Date created: | Last Updated:

Category: Project

Description: In this study we investigated contextual modulation of naturalistic texture perception, namely, how perception of naturalistic texture in a given region is affected by its surrounds. For this we used a texture discrimination task involving target texture patches, and measured performance as different texture surrounds were included around the targets. We compare our results to the predictions from the feedforward summary-statistics (SS) representation model of peripheral vision with fixed pooling windows (standard SS model), as obtained from simulated observers with an architecture based on this model. We tested the effects of segmentation, surround statistics, and surround position on contextual modulation. Our results show that segmentation effects escape the standard SS model. Furthermore, our results on the effects of surround statistics, as well as on surround position allow us to relate our results to multiple lines of previous perceptual and physiological work, offering solid evidence to start extending the standard SS model. All data and code used in the project is available in this repository.