Cite as:
Bècle P, Monot A, Viénot F, Eyquem D, 1998, "Detecting gradient and discriminating shape from texture" Perception 27 ECVP Abstract Supplement

Detecting gradient and discriminating shape from texture

P Bècle, A Monot, F Viénot, D Eyquem

Texture is a surface parameter. It characterises materials and can give information about the shape of objects. Surface textural variations can be recognised either as a local gradient or as a texture boundary (Wolfson and Landy, 1998 Vision Research 38 439 - 446). We have investigated the effect of two variables--the size and the contrast of the texture components--on the perception of texture. A visual stimulus generator was used to display 2-D textured images on a monitor. Eleven subjects were shown a test textured square (7.7 deg × 7.7 deg) against a reference textured background. The size of the texture components was chosen with reference to previously measured psychometric curves for detection threshold. Observers were asked to classify the difference between the two textures within three categories: not perceptible (weight 1), perceptible (weight 2), and highly perceptible (weight 3). Limits between categories differed significantly. This demonstrates the existence of two responses, one for the gradient detection and one for the edge detection. Average responses were plotted versus texture size ratio (reference/test). The curves exhibited an asymmetry: when the elements of the test were smaller than the elements of the background, gradient detection and boundary detection were more difficult to assess than in the opposite configuration. The former case suggests a filling-in phenomenon. Preliminary results show the same tendency for contrast. It has also been reported by the observers that the latency for detecting texture gradient was greater than for discriminating shape from texture.

These web-based abstracts are provided for ease of seaching and access, but certain aspects (such as as mathematics) may not appear in their optimum form. For the final published version of this abstract, please see
ECVP 1998 Abstract Supplement (complete) size: 2107 Kb