Texture segmentation: do the processing units on the saliency map increase with eccentricity?

Area of Science:

• Computational neuroscience
• Visual perception
• Human-computer interaction

Background:

• Saliency maps simulate human visual attention and target detection.
• Previous models did not fully account for the spatial structure of saliency processing.

Purpose of the Study:

• Investigate the spatial organization of saliency maps.
• Determine how processing unit size relates to retinal eccentricity.
• Understand the factors influencing interference between saliency signals.

Main Methods:

• Two experiments systematically varied the distance between a target and a masking element.
• Texture segmentation tasks were used to assess saliency signal interference.
• Analysis focused on critical spatial distances and their relation to target eccentricity.

Main Results:

• Saliency signals interfere when texture irregularities are within a critical spatial distance.
• This critical distance increases significantly with target retinal eccentricity.
• Eccentricity-dependent critical distances suggest crowding effects in visual processing.

Conclusions:

• Saliency maps are structured into processing units with size increasing with eccentricity.
• Crowding effects, influenced by target eccentricity and saliency signal strength, impair visual processing.
• Findings advance computational models of human visual attention.