Seeing the tipping point: Balance perception and visual shape.

Area of Science:

• Cognitive Psychology
• Computational Vision
• Perception Science

Background:

• Humans can quickly perceive functional properties of objects, like grasp points and center of mass.
• Perception of balance is often biased, leading to misjudgments of an object's center of mass.
• These biases may indicate fundamental aspects of how objects are represented in the mind.

Purpose of the Study:

• To investigate systematic biases in center of mass estimation for two-dimensional (2D) shapes.
• To propose and test the "shape inflation" hypothesis as an explanation for these biases.
• To explore how manipulating shape representation affects balance perception.

Main Methods:

• Systematic center of mass estimation tasks for 2D shapes (Study 1).
• Modeling biases using a computer graphics "shape inflation" technique.
• Experimental manipulations to reduce the need for inflated shape representations (Studies 2 & 3).

Main Results:

• Biased center of mass estimation was consistently observed for 2D shapes.
• A model based on "shape inflation" accurately predicted the observed estimation biases.
• Reducing cues to shapehood and depicting explicit depth diminished the biases.

Conclusions:

• The human mind implicitly represents 2D shapes as inflated, volumetric 3D objects.
• This "shape inflation" intrudes on and biases the perception of geometric properties like center of mass.
• Biases in balance perception reflect a general functional characteristic of object representation in the mind.