Year: 2006
Type: article
Abstract: In previous work, we modeled Mondloch's behavioral data [Mondloch et al., 2002] on adult face discrimination [Zhang and Cottrell, 2004].We found that our standard model [Dailey and Cottrell, 1999, Dailey et al., 2002] was overly holistic and that by adding local feature processing we could qualitatively match the human data.However, further investigation has lead us to reconsider our conclusions.In particular, we have found that the form of the Gabor filters we used were not biologically realistic, in that the spatial frequency range was overly limited.In this paper, we show that biologically realistic Gabor filters allow us to match the adult data using only the first two processing steps of our model.No neural network is necessary.In addition, we find that we can also model the developmental data qualitatively using a very simple manipulation of this preprocessing, i.e., how many people the model knows, and how many principal components it uses to represent the data.These two variables are sufficient to account for the slower development of configural processing compared to featural processing, and how the child's performance improves with age.
Authors: Cottrell, Garrison W., Zhang, Lingyun
Institution: University of California San Diego
Language: English
Subfield: Cognitive Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: green
Award: Modeling Face Perception
OpenAlex ID: W2626449169