Purchase this article with an account.
Iris Groen, Sennay Ghebreab, Victor Lamme, Steven Scholte; The role of Weibull image statistics in rapid object detection in natural scenes. Journal of Vision 2010;10(7):992. doi: https://doi.org/10.1167/10.7.992.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The ability of the human brain to extract meaningful information from complex natural scenes in less amount of time than from simple, artificial stimuli is one of the great mysteries of vision. One prominent example of this ability is that natural scenes containing animals lead to a frontal ERP difference compared to scenes without animals as soon as 150 ms after stimulus-onset (Thorpe et al., 1996). Whereas these findings make clear that the brain is able to very rapidly distinguish these types of images, it is unclear on the basis of what information this distinction is made - in other words, whether early differences in the ERP between natural stimuli are related to low-level or high-level information in natural images. We have shown previously that the early animal vs. non-animal difference is driven by low-level image statistics of local contrast correlations, as captured by two parameters (beta and gamma) of the Weibull fit to the edge histogram of natural images (Scholte et al., 2009). These parameters can be estimated in a physiologically plausible way and explain 85% of the variance in the early ERP. We are currently expanding on this work by investigating to what extent low-level image statistics, as measured by beta and gamma, are involved in determining the latencies of target vs. non-target ERP differences in those cases where other types of stimuli than animals (vehicles) are used and where the specific task the subject is performing is varied (from simple detection to subordinate categorization). Early results confirm our previous findings and expand them to other types of stimuli and tasks.
Thorpe et al., (1996). Speed of processing in the human visual system. Nature, 381(6582):520-2.
Scholte et al., (2009). Visual gist of natural scenes derived from image statistics parameters [Abstract]. Journal of Vision, 9(8):1039, 1039a.
This PDF is available to Subscribers Only