Abstract
Improving visual resolution after prolonged exposure to blurred images has been known as one of the adaptation phenomena, called as blur adaptation (Pesudovs & Brennan, 1993). Mon-Williams et al. (1998) examined temporal variation of spatial frequency sensitivity after a period of time without refractive correction, and concluded that this variation (i.e. adaptation) reflects changes of gain in spatial frequency channels in the visual system. If their hypothesis is true, there exist differences in the aftereffects of adaptation to different spatial frequency structures.
In this study we investigated whether the visual system can appropriately adapt to natural images blurred by two kinds of filters. The first case was blurred by the filter that the slope of amplitude spectra becomes steeper than -1 which is the typical slope of amplitude spectra of natural image. Webster et al. (2002) demonstrated the existence of the effect of blur adaptation by using this kind of filter. Other case was blurred by the Gaussian filter. Images blurred by these two filters have different spatial frequency structures and appearance. Subjects adapted to either blurred image and amounts of the aftereffect were measured with both blurred images by matching tasks. As a result, the effects of adaptation were superior when the adaptation and test stimuli were blurred by the same type of filter. Results were duplicated by the blur estimation model with Laplacian pyramid structure (Burt & Adelson, 1983) as changes in spatial frequency gains of the model according to adaptation states.
PesudovsK.BrennanN. A. (1993). Decreased Uncorrected vision after a period of distance fixation with spectacle wear. Optom. Vis. Sci., 70, 528–531.
Mon-WilliamsM.TresilianJ. R.StrangN. C.KochharP.WannJ. P. (1998). Improving vision: neural compensation for optical defocus. Proc R. Soc. Lond. B, 265, 71–77.
WebsterM. A.GeorgesonM. A.WebsterS. M. (2002). Neural adjustments to image blur. Nat. Neurosci., 5, 839–840.
BurtP. J.AdelsonE. H. (1983). The Laplacian pyramid as a compact image code. IEEE Transactions on Communications, COM-31, 532–540.