Inspiration

The Health industry has a neglected but major problem, ocular disease. Ocular diseases affect over 2 Billion people around the world according to the WHO. Common eye diseases include Cataracts, Age-related macular degeneration, Diabetic retinopathy and Glaucoma.

Because, patients only begin to notice ocular diseases around the age of 35-40 many are too late to get accustomed to a regular lifestyle with improper to near-blind vision. In extreme circumstances if diagnosis is not done in advance correctly and ocular diseases are left untreated severe blindness can occur.

Each year, approximately 6 million people with undiagnosed ocular diseases are too late leading to blindness. We intend to solve this with horus.

What it does

Our app helps serve as a doctor patient management system helping ease the communication in an online setting. The app is tailored to suit all the of an Ocular Disease patient centering its focus on eye health.

The applications core features include:

  • Disease Diagnosis
  • Doctor Patient Communication
  • Treatment/Medication Centers
  • Eye Lifestyle, Care tips

This application is designed to suit the needs of both the patient as well as a doctor helping serve as a complementary application for all to use.

How we built it

Ocular Disease Detection Algorithm: The artificial intelligence skills used in this project is primarily for the machine learning algorithm developed to diagnose the category of corneal ulcers based on images of the human eye.

The different classifications include:

  • Point-like corneal ulcers.
  • Point-flaky mixed corneal ulcers.
  • Flaky corneal ulcers.
  • The application collects the image data from a user and parses it through a convolutional neural network in order to return the possibility of different ulcers. The different API’s used include: Tensorflow, Matplotlib, CV2, OS, Numpy, and Pandas.

Web Application: In order to completely integrate the application with our artificial intelligence technology to be user-friendly and accessible to our users, we developed a graphical front end. This will be accessible by both our doctors and patients to streamline effective communication and for all our additional services. The technology used to create this part of the application include: React, react router, axios, tailwind In order to integrate our system with our machine learning algorithm in order to make it more dynamic, we used additional API’s as well including Spring boot, hibernate, deeplearning4j, h2.

Challenges we ran into

We ran into challenges with the machine learning model. The complexity of the problem required a deeper network with more advanced machine learning techniques and more complexity in data augmentation.

Accomplishments that we're proud of

We are proud of our ability to work as a team and communicate effectively to build the product with 3 different interconnected codebases in 3 different languages to create a comprehensive product/tech stack.

What we learned

We learned how to connect our model to a backend in java, and how to use tensors in java, and create a frontend in react. We additionally learned more about using transfer learning and other deep learning approaches.

What's next for Horus

This application has the potential to revolutionize the Ocular Health Care industry and solve major holes and problems prevalent in the sector. Our core features of the application and the artificial intelligence technology allows us to effectively use our application for Health good. The magnitude of the problem is drastic with over 2 billion people suffering from Ocular Diseases furthermore, the Ocular vision industry is worth over $38.82 billion according to studies. With the huge market for our product, our application distinguishes itself from other competitors and solves the huge problems at hand in the industry. Diagnosis of corneal ulcers is now much more accessible to everyone around the world no matter where they are.

Bringing a service of such magnitude to help the low income- marginalized and disproportionately affected communities truly makes us strive towards our mission.

Built With

Share this project: