Corneal Topography

Using a smartphone's camera to build a mass screening device to generate the corneal topography through which corneal diseases can be screened out, especially in villages and remote areas

Teammates:

Guruswamy Revana Navya Tatikonda

Sabari Prabaaker

Sahithi Kasaraneni

Vivek Kiran

Mentors:

Dr. Pravin Krishna Vaddavalli

Ashish Jain

Chinmae Gorla

Koteswara Rao

Suril Mehta

My Role:

Software Developer (Image Processing)

Deliverables:

A functional hardware prototype with software integration for detecting the deviated pattern in the eye's corneal ring through which corneal diseases can be diagnosed.

Time Period:

June 2017 [7 Days - Full time]

@ Engineering the Eye 2017 - By L.V. Prasad Eye Institute & MIT Media Lab

Background

Corneal Topography can be used to detect astigmatism, corneal diseases such as Keratoconus - a progressive, non-inflammatory corneal disorder resulting in impaired vision, which can be caused by overexposure to UV rays from the sun, excessive eye rubbing, etc., and is very common in children of age 14-18 years.

The medical device which is used to screen corneal topography at eye hospital would cost ~ 20,00,000 INR (~30000 USD). Planting and maintaining these devices in villages and remote areas are nearly impossible.

Vision

Using a smartphone's camera to build a mass screening device to generate the corneal topography (mapping the surface curvature of the cornea, the outer structure of the eye) through which corneal diseases can be screened out. This can be used for initial screening at remote villages and at primary health centers by workers who don't need much technical expertise.

Proposed Solution

A mobile application which the ground-level workers can easily use to get a clear picture of the eye, which can be processed to find the possibility of corneal disease. As the device is aimed to solve the initial screening at villages, the image processing algorithm is focused on reducing the false-positive cases rather than 100% perfection which the medical devices deliver.

Stepwise approach -

1. Test cases - Collected about 54 test cases with different eye conditions, some of which include corneal infection and damages.

2. Generation of test images - Through algorithms, generated around 500 images for testing purposes.

3. Sharpness - Images were sorted based on their sharpness with an algorithm in MatLab

4. Image Processing - Using MatLab

5. Efficiency - For better optimization, worked on Ellipse detection through which a more accurate center could be found.

6. Hardware Design - To achieve uniform illumination, a hardware prototype was fabricated using a 3D printer that can be attached to any smartphone camera and used as a projection to the eyes.

7. Mobile Application -

To maintain patient records, the application is designed to be intuitive and to be used by field-level workers without much training.

After adding the patient details, the fieldworker can take a picture of the patient's eye using the smartphone camera attached with a Placido ring projection.