Contrast Polarity Differences Across Indoor Subscenes and Outdoor Visual Environments: Implications for Myopia Risk

Outdoor exposure has been shown to help protect against myopia; however, low spatial frequency exposure alone cannot fully explain myopia risk. This study aimed to explore contrast polarity differences between indoor and natural environments.

A total of 53,620 images from the Massachusetts Institute of Technology Indoor Scenes, Mapillary Vistas and Places365 databases were categorised as natural, urban or indoor (with further subdivisions). Local contrast polarity was quantified by calculating the proportions of positive-contrast (ON) and negative-contrast (OFF) pixels after Difference-of-Gaussians filtering at 0.5, 2 and 6 cycles per degree (cpd). Additionally, image features with a 0.4 Bangerter foil, highly aspherical lenslets lens and diffusion optics technology (DOT) lens were also evaluated.

Natural environments exhibited high OFF proportions (0.555 at 0.5 cpd; 0.559 at 2 cpd and 0.555 at 6 cpd). Indoor environments had lower OFF dominance (0.536 at 0.5 cpd; 0.533 at 2 cpd and 0.531 at 6 cpd) across all frequencies (all p < 0.001). Among indoor subscenes, the OFF proportion was lowest in the nursery at 0.5 cpd (0.493), in the hospital room at 2 cpd (0.505) and in the bathroom at 6 cpd (0.508). Contrast-reducing optical interventions increased indoor OFF proportions at 0.5 cpd with scene-dependent magnitudes; however, this effect diminished at higher spatial frequencies.

These findings underscore that built environments, particularly specific indoor subscenes, exhibit lower OFF proportions than natural environments. Enhancing low-frequency negative contrast indoors may guide myopia prevention and public health initiatives.