Affiliation:
1. Institute of Psychology, University of Tartu, 50409 Tartu, Estonia endel.poder@ut.ee
Abstract
Abstract
Deep convolutional neural networks (CNN) follow roughly the architecture of biological visual systems and have shown a performance comparable to human observers in object classification tasks. In this study, three deep neural networks pretrained for image classification were tested in visual search for simple features and for feature configurations. The results reveal a qualitative difference from human performance. It appears that there is no clear difference between searches for simple features that pop out in experiments with humans and for feature configurations that exhibit strict capacity limitations in human vision. Both types of stimuli reveal comparable capacity limitations in the neural networks tested here.
Subject
Cognitive Neuroscience,Arts and Humanities (miscellaneous)
Reference44 articles.
1. Perceptual integration and post-categorical filtering;Allport,1985
2. Parallel versus serial processing in rapid pattern discrimination;Bergen;Nature,1983
3. Perception and communication.
4. Does attention have different effects on line orientation and line arrangement discrimination?;Cheal;Quarterly Journal of Experimental Psychology,1991
5. Multi-column deep neural networks for image classification;Ciresan,2012
Cited by
2 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献