Optimization Of The Efficient Sub-Pixel Convolutional Network Model For Satellite Image Super-Resolution: Study Of Epoch And Batch Size Hyperparameters

Harimino Andriamalala RAJAONARISOA, Adolphe Andriamanga RATIARISON

Abstract


This study applied the ESPCN model to the super-resolution of geostationary meteorological satellite images. Using 100 pairs of low- and high-resolution images, the model was trained and tested the by optimizing the epoch and batch-size hyperparameters. The analysis of the PSNR between the reconstructed images and the target images made it possible to identify the optimal values that ensure the best reconstruction quality. The results demonstrated the importance of hyperparameter tuning to improve model performance in the context of satellite image processing.

Keywords


Batch size, Deep learning, Epoch, ESPCN, Image super-resolution, PSNR.

Full Text:

PDF

References


Q. Ha, M. Dang, T. D. Hoang, and Y. Kim, “Image super-resolution using deep learning: A comprehensive survey,” International Journal of Automation and Computing, vol. 16, no. 6, pp. 663–685, 2019, doi: 10.1007/s11633-019-1183-x

Z. Wang, Z. Zeng, Y. Zhang, L. Wang, and L. Zhang, “A comprehensive review of super-resolution in remote sensing: From classical to deep learning era,” Earth-Science Reviews, vol. 232, p. 104110, 2022, doi: 10.1016/j.earscirev.2022.104110

X. Lu, X. Zhang, Y. Liu, and J. Wu, “Remote sensing image super-resolution via a dual-path network,” Remote Sensing, vol. 11, no. 13, p. 1588, 2019, doi: 10.3390/rs11131588

W. Shi, J. Caballero, F. Huszár, J. Totz, A. P. Aitken, R. Bishop, et al., “Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2016, pp. 1874–1883, doi: 10.1109/CVPR.2016.207.

Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, “Image quality assessment: From error visibility to structural similarity,” IEEE Trans. Image Process., vol. 13, no. 4, pp. 600–612, 2004, doi: 10.1109/TIP.2003.819861

Goodfellow, Y. Bengio, and A. Courville, Deep Learning. Cambridge, MA, USA: MIT Press, 2016

C. Dong, C. C. Loy, K. He, and X. Tang, “Image super-resolution using deep convolutional networks,” IEEE Trans. Pattern Anal. Mach. Intell., 2016.

N. Ahn, B. Kang, and K. Sohn, “Fast, accurate, and lightweight super-resolution with cascading residual network,” in Proc. European Conf. Comput. Vis. (ECCV), 2018.

S. L. Smith and Q. V. Le, “A Bayesian perspective on generalization and stochastic gradient descent,” in Proc. Int. Conf. Learn. Representations (ICLR), 2018.

B. Lim, S. Son, H. Kim, S. Nah, and K. M. Lee, “Enhanced deep residual networks for single image super-resolution,” in Proc. CVPR Workshops, 2017.

N. Ahn, B. Kang, and K. Sohn, “Fast, accurate, and lightweight super-resolution with cascading residual network,” in Proc. European Conf. Comput. Vis. (ECCV), 2018.

S. L. Smith and Q. V. Le, “A Bayesian perspective on generalization and stochastic gradient descent,” in Proc. Int. Conf. Learn. Representations (ICLR), 2018.

W. Shi, J. Caballero, F. Huszár, J. Totz, A. P. Aitken, R. Bishop, et al., “Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2016, pp. 1874–1883, doi: 10.1109/CVPR.2016.207.

C. Dong, C. C. Loy, K. He, and X. Tang, “Image super-resolution using deep convolutional networks,” IEEE Trans. Pattern Anal. Mach. Intell., 2016.

Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, “Image quality assessment: From error visibility to structural similarity,” IEEE Trans. Image Process., vol. 13, no. 4, pp. 600–612, 2004, doi: 10.1109/TIP.2003.819861.




DOI: http://dx.doi.org/10.52155/ijpsat.v55.1.7763

Refbacks

  • There are currently no refbacks.


Copyright (c) 2025 Harimino Andriamalala RAJAONARISOA

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.