement on the project and business model

What we learned

  • teamwork - team effort is important while time is limited
  • Business model - how to think commercially for marketing and make a profitable project
  • presentation - how to present interactively, explain the idea and concept, attract audience

What's next for Amsler Track

We will approach the Department of Health and the Department of Human Services for government support. Since the ultimate target of our application is to provide a connection for patients and practitioners, we need government intranet to build a platform on which patients can store and pass their self-monitoring records. Similarly, practitioners can access the patient records via the same channel. It helps to ensure confidentiality, authenticity and integrity of the medical records.

Currently, the interface is built for Android and we build on the iOS platform later.

In the future, the tech applied in this APP such as machine learning to other optical devices for medical purposes.

References

The Department of Health 2008, ‘Visual impairment and blindness in Australia’, viewed 14 July 2018, https://www.health.gov.au/internet/publications/publishing.nsf/Content/CA25774C001857CACA257560001B95AB/$File/4.1.1Vis.pdf.

Boyd, K 2018, ‘What Is Macular Degeneration?’, viewed 14 July 2018, https://www.aao.org/eye-health/diseases/amd-macular-degeneration.

Macular Disease Foundation Australia 2018, ‘Amsler Grid’, viewed 14 July 2018, https://www.mdfoundation.com.au/content/testing-amsler-grid.

Mitchell, P, Smith, W, Attebo, K & Wang, JJ 1995, ‘Prevalence of age-related maculopathy in Australia: the Blue Mountains Eye Study’, Ophthalmology, 102(10), 1450-1460.

Dagnelie, G & Massof, RW 1996，‘Toward an artificial eye’， IEEE Spectrum, 33(5), 20-29.

Share this project:

Updates