Visual discrimination of optical material properties: A large-scale study.

Area of Science:

• Visual Perception
• Material Science
• Computer Graphics
• Psychophysics

Background:

• Understanding complex visual processing of object materials requires diverse research approaches.
• Sharing stimulus and response data is crucial for comparing results across studies and facilitating interdisciplinary research.

Purpose of the Study:

• To construct a comprehensive database of material images with annotated visual discrimination performance data.
• To enable direct comparison of results from different studies on material perception.
• To provide a resource for researchers from various backgrounds to enter the field of material perception.

Main Methods:

• Material images were generated using physically based computer graphics.
• Psychophysical experiments were conducted in both laboratory and crowdsourcing settings.
• A nonverbal oddity task was used for material discrimination across dimensions like gloss, translucency, and material type (metal, glass, plastic).

Main Results:

• Material discrimination performance was influenced by variations in illumination and object geometry.
• Discriminating the spatial consistency of specular highlights in gloss perception exhibited significant individual differences.
• Higher-order color texture statistics partially explained task performance, but not completely.

Conclusions:

• The developed database is a valuable resource for studying visual material perception.
• Crowdsourced results strongly correlated with laboratory findings, indicating the database's utility even with less controlled experimental conditions.
• The nonverbal task design supports broad applicability across diverse populations and research areas.