👁 VisionCare® Platform

During my stay at the Fuller Lab in the Chemical Engineering Department at Stanford University, I was in charge of a project in collaboration with Johnson & Johnson VisionCare®. As Johnson & Johnson has a large market segment for contact lenses in the United States, it seemed plausible they asked for a thorough tear analysis and the search for the root of the question "Why do people get dry eyes while wearing contact lenses?" For this reason, my colleague Sophie Lohmann and I developed a hardware-software system that would diagnose fluid movement on the surface of the eye while wearing contact lenses.

About 125 million people worldwide wear contact lenses [1]. Each year, a considerable fraction of these visit their doctor with the complaint of dry eyes, mentioning everything from burning, redness, grittiness, itching, and even excessive tearing. Considering that dry eye disease has increased exponentially over the past number of years, the demand for diagnosing and treating patients is equally rising. Efforts from industry and research facilities have been taken to attempt combatting this issue known as ”contact lens-associated dry eye” (CLADE) [2].

In general, tears have to cover the entire preocular surface to function properly. After a blink, it is re-established but short-lived and takes about 15–40 seconds to rupture and dry spots to appear [3]. When wearing contact lenses, the overevaporation of tears dramatically reduces this number and causes many wearers discomfort and itchiness due to lack of moisture [2]. Changing the lens material, care systems, and lubrication agents so far in an effort to alleviate wearer discomfort has, however, not deemed successful [4]. In other words, there is no cure for CLADE yet. Therefore it is key to enable a fast and quantitative way to characterize the tear film thickness throughout space and time.

In this project, the development of a setup and a user-friendly software to enable time-efficient visualization of tear film dynamics for contact lens wearers is thoroughly assessed. In a first step, the optical setup for capturing preocular thin film interference patterns on the human eye is exhibited. Compared to a previous setup, we demonstrate the dramatic improvement of video quality for identifying the interference patterns. Moreover, the design of a novel GUI-based software prototype to semi-automatically pinpoint a color from the region to a tear film thickness is described. So far, noise-corrupted images and head movement have prevented the time-efficient temporal analysis of these patterns. However, the new pipeline offers integrated de-noising and stabilizing features, an interactive process to create unique color-thickness correspondences, and an efficient plotting method not only to display the spatiotemporal thickness evolution over entire regions, but also of individual points.

As seen in the animation, a 15-cm-diameter dome light is used to illuminate the pre-ocular surface of a human subject resting their head on the head-chin rest. A camera with a high sensitivity combined with an appropriate focusing lens – records the images. The zoom and focus module result in a working distance (i.e. distance between lens and focus plane) of approximately 25 cm. In order to increase the contrast of fringe patterns, a tri-band pass filter is placed in front of the camera and ambient light is blocked using a curtain covering the test subject. Adjusters in the vicinity of the camera as well as a rotation module between post and camera allow to align eye, light and camera.

In the following, you'll see a screenshot of what the GUI of the "Fringalyzer"I designed looked like. In principal, one would have to upload a video of the interference on an eye to the program and by some image segmentation, it would be able to infer the thickness of the specific fringe on an eye according to its color on the spectrum. This would allow the creation of a 3D model of the tear film and ultimately provides the ophthalmologist with the profile to help him with diagnosis.

Ultimately, a long-term goal of this project would be a live capturing of a patient’s tear film profiles with an inexpensive device at the doctor’s office for dry eye diagnosis. For this purpose, the data would not only need to be parsed more efficiently, but also implemented into the frame acquisition software of the respective camera manufacturer using the API’s of the Software Development Kit. After a sufficient amount of diagnoses have been made, efforts in research and industry could be weighted toward designing novel contact lenses that inhibit the effect of dry eye and would attract more users on an annual basis.

References

[1] S. B. Eiden. Consider Treating Lens-related Dry Eye With Pharmaceuticals. Contact Lens Spectrum, 2011.

[2] P. Ramamoorthy, L. T. Sinnott, and J. J. Nichols. Contact lens material characteristics associated with hydrogel lens dehydration. ... and Physiological Optics, 30(2):160–166, Mar. 2010.

[3] An application of detection function for the eye blinking detection. IEEE, 2008.

[4] J. J. Nichols and L. T. Sinnott. Tear film, contact lens, and patient-related factors associated with contact lens–related dry eye. Investigative ophthalmology & visual ..., 47(4):1319, 2006.