Shape Constancy

Shape constancy

The percept of an object's 3D shape remains unchanged despite variations in the 2D retinal image due to changes in viewing direction.

Perceptual constancy

The perception of an object's geometric or physical characteristics remains stable even when the retinal image changes due to 3D viewing conditions.

Interpretation of 3D shapes

The visual system extracts non-accidental symmetrical interpretations to maintain shape constancy.

Shepard's Boxes

Even though the top surfaces of the boxes look identical in 2D, in 3D perception, they are interpreted differently.

Cortical pathways for visual processing

Temporal cortex processes shapes; Parietal cortex processes distances and sizes.

Mishkin, Ungerleider, and Macko (1983) findings

Object vision and spatial vision are handled by different cortical areas.

Mental rotation

The percept of a rotated shape is a mental transformation of the original shape.

Shepard & Metzler (1971) study

They demonstrated mental rotation using symmetrical 3D cube objects with right angles.

Mental rotation and perspective-taking

It allows us to understand what another person can or cannot see from their viewpoint.

Biederman (1987) proposal

Objects are recognized by their symmetrical parts (geons), which allow for qualitative object recognition.

Biederman & Gerhardstein (1993) key finding

Humans exhibit perfect shape constancy with 2D line drawings of 3D objects made of simple parts.

Conclusion of Biederman & Gerhardstein's study

Humans can perfectly perceive qualitative aspects of objects, such as part presence and shape.

Rock & DiVita (1987) study

Binocular viewing of 3D bent wire objects without symmetry.

Change in viewing orientation by 45°

Recognition dropped to chance level, indicating failure in forming a 3D mental representation.

Conclusion by Rock & DiVita

Humans do not form 3D visual representations of 3D shapes without symmetry.

Buelthoff & Edelman (1992) study

They repeated Rock & DiVita's study using VR and a large number of random objects but failed to use symmetrical ones.

Conclusion by Buelthoff & Edelman

Humans store 2D images of 3D objects rather than forming full 3D mental representations.

Edelman's (1995) approach

Not provided in the notes.

Animal models

He used animal models instead of unstructured objects.

Significance of Edelman's results

Shape constancy was reliable, showing the importance of symmetry in perception.

Performance of symmetrical shapes

Symmetrical shapes maintained better recognition even after rotation.

Lee & Saunders' experiment

Shape discrimination for sequentially presented objects at different rotation angles (0°, 15°, 30°, 60°).

Types of shape cues used

Shading only, stereo only, and combined shading & stereo.

Key result at 60° rotation

Symmetrical shapes performed twice as well as asymmetrical ones.

Li & Pizlo's main hypothesis

Shape perception is based on symmetry rather than depth perception.

Method used in Li & Pizlo's experiment

Subjects identified whether two sequentially presented 3D stimuli were the same or different, despite being shown from different angles.

Stimulus regularity manipulation

Objects ranged from symmetrical polyhedra to irregular polygonal lines.

Key results of Li & Pizlo's study

Asymmetrical stimuli led to chance performance; symmetrical polyhedra maintained reliable shape constancy; monocular and binocular performance were highly correlated.

Role of binocular viewing

It refined monocular 3D shape perception but was not reliable without it.

Additional factors in Experiment 2

Occluding contour, edges, shading, texture, and binocular viewing.

Main takeaway from Experiment 2

Shape constancy is influenced by these visual cues, but symmetry remains crucial.

Humans forming 3D mental representations

Yes, but only for certain objects, primarily those with symmetry.

Importance of symmetry for shape constancy

It provides redundancy that helps the visual system maintain a stable perception.

Role of ecological validity in shape perception research

Studies should use stimuli that reflect real-world objects, which are typically symmetrical and have distinct parts.

Significance of shape perception as a perceptual invariant

Shape constancy is a fundamental feature of human vision, ensuring consistent object recognition.

Impact of transformation from 3D to 2D on perception

The loss of structure in unstructured objects (e.g., wire objects) prevents reliable shape constancy.