English
 
User Manual Privacy Policy Disclaimer Contact us
  Advanced SearchBrowse

Item

ITEM ACTIONSEXPORT
  Of Human Observers and Deep Neural Networks: A Detailed Psychophysical Comparison

Geirhos, R., Jannsen, D., Schütt, H., Bethge, M., & Wichmann, F. (2017). Of Human Observers and Deep Neural Networks: A Detailed Psychophysical Comparison. Poster presented at 17th Annual Meeting of the Vision Sciences Society (VSS 2017), St. Pete Beach, FL, USA.

Item is

Basic

show hide
Item Permalink: http://hdl.handle.net/21.11116/0000-0000-C445-5 Version Permalink: http://hdl.handle.net/21.11116/0000-0006-B4B6-1
Genre: Poster

Files

show Files

Locators

show
hide
Locator:
Link (Any fulltext)
Description:
-

Creators

show
hide
 Creators:
Geirhos, R, Author
Jannsen, D, Author
Schütt, H, Author
Bethge, M1, Author              
Wichmann, F, Author              
Affiliations:
1External Organizations, ou_persistent22              

Content

show
hide
Free keywords: -
 Abstract: Deep Neural Networks (DNNs) have recently been put forward as computational models for feedforward processing in the human and monkey ventral streams. Not only do they achieve human-level performance in image classification tasks, recent studies also found striking similarities between DNNs and ventral stream processing systems in terms of the learned representations (e.g. Cadieu et al., 2014, PLOS Comput. Biol.) or the spatial and temporal stages of processing (Cichy et al., 2016, arXiv). In order to obtain a more precise understanding of the similarities and differences between current DNNs and the human visual system, we here investigate how classification accuracies depend on image properties such as colour, contrast, the amount of additive visual noise, as well as on image distortions resulting from the Eidolon Factory. We report results from a series of image classification (object recognition) experiments on both human observers and three DNNs (AlexNet, VGG-16, GoogLeNet). We used experimental conditions favouring single-fixation, purely feedforward processing in human observers (short presentation time of t = 200 ms followed by a high contrast mask); additionally, we used exactly the same images from 16 basic level categories for human observers and DNNs. Under non-manipulated conditions we find that DNNs indeed outperformed human observers (96.2 correct versus 88.5; colour, full-contrast, noise-free images). However, human observers clearly outperformed DNNs for all of the image degrading manipulations: most strikingly, DNN performance severely breaks down with even small quantities of visual random noise. Our findings reinforce how robust the human visual system is against various image degradations, and indicate that there may still be marked differences in the way the human visual system and the three tested DNNs process visual information. We discuss which differences between known properties of the early and higher visual system and DNNs may be responsible for the behavioural discrepancies we find.

Details

show
hide
Language(s):
 Dates: 2017-08
 Publication Status: Published in print
 Pages: -
 Publishing info: -
 Table of Contents: -
 Rev. Type: -
 Identifiers: DOI: 10.1167/17.10.806
BibTex Citekey: GeirhosJSBW2017
 Degree: -

Event

show
hide
Title: 17th Annual Meeting of the Vision Sciences Society (VSS 2017)
Place of Event: St. Pete Beach, FL, USA
Start-/End Date: 2017-05-19 - 2017-05-24

Legal Case

show

Project information

show

Source 1

show
hide
Title: Journal of Vision
Source Genre: Journal
 Creator(s):
Affiliations:
Publ. Info: Charlottesville, VA : Scholar One, Inc.
Pages: - Volume / Issue: 17 (10) Sequence Number: - Start / End Page: 806 Identifier: ISSN: 1534-7362
CoNE: https://pure.mpg.de/cone/journals/resource/111061245811050