Differentiating global and local contour completion using a dot localization paradigm

Area of Science:

• Cognitive Psychology
• Computational Neuroscience
• Visual Perception

Background:

• Amodal completion theories for occluded objects focus on local contours or global factors like symmetry.
• These theories may represent separate low-level contour interpolation and higher-order global recognition processes.
• Distinguishing these processes requires precise boundary representations from only one.

Purpose of the Study:

• To experimentally differentiate between local contour interpolation and global symmetry-based processes in amodal completion.
• To assess the precision and accuracy of perceived object boundaries under different completion interpretations.
• To determine if local interpolation yields more precise representations than global completion.

Main Methods:

• A dot localization paradigm was employed to measure perceived object boundary accuracy and precision.
• Participants completed partially occluded objects, guided by either local or global interpretations.
• Psychometric functions were estimated using interleaved 2-up, 1-down staircases to find .707 probability points.

Main Results:

• Local contour interpolation resulted in precise and accurate representations of occluded object boundaries.
• These local process findings were consistent across different observers.
• Completion based on global symmetry did not yield similarly precise or accurate boundary representations.

Conclusions:

• The findings support a distinction between automatic, local contour interpolation and recognition-based global processes.
• Local processes generate more precise representations of occluded object boundaries than global ones.
• This suggests separate neural mechanisms underlie local and global aspects of amodal completion.