Now showing items 1-2 of 2

  • .

    .

    .

    Deep convolutional networks do not classify based on global object shape 

    Baker, Nicholas; Lu, Hongjing; Erlikhman, Gennady; Kellman, Philip J. (2018)
    Deep convolutional networks (DCNNs) are achieving previously unseen performance in object classification, raising questions about whether DCNNs operate similarly to human vision. In biological vision, shape is arguably the ...
  • .

    .

    .

    On the Legibility of Mirror-Reflected and Rotated Text 

    Erlikhman, Gennady; Strother, Lars; Barzakov, Iskra; Caplovitz, Gideon Paul (2017)
    We happened to observe that text that was reflected about either the horizontal or vertical axis was more difficult to read than text that was reflected about first one and then the other, which amounts to a 180-degree ...