Abstract
Understanding internal process of ConvNets is commonly done using visualization techniques. However, these techniques do not usually provide a tool for estimating stability of a ConvNet against noise. In this paper, we show how to analyze a ConvNet in the frequency domain. Using the frequency domain analysis, we show the reason that a ConvNet might be sensitive to a very low magnitude additive noise. Our experiments on a few ConvNets trained on different datasets reveals that convolution kernels of a trained ConvNet usually pass most of the frequencies and they are not able to effectively eliminate the effect of high frequencies. They also show that a convolution kernel with more concentrated frequency response is more stable against noise. Finally, we illustrate that augmenting a dataset with noisy images can compress the frequency response of convolution kernels. Copyright © 2017 by SCITEPRESS - Science and Technology Publications, Lda. All rights reserved.
Originalsprache | englisch |
---|---|
Titel | 12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, VISIGRAPP 2017; Porto; Portugal |
Seiten | 362-369 |
ISBN (elektronisch) | 978-989758226-4 |
DOIs | |
Publikationsstatus | Veröffentlicht - 2017 |
Extern publiziert | Ja |
Veranstaltung | 12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications: VISIGRAPP 2017/VISAPP 2017 - Porto, Portugal Dauer: 27 Feb. 2017 → 1 März 2017 Konferenznummer: 12 http://www.grapp.visigrapp.org/?y=2017 |
Konferenz
Konferenz | 12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications |
---|---|
Kurztitel | VISIGRAPP |
Land/Gebiet | Portugal |
Ort | Porto |
Zeitraum | 27/02/17 → 1/03/17 |
Internetadresse |