Number is more than meets the eye: Unveiling segmentation mechanisms in numerosity perception with visual illusions

Area of Science:

• Visual perception
• Cognitive neuroscience
• Computational vision

Background:

• Humans and animals can estimate approximate numerosity, but the underlying visual mechanisms are unclear.
• Perceived numerosity is affected by object connectedness, with grouping cues influencing segmentation.
• Illusory contours (ICs) mimic real object perception, making them valuable for studying visual processing.

Purpose of the Study:

• To investigate if approximate numerosity perception of illusory contours (ICs) is affected by connectedness.
• To determine if visual mechanisms for processing real and illusory objects share similarities.

Main Methods:

• A psychophysical forced-choice task was used to compare numerosities of Ehrenstein-like IC ensembles.
• The number of connections between ICs was manipulated (0, 2, or 4) using task-irrelevant lines.
• Low-level visual features were controlled across conditions to isolate the effect of connectedness.

Main Results:

• Numerosity was monotonically underestimated as the number of connections increased.
• Perceptual precision remained constant, consistent with Weber-like encoding.
• Reaction times increased proportionally with connectedness, suggesting recurrent neural processing.

Conclusions:

• Illusory contour ensembles are subject to the same connectedness effect as real objects.
• This supports a shared visual mechanism for numerosity extraction across real and illusory stimuli.
• Findings offer insights into visual segmentation and inform models of artificial intelligence and perception.