Two-dimensional aspect ratio discrimination for shape defined by orientation texture

Vision Res. 1996 Nov;36(22):3695-702. doi: 10.1016/0042-6989(96)00083-1.

Abstract

A pattern of 12,860 short (0.15 x 0.05 deg) texture lines contained an orientation texture-defined (OTD) rectangle of aspect ratio a/b (a was the rectangle's height and b was its width). All the lines within the rectangle had the same orientation and all the lines outside the rectangle had the same orientation. These two orientations were theta deg symmetrically clockwise and anticlockwise of the vertical. The rectangle's visibility was created entirely by the orientation difference 2 theta. Aspect ratio discrimination threshold for the texture-defined rectangle was a U-shaped function of theta that was approximately symmetrical about theta = 45 deg. The lowest values of aspect ratio discrimination threshold were 2.8% (SE = 0.1%), 2.7% (SE = 0.1%) and 5.1% (SE = 0.3%) for three observers. A luminance-defined (LD) rectangle with the same spatial sampling as the OTD rectangle was created by removing all texture lines outside the rectangle. Aspect ratio discrimination threshold for the LD rectangle was 1.1% (SE = 0.1%), 1.7% (SE = 0.1%) and 2.2% (SE = 0.1%)) for the same three observers. Although these discrimination thresholds were not greatly lower than discrimination thresholds for the OTD rectangle, they were significantly lower at the P < 0.001 level. Discrimination thresholds for the OTD rectangle are comparable with the lowest aspect ratio discrimination thresholds for motion-defined (MD) rectangles (2 and 3% for two observers), and for disparity-defined (DD) rectangles (3.1, 3.4, 4.0 and 7.4% for four observers), even though the MD and DD rectangles were much smaller than the 185 deg2 OTD rectangle used in the present study.

Publication types

  • Research Support, Non-U.S. Gov't
  • Research Support, U.S. Gov't, Non-P.H.S.

MeSH terms

  • Adult
  • Discrimination, Psychological*
  • Female
  • Form Perception*
  • Humans
  • Male
  • Pattern Recognition, Visual*
  • Regression Analysis
  • Sensory Thresholds