Scope of super-resolution in central vision

The term super-resolution refers to computational enhancement of detail in sets of similar images, particularly undersampled images. Undersampling is a key feature of visual disorders associated with neuroretinal matrix damage. It is conceivable that a neurophysiological equivalent of super-resolution might be recruited in these settings. The need for sets of similar images might be satisfied by augmenting retinal image movements.

Monocular letter acuity thresholds were measured by computer graphics in normal subjects, with and without controlled test target movements. Various degrees of neuroretinal matrix defects were simulated by superposing stationary masks on the test targets.

Static acuity thresholds increased monotonically with increased mask densities. The addition of target movement partially counteracted the threshold increment. The effect was barely measurable at small mask densities and increased monotonically to a gain of approximately two lines on a decimal acuity scale with masks simulating loss of 90% of neural matrix elements. Similar results were obtained from a patient with an acquired neuroretinal matrix defect, attesting to the validity of the mask experiments.

Super-normal retinal image movements have subtle effects on visual resolution in normal eyes, whereas eyes with neuroretinal matrix defects may improve markedly. This finding may have implications for the development of aids for the visually handicapped.