Glaucosim Tonometer estimates eye pressure from a 60-second front-camera video. Glaucosim VF runs a 24-2 visual field on the patient's own laptop. Both are made for weekly use with a clinician watching the trend, and both build on tests clinics already trust.
Standard care sees a glaucoma patient twice a year. By the time progression shows up on the next visual field, fibres are already gone. Glaucosim runs two of the most informative tests every week, on devices the patient already owns.
The front camera follows two pulses at once: the faint beat of the iris as blood enters the eye, and the pulse in the facial skin. Their ratio holds steady when lighting or heart rate drift, and a model trained on many patients turns it into a pressure reading in mmHg.
The same 54-point central grid the clinic uses, with a smart staircase (ZEST) that finds each threshold in fewer flashes. It returns the standard summary scores and reliability checks, from the same family as the SITA test on the clinic's perimeter.
Every reading carries its own quality score. A capture the camera cannot trust never reaches the clinician's trend line, so testing often stays honest instead of just frequent.
No headset, no chin rest, no kiosk. The patient runs the test from the laptop they already use at home, with reliability indices the clinician can audit.
Hold a credit card to the screen to anchor physical pixel size. Stored per device. Never repeated unless the device changes.
The front camera measures the patient's distance to the screen with a face-mesh pinhole model. Stimuli appear on the calibrated grid; the patient presses the spacebar at each one.
Overall sensitivity, the pattern of any loss, and the reliability checks sync to the clinician's dashboard, with a trend line against earlier weeks.
An ordinary front camera. The iris is tracked frame by frame and its faint, heartbeat-driven pulse is pulled out, then read against the pulse in the face to give a pressure estimate.
The face-mesh distance tracker keeps the patient in the capture band. The eye-target oval offsets to whichever eye is being tested so the patient doesn't drift the camera sideways.
One signal is the pulse of the iris; the other is the pulse in the cheek and forehead. Reading them as a ratio cancels the day-to-day swings in lighting and blood pressure.
The model converts the ratio to millimetres of mercury and scores the take. Blurry or shaky captures are thrown out before they reach the trend.
Both tests rest on measurements that already have an evidence base. The ask is not to believe a new biological claim. It is to accept that a known measurement can run on a device the patient already owns.
Phase 1 comparison studies for both tests are in design at an academic glaucoma center: agreement against the clinic standards (HFA SITA-Fast for the field, Goldmann for pressure), plus repeat in-clinic testing for reproducibility.
A single dashboard where every patient's weekly Glaucosim VF and Tonometer captures land. Trend lines, reliability indices, and quality gates exposed so you can decide what to trust before you decide what to do.
Master dashboard with code-only enrolment, demographic capture, training vs test cohort split, and per-subject capture archives. Every variable saved with the capture; nothing reconstructed from logs.
"Two appointments a year cannot catch a disease that damages the optic nerve silently, fibre by fibre, between visits. We built two home tests for the cadence the disease actually needs."
