Surface Attitude Judgements in monocular and stereo textures: a method evaluation

People can judge the attitude of surfaces in the both ecological valid 3D environments and in 2D representations with no binocular disparity. Many methods have been proposed for measuring perceived surface attitude but the reliability and bias vulnerability of such methods is seldom compared. One concern is that test methods which use the same visual modality as the surfaces under test may introduce bias. We compared 2D, 3D gauge figures which use perspective only and perspective plus disparity cues respectively, and a perspective and disparity free dial method wherein observers set the orientation of two lines to represent perceived tilt and slant. Observers judged tilt and slant for artificial sloped surfaces patterned with randomly orientated squares providing strong perspective cues. Surfaces were presented dichoptically with perspective only (2D) and perspective plus disparity (3D), and monocularly with perspective only. All response methods showed reliable and unbiased responses in tilt judgements. For slant, bias varied between conditions. The dial method produced estimates that correlated best with ground truth and produced the least bias overall, although it tended to underestimate small slants and overestimate large slants. The gauge figure methods were similar and tended to overestimate all slants but were worse for small slants. These results suggest that our textures provide completing cues to slant that resulted in bias when tested within the same visual modality. Thus the dial method introduced least bias. This method is difficult, and introduces individual differences but practice might reduce this making it preferable to gauge figure methods. [This project was funded by the Engineering and Physical Sciences Research Council; grant number EP/S016260/1.]