Purchase this article with an account.
Paul B. Hibbard, Samira Bouzit; Binocular energy responses to natural images. Journal of Vision 2006;6(6):833. doi: https://doi.org/10.1167/6.6.833.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The responses of many binocular cells in the primary visual cortex are well described by the binocular energy model (Fleet et al., 1996). If we are to understand how these cells encode and interpret the binocular information in natural images, it is important to consider how they respond to such images. We therefore analysed the responses of the binocular energy model to a collection of natural, binocular images, that were taken with a pair of digital cameras with an inter-camera distance of 6.5 cm. Images were either indoor “still-lifes” or outdoor scenes. The responses of model binocular energy units to these stimuli were computed. Responses were calculated over a range of spatial frequencies and orientations, for units that encoded binocular disparity via positional shifts in their receptive fields. The results show (i) a strong peak in responses for disparities around zero for locations close to fixation; (ii) a rapid broadening of the response distribution with increasing eccentricity; (iii) responses that are much more tightly tuned around zero for vertical than for horizontal disparity and (iv) marked differences in responses to indoor and outdoor scenes. These results are consistent with physiological and psychophysical studies of disparity encoding (Prince et al, 2002; Cumming, 2002; Ogle, 1950). They also show that a coarse estimate of distance may be obtained from the responses of a population of binocular neurons, without needing to scale disparities, or even to solve the correspondence problem.
This PDF is available to Subscribers Only