Home | Research | CV | Published Work | Research Facilities | Teaching



Visual Perception of Shape from Texture

Human beings perceive depth and shape in projections of 3D scenes, such as photographs and pictures. This demonstrates how effortlessly the human visual system extracts 3D shape from 2D images. The question arises, which image information the human visual system uses to accomplish this feat.

2D projections of 3D surfaces provide multiple sources of information. For example when a textured surface is projected onto an image plane, the texture pattern in the image is systematically distorted. Because the texture distortion in the image and the 3D structure of the underlying scene are related by means of projective geometry, one can exploit variations in 2D image texture to determine 3D structure of the underlying scene.


Patterns of optical texture. Variations in size, density and shape of texture elements give the impression of changes in 3D surface structure.


Computational models of shape from texture face two main challenges. First, surfaces can belong to either of three surface types (planar, singly and doubly curved ). Second, textures come in a huge variety.


(a) Picture of several geometric primitives that can be classified into planar, singly and doubly curved surfaces. Planes have no curvature. Singly curved surfaces, such as cones and cylinders, can be obtained by bending a sheet of paper, without tearing or folding it. In contrast, doubly curved surfaces, such as spheres, cannot be obtained from a sheet of paper, without tearing or folding. (b) A picture of several surfaces and textures that we encounter on an everyday basis.


We proposed a computational model to explain people’s perceptual reports about 3D shape from texture (Thaler, L. & Todd, J.T. (2010). The perception of 3D shape from texture based on directional width gradients. Journal of Vision, 10(5), 1-13. [link to free article]).

In future work I plan to test the validity of this model for goal-directed hand movements.

Thaler, L. , Chaudry, R.& Nikolov, M.B. Texture Width Gradients and Binocular Disparity are not integrated for shape perception with Real 3D Objects.

Thaler, L. & Todd, J.T. (2010). The perception of 3D shape from texture based on directional width gradients. Journal of Vision, 10(5), 1-13. [link to free article]

Todd, J.T., Thaler, L., Dijkstra, T.M.H, Koenderink, J.J. & Kappers, A.M.L. (2007). The Effects of Viewing Angle, Camera Angle and Sign of Surface Curvature on the Perception of 3D Shape from Texture. Journal of Vision, 7(12), 1-16. [link to free article]

Thaler, L., Todd, J.T. & Dijkstra, T.M.H. (2007). The Effects of Phase on the Perception of 3D Shape from Texture: Psychophysics and Modeling. Vision Research, 47, 411-427. [download as pdf]

Todd, J.T., Thaler, L. & Dijkstra, T.M.H (2005). The Effects of Field of View on the Perception of 3D Slant from Texture. Vision Research, 45, 1501-1517. [download as pdf]