OcularEVM

OcularEVM is developing a non-contact, non-invasive way to estimate intraocular pressure (IOP) from ordinary video. Rather than touching the eye, it reads the pressure-dependent dynamics of intraocular structures and inverts them to recover a clinically meaningful quantity.

It is being developed by Leon Moriguchi — Investigator & Founder, Institute of ΛάΓ — selected for the 2026 MITOU IT Program (IPA), Japan's national program for exceptional young software creators.

When the eye finishes a saccade, its internal structures don't stop instantly. The crystalline lens, suspended by zonular fibers, overshoots and settles in a brief damped oscillation — a post-saccadic oscillation (PSO). High-speed Purkinje-image measurements show this settling behaves like a classical damped harmonic oscillator, with the lens oscillating at roughly 20 Hz over the 50–60 ms following a gaze shift [1].

The shape of that oscillation — its amplitude and how quickly it decays — is governed by the mechanical state of the eye, which depends on intraocular pressure. Recent fluid–structure-interaction simulations make the link explicit: IOP plays a central role in how the lens overshoots, and the same work points to Purkinje-image sensing as a route to non-invasive IOP estimation [2].

OcularEVM treats this as an inverse problem: given the observed oscillation, recover the pressure that shaped it.

The system pairs a physically grounded model of intraocular dynamics with a learned component that absorbs individual variation, and is being validated against clinical reference measurements. The specifics of the sensing setup and inference pipeline are held as proprietary intellectual property.

Glaucoma is the leading cause of irreversible blindness worldwide, and it progresses silently — the visual field narrows without subjective symptoms until damage is advanced. An estimated 76 million people were affected in 2020, a figure projected to reach roughly 112 million by 2040 [3].

Intraocular pressure is glaucoma's principal modifiable risk factor, yet IOP fluctuates through the day and its peaks often fall outside clinic hours — so a single in-office reading can miss them entirely. A camera-based screen that works outside the clinic would change the cadence of measurement itself: catching diurnal peaks, supporting quantitative monitoring of treatment, and — at scale — building population-level data on IOP dynamics that barely exists today.

OcularEVM is at an early validation stage. The current focus is stabilizing the capture setup and testing whether the camera-derived signal tracks reference measurements across a range of pressures.