In eye clinics, Contrast Sensitivity Function (CSF) offers a better assessment of visual function loss than traditional visual chart tests, but its lengthy testing process has made widespread assessment difficult. We developed an Artificial Intelligence Enhanced CSF (AIECSF) method using Bayesian adaptive procedures to efficiently measure CSF. By employing active learning to select the most informative visual stimuli, AIECSF can rapidly estimate four CSF parameters and predict CSF curves across a broad range of spatial frequencies in simulated and psychophysical tests. This approach enhances vision screening, monitoring, and diagnostics in clinical settings, potentially transforming vision evaluation within routine visit timeframes.
