Abstract
Texture gradients provide multiple potential cues to the 3D slant. A number of studies have demonstrated that texture scaling and compression contribute to slant perception. Textures with oriented structure provide an additional perspective convergence cue, which would be most informative when the oriented structure is aligned with the tilt direction. Some findings suggest that aligned spectral components are important for 3D perception from texture, but this has only been demonstrated in restricted situations where other texture cues are unreliable. In this study, we tested the contribution of oriented spectral components in conditions where texture and scaling are more informative. Observers viewed simulated planar patches of textured surfaces with varied slant (0°–80°) and field-of-view (16° and 6°) and performed an adjustment task to estimate slant. The simulated textures were octotropic plaids with a full range of orientations, or with either the aligned or perpendicular plaid components removed. Removing these plaid components would not degrade information from texture scaling and compression, so any differences could be attributed to the oriented components. We measured perceptual bias in monocular conditions and also the relative weighting of texture and stereo information in binocular conditions. For all textures, we observed more underestimation of slant and less weighting of texture information for surfaces with low slant, replicating previous findings. Comparing across textures, we found that aligned spectral components produced some improvement in slant estimates, but differences were small and only observed in some conditions, and there were no detectable effects on texture cue weights. These results demonstrate that aligned spectral components contribute to perception of slant from texture, but suggest that the contribution is limited when other texture cues are informative. Our findings are consistent with the notion that the visual system utilizes multiple texture cues for 3D perception.
Acknowledgement: Supported by the Hong Kong Research Grants Council, GRF 752412H