Harimino Andriamalala RAJAONARISOA1,*, Solonjaka RAKOTONASY2, Fanilo RANDRIAMAHALEO3, Adolphe Andriamanga RATIARISON4
Harimino Andriamalala RAJAONARISOA
1 Dynamic laboratory of Atmosphere, Climate, and Ocean, Physics and Applications, University of Antananarivo, Madagascar
Email: [email protected]
2 Mathematics and Computer Science Mention, Faculty of Science, University of Antananarivo, Madagascar
Email: [email protected]
3 Mathematics and Computer Science Mention, Faculty of Science, University of Antananarivo, Madagascar
Email: [email protected]
4 Dynamic laboratory of Atmosphere, Climate, and Ocean, Physics and Applications, Sciences and Technologies, University of Antananarivo, Madagascar
Email: [email protected]
The purpose of this study was to determine the "batch size" and "epoch" parameters of the SRGAN (Super Resolution Generative Adversarial Network) Model that best improve the quality of images from the Second Generation Geostationary Meteorological Satellites. The datasets used for training the model consisted of two hundred images from this satellite, including one hundred low-resolution images and one hundred corresponding high-resolution images. The images were captured by the same meteorological satellite at the same moments but in two different resolutions (low and high resolution). 80 pairs of images were used for model training, and the remaining 20 pairs were used for testing. For each combination of batch size and epoch parameters, the trained model generated an image, which was then compared to the expected high-resolution image using the PSNR metric. The PSNR evolution curves as a function of these parameters taken independently (the other parameters of the SRGAN model were set to default) helped find the optimal values for batch size and epoch.
Artificial Intelligence, Batch size, Epoch, PSNR, SRGAN, Upscaling
Harimino Andriamalala RAJAONARISOA, Solonjaka RAKOTONASY, Fanilo RANDRIAMAHALEO, Adolphe Andriamanga RATIARISON (2023). Super Resolution Generative Adversarial Network Model parameter optimization to improve the quality of geostationary meteorological satellite images. Journal of Artificial Intelligence and Systems, 5, 115–124. https://doi.org/10.33969/AIS.2023050108.
[1] Renaut, D. (2004). Les satellites météorologiques. La meteorologie, 2004(45), 33-37.
[2] Capderou, M. (2017). Les orbites des satellites météorologiques. La Météorologie, 2017(97), 94-101.
[3] 52è Conseil d’EUMETSAT, Résolution du Conseil EUM/C/03/Rés. I du Juin 2004. https://www-cdn.eumetsat.int/files/2020-04/pdf_leg_res_2003_fr.pdf
[4] B. Lim, S. Son, H. Kim, S. Nah and K. M. Lee, "Enhanced Deep Residual Networks for Single Image Super-Resolution," 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)
[5] Honolulu, HI, USA, 2017, pp. 1132-1140, doi: 10.1109/CVPRW.2017.151.
[6] Wei-Sheng Lai, Jia-Bin Huang, Narendra Ahuja, and Ming-Hsuan Yang, "Deep Laplacian Pyramid Networks for Fast and Accurate Super-Resolution", in IEEE Conference on Computer Vision and Pattern Recognition, 2017.
[7] Ledig C, Theis L, Huszár F, et al. Photo-realisticsingle image super-resolution using a generativeadversarial network[C]//Proceedings of the IEEEconference on computer vision and patternrecognition. 2017: 4681-4690.
[8] EUMETSAT, Meteosat 0 degree Natural Colour Full disc low resolution. https://eumetview.eumetsat.int/static-images/MSG/RGB/NATURALCOLOR/FULLDISC/
[9] EUMETSAT, Meteosat 0 degree Natural Colour Full disc high resolution https://eumetview.eumetsat.int/static-images/MSG/RGB/NATURALCOLOR/FULLRESOLUTION/
[10] EUMETSAT, Monitoring weather and climate from space.
https://eumetview.eumetsat.int/static-images/MSG/RGB/NATURALCOLOR/FULLDISC/
[11] Singh, M. K., Baluja, G. S., & Sahu, D. P. Understanding the Convolutional Neural Network &it’s Research Aspects in Deep Learning. Vol, 5, 867-871.
[12] Krizhevsky, A., Sutskever, I. and Hinton, G.E. (2017) Imagenet Classification with Deep Convolutional Neural Networks. Communications of the ACM, 60, 84-90. https://doi.org/10.1145/3065386
[13] (K. He, X. Zhang, S. Ren and J. Sun, "Deep Residual Learning for Image Recognition," 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
[14] Las Vegas, NV, USA, 2016, pp. 770-778, doi: 10.1109/CVPR.2016.90.
[15] Ledig C, Theis L, Huszár F, et al. Photo-realisticsingle image super-resolution using a generativeadversarial network[C]//Proceedings of the IEEEconference on computer vision and patternrecognition. 2017: 4681-4690.
[16] A. M. Eskicioglu and P. S. Fisher, "Image quality measures and their performance," in IEEE Transactions on Communications, vol. 43, no. 12, pp. 2959-2965, Dec. 1995, doi: 10.1109/26.477498.
[17] Palubinskas, G. (2017). Image similarity/distance measures: what is really behind MSE and SSIM?. International Journal of Image and Data Fusion, 8(1), 32-53.
[18] Hore, A., & Ziou, D. (2010, August). Image quality metrics: PSNR vs. SSIM. In 2010 20th international conference on pattern recognition (pp. 2366-2369). IEEE.