Generative adversarial networks recover features in astrophysical images of galaxies beyond the deconvolution limit
Citations
1,301 citations
544 citations
Cites background from "Generative adversarial networks rec..."
..., for super-resolution in [30]* and deblurring in [14]....
[...]
...The same is true for deblurring, where training pairs can be generated by blurring [12]–[14]....
[...]
...[6]–[11] [10], [12]–[14] [9], [15]–[20] [21]–[23] [24]–[27]...
[...]
383 citations
275 citations
211 citations
References
3,869 citations
"Generative adversarial networks rec..." refers methods in this paper
...…achieve a PSNR of 37.2dB. Blind deconvolution (Bell & Sejnowski 1995) 2 achieves a PSNR of 19.9dB and Lucy- 2 We compare with blind deconvolution in Matlab. https://www. mathworks.com/help/images/ref/deconvblind.html Richardson deconvolution (Richardson 1972; Lucy 1974) 3 achieves a PSNR of 18.7dB....
[...]
1,827 citations
1,656 citations
"Generative adversarial networks rec..." refers background or methods in this paper
...1 It is known that if one uses Euclid distance for image recovery, this often produces blurred images because Euclid distance is uniform over the whole image (Reed et al. 2016), thus a more sophisticated loss function could improve the system....
[...]
...We train the GAN using open source code released by Reed et al. (2016) with TITAN X PASCAL GPUs....
[...]
...In this work, we adopted a standard GAN architecture; therefore we only briefly introduce GAN and interested readers can consult Reed et al. (2016) and Goodfellow et al. (2014) for details....
[...]
1,213 citations