Purchase this article with an account.
Gregory Daniel, Tribhawan Kumar, Donald A. Glaser; The matching of vertical lines in the presence of stereoscopic interpolation. Journal of Vision 2003;3(12):73. doi: https://doi.org/10.1167/3.12.73.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
It has been shown in a stereogram consisting of several horizontal rows of regularly spaced points that the unambiguous outside points are matched first. Subsequently, the intervening points are matched as if a plane had been interpolated between the outside points, and this potentially ambiguous matching is ‘unstable’ . Thus, the perceived depth of a bar is dependent not only on the disparity, but also on the ‘texture’ formed by the number of bars and separation of bars in the ambiguous region. We investigated this phenomenon using a stereogram made of a horizontal row of equally spaced thin vertical lines, with the outermost lines on each side both having either a positive or negative disparity. The number of lines, the spacing between the lines, the stimulus duration, and the onset time of the outermost lines were varied separately during the experiment. Our results show the size of the interpolation effect is dependent on the number of bars being shown, and the results agree with those reported by the earlier researchers showing a weak effect for widely spaced lines. On average, the observers showed little effect for very brief and very long presentations, with the maximum effect occurring at durations of about 1 second. In addition, the effect is weaker when the outermost lines appear 1 second after the onset of the intervening lines than when all the lines are shown simultaneously. We interpret these results to indicate that ‘matching’ of the features does not proceed simultaneously across the visual field, but may be influenced initially by the unambiguous features of the image and the fixation point.
MitchisonG. J.McKeeS. P.(1987). The resolution of ambiguous stereoscopic matches by interpolation. Vision Research, 27, 285–294.
This PDF is available to Subscribers Only