Astrid A. Zeman

This paper compares the representational hierarchy of human and artificial visual systems, finding similarities in object-scene co-occurrence representations but differences in information processing strategies.

This study compares the ability of deep convolutional neural networks and humans to process Mooney face stimuli, finding that CNNs exhibit the face inversion effect and perform similarly to humans in some conditions.

This paper compares deep learning and human vision by training CNNs with increasing spatial frequency and finds improved classification of low-pass images.

This study investigates how animacy organization in occipitotemporal cortex relates to face-body division and taxonomic hierarchy, finding that animal resemblance to faces and bodies creates an apparent animacy continuum.

This study compares how deep CNNs and human visual cortex represent object shape and category, finding that CNNs and the brain encode these features separately, with CNNs peaking in final layers and the brain in anterior ventral temporal cortex.