Local masking in natural images: a database and analysis

Area of Science:

• Visual perception
• Computational neuroscience
• Image processing

Background:

• Visual masking studies offer insights into visual coding.
• Few studies use natural scenes as masks, limiting understanding of their effects on detection thresholds.

Purpose of the Study:

• To investigate how natural scene properties influence local contrast detection thresholds.
• To compare results from natural masks with those from unnatural masks.
• To develop a ground-truth dataset for human visual system research.

Main Methods:

• A psychophysical study measured contrast detection thresholds for log-Gabor noise targets in natural image patches.
• A three-alternative forced-choice experiment was used with 30 natural images from the CSIQ database.
• Masking maps were created, detailing detection thresholds at specific spatial locations.

Main Results:

• Detection thresholds were influenced by patch properties like visual complexity, texture fineness, sharpness, and luminance.
• Low-level mask features (except sharpness) showed weak correlation with detection thresholds (r ≤ 0.52).
• A computational contrast gain control model predicted thresholds well (average r = 0.79).

Conclusions:

• Natural scene complexity, beyond simple features, significantly affects visual detection.
• Computational models can effectively predict human visual performance in naturalistic conditions.
• The created dataset facilitates further research into the human visual system with natural masks.