Semantic Segmentation of Satellite Images Based on an Improved U-Net Model with Dual Inputs 


Vol. 48,  No. 10, pp. 1219-1222, Oct.  2023
10.7840/kics.2023.48.10.1219


PDF
  Abstract

This paper proposes a new deep learning-based model for performing semantic segmentation using high-resolution satellite images. Because the feature information obtained from the encoder is limited in existing deep learning-based semantic segmentation techniques, they have a problem in that this information is transmitted to the decoder and accuracy is impaired. This limited feature extraction and degradation in accuracy causes inefficiency in prediction, which leads to inaccurate results. To overcome these limitations, this work proposes an encoder-decoder structure that is improved over existing architectures by designing inputs of various sizes to be processed at the same time. This effectively extracts a variety of rich feature information, enables more efficient delivery to decoders, and improves the accuracy of semantic segmentation.

  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.


  Related Articles
  Cite this article

[IEEE Style]

Y. Lee, J. Jung, Y. Shin, "Semantic Segmentation of Satellite Images Based on an Improved U-Net Model with Dual Inputs," The Journal of Korean Institute of Communications and Information Sciences, vol. 48, no. 10, pp. 1219-1222, 2023. DOI: 10.7840/kics.2023.48.10.1219.

[ACM Style]

Yelim Lee, Jin-won Jung, and Yoan Shin. 2023. Semantic Segmentation of Satellite Images Based on an Improved U-Net Model with Dual Inputs. The Journal of Korean Institute of Communications and Information Sciences, 48, 10, (2023), 1219-1222. DOI: 10.7840/kics.2023.48.10.1219.

[KICS Style]

Yelim Lee, Jin-won Jung, Yoan Shin, "Semantic Segmentation of Satellite Images Based on an Improved U-Net Model with Dual Inputs," The Journal of Korean Institute of Communications and Information Sciences, vol. 48, no. 10, pp. 1219-1222, 10. 2023. (https://doi.org/10.7840/kics.2023.48.10.1219)
Vol. 48, No. 10 Index

Search
(IN TITLE, AUTHOR, ABSTRACT,KEYWORDS)

Advanced Search ▼

POPULAR KEYWORDS
(TOP 10 KEYWORDS)

Recent Publications
(LAST 3 YEARS)






pISSN : 1226 - 4717
eISSN : 2287 - 3880