Saliency metric for subadditive dissimilarity judgments of rectangles

Area of Science:

• Cognitive Psychology
• Psychophysics
• Visual Perception

Background:

• Previous research indicated strong subadditivity in rectangle dissimilarity judgments and the presence of individual differences.
• Four-parameter psychophysical maps were previously effective in reproducing average dissimilarity ratings.

Purpose of the Study:

• To replicate previous findings on rectangle dissimilarity judgments.
• To investigate the effect of bounded response transformation on psychophysical map fits.
• To identify shifts in subject response strategies based on stimulus pair types.

Main Methods:

• Replication of previous experimental findings on dissimilarity judgments for rectangles.
• Application of psychophysical maps with a bounded response transformation metric.
• Analysis of response times to differentiate subject strategies.

Main Results:

• Previous findings regarding subadditivity and individual differences in dissimilarity judgments were replicated.
• The bounded response transformation metric had a modest effect on the fits of psychophysical maps.
• A subadditive saliency metric was defined, weighting larger coordinates more heavily.
• Response times indicated a strategy shift: subjects focused on shape/size for rectangle-rectangle pairs and height/width for square-rectangle pairs.

Conclusions:

• The study confirms subadditivity and individual differences in visual perception of rectangles.
• A novel subadditive saliency metric was identified, reflecting differential weighting of stimulus dimensions.
• Distinct cognitive strategies are employed when judging dissimilarity between different types of rectangular stimuli.