Natural interaction based visual field prediction in virtual reality - a pilot study

PVFM predicts visual field boundaries from natural VR interactions in retinitis pigmentosa patients, matching kinetic perimetry for largely concentric fields with radial-weighted IoU 75.4%–91.0%.

• Why it matters: Conventional kinetic perimetry is burdensome, limiting engagement and feasibility for preventive exercises or rehabilitation; PVFM targets an intuitive, potentially passive VR-based alternative.
• What they did: In a pilot study with n=4 retinitis pigmentosa patients (eight eyes), PVFM was developed and compared against kinetic perimetry on the Octopus 900 using radial-weighted Intersection over Union and boundary angular distances.
• The result: For concentric fields (n=6 eyes), mean absolute angular distances were 0.31°–2.68°; for pronounced peripheral non-concentric areas (n=2 eyes), IoU dropped to 29.6%–35.7° with 17.5°–22.3° distances, indicating feasibility with a key limitation.