Impacts of Image Resizing on the Performance of Deep Neural Network-Based Image Classifiers 


Vol. 44,  No. 7, pp. 1299-1302, Jul.  2019
10.7840/kics.2019.44.7.1299


PDF
  Abstract

In this letter, we investigate the impact of image resizing on the performance of deep neural network-based image classifiers. Since most deep neural network-based image classifiers require a fixed-size input dimension, an input image needs to be resized before testing. In this letter, we use five image resizing operators to empirically investigate the impact of each resizing operator on the performance of the image classifiers. For quantitative evaluation, we report Top-5 and Top-1 accuracies of five image classifiers trained by the ImageNet dataset. We believe that this study serves as a practically useful benchmark for researchers and practitioners interested in utilizing deep neural network-based image classifiers.

  Statistics
Cumulative Counts from November, 2022
Multiple requests among the same browser session are counted as one view. If you mouse over a chart, the values of data points will be shown.


  Cite this article

[IEEE Style]

Y. Kim, C. Jung, C. Kim, "Impacts of Image Resizing on the Performance of Deep Neural Network-Based Image Classifiers," The Journal of Korean Institute of Communications and Information Sciences, vol. 44, no. 7, pp. 1299-1302, 2019. DOI: 10.7840/kics.2019.44.7.1299.

[ACM Style]

Yoonhyung Kim, Chanho Jung, and Changick Kim. 2019. Impacts of Image Resizing on the Performance of Deep Neural Network-Based Image Classifiers. The Journal of Korean Institute of Communications and Information Sciences, 44, 7, (2019), 1299-1302. DOI: 10.7840/kics.2019.44.7.1299.

[KICS Style]

Yoonhyung Kim, Chanho Jung, Changick Kim, "Impacts of Image Resizing on the Performance of Deep Neural Network-Based Image Classifiers," The Journal of Korean Institute of Communications and Information Sciences, vol. 44, no. 7, pp. 1299-1302, 7. 2019. (https://doi.org/10.7840/kics.2019.44.7.1299)

Search
(IN TITLE, AUTHOR, ABSTRACT,KEYWORDS)

Advanced Search ▼

POPULAR KEYWORDS
(TOP 10 KEYWORDS)

Recent Publications
(LAST 3 YEARS)






pISSN : 1226 - 4717
eISSN : 2287 - 3880