English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Deep convolutional neural networks (DCNNs) have attracted much attentionrecently, and have shown to be able to recognize thousands of object categoriesin natural image databases. Their architecture is somewhat similar to that ofthe human visual system: both use restricted receptive fields, and a hierarchyof layers which progressively extract more and more abstracted features. Yet itis unknown whether DCNNs match human performance at the task of view-invariantobject recognition, whether they make similar errors and use similarrepresentations for this task, and whether the answers depend on the magnitudeof the viewpoint variations. To investigate these issues, we benchmarked eightstate-of-the-art DCNNs, the HMAX model, and a baseline shallow model andcompared their results to those of humans with backward masking. Unlike in allprevious DCNN studies, we carefully controlled the magnitude of the viewpointvariations to demonstrate that shallow nets can outperform deep nets and humanswhen variations are weak. When facing larger variations, however, more layerswere needed to match human performance and error distributions, and to haverepresentations that are consistent with human behavior. A very deep net with18 layers even outperformed humans at the highest variation level, using themost human-like representations.

Deep Networks Can Resemble Human Feed-forward Vision in Invariant Object Recognition