Image Statistics and the Fine Lines of Material Perception

Area of Science:

• Visual perception
• Computational neuroscience
• Material appearance

Background:

• Perception of object properties like gloss is complex.
• Gloss perception theories often focus on image statistics, such as luminance variations.
• Existing models suggest gloss perception depends on image properties like sub-band skew.

Purpose of the Study:

• To investigate the role of image statistics versus specular edge structure in gloss perception.
• To test whether adaptation to image skew affects perceived gloss.
• To determine if adaptation to contours aligned with specular edges influences gloss perception.

Main Methods:

• Observers were adapted to stimuli with varying image skew (positive or negative).
• Observers were adapted to contours geometrically aligned with simulated specular edges.
• Aftereffects on perceived gloss were measured following adaptation.
• Adaptation effects were compared between aligned and rotated contours.

Main Results:

• Adaptation to positive or negative image skew decreased perceived gloss, contradicting statistical theories.
• Adaptation to contours aligned with specular edges also reduced perceived gloss.
• The gloss aftereffect was significantly stronger when contours matched specular edge alignment.

Conclusions:

• Gloss perception critically depends on the visual system's ability to encode specular edge structure.
• Image statistics, such as skew, are not the primary drivers of gloss perception.
• Findings support a model where specular edge encoding is fundamental to perceiving material gloss.