CODAZEN_
Insights/Prototypes/Gaze-weighted acuity field
In buildPerception Simulation

Gaze-weighted acuity field

A running model of foveal falloff. Everything outside the simulated gaze point degrades the way real peripheral vision degrades — in resolution and in positional certainty.

Canvas 2DWebGL (planned)
Simulated fovea · 42% of frame width
Sample grid · 22px pitch
Controls

How it works

A sample point’s radius scales with a Gaussian centred on the gaze position, approximating cone density across the retina. Away from centre, samples also gain positional jitter — peripheral vision loses location accuracy before it loses the ability to detect that something is there.

The gaze point follows a slow Lissajous path rather than recorded eye-tracking data, so the motion is illustrative, not empirical.

What we are testing

Whether teams can predict missed interface elements from a simulation instead of a user session.

How far acuity falloff alone explains attention failures, before adding saliency or task models.

Whether the output is legible enough for designers to act on without a human-factors background.

Known limitations

Gaze path is synthetic. Real fixations are saccadic and task-driven, not smooth.

No colour-sensitivity or contrast-adaptation model yet — falloff is monochrome in effect.

Not validated against eye-tracking data. Treat output as a hypothesis generator, not a measurement.