A Super-Resolution Generative Adversarial Network with Simplified Gradient Penalty and Relativistic Discriminator

Sa, Haitao; Zou, Dafang; Mao, Jiafa; Sheng, Weiguo

doi:10.1109/ijcnn.2019.8852251

Cited by 2 publications

(1 citation statement)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The training process of standard GAN (StdGAN) is unstable because the architecture of the discriminative network is nontransformed and saturated [46,52]. RaGAN is an extension of StdGAN where the discriminative network takes a relativistic and non-saturating form [52][53][54]. Therefore, rather than measuring the probability that the input land cover map M is realistic and generated in StdGAN, RaGAN predicts whether a land cover map Mr from the training dataset is more realistic than a fake one, Mf (Mf=G(I)).…”

Section: Discriminative Network Dsgsmentioning

confidence: 99%

Spectral–Spatial Generative Adversarial Network for Super-Resolution Land Cover Mapping With Multispectral Remotely Sensed Imagery

Shang

Jiang

Ling

et al. 2023

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Super-resolution mapping (SRM) can effectively predict the spatial distribution of land cover classes within mixed pixels at a higher spatial resolution than the original remotely sensed imagery. The uncertainty of land cover fraction errors within mixed pixels is one of the most important factors affecting SRM accuracy. Studies have shown that SRM methods using deep learning techniques have significantly improved land cover mapping accuracy but have not coped well with spectral-spatial errors. This study proposes an end-to-end SRM model using a spectral-spatial generative adversarial network (SGS) with the direct input of multispectral remotely sensed imagery, which deals with spectral-spatial error. The proposed SGS comprises three parts: (1) Cube-based convolution for spectral unmixing is adopted to generate land cover fraction images. (2) A residual-inresidual dense block fully and jointly considers spectral and spatial information and reduces spectral errors. (3) A relativistic average GAN is designed as a backbone to further improve super-resolution performance and reduce spectral-spatial errors. SGS was tested in one synthetic and two realistic experiments with multi-/hyper-spectral remotely sensed imagery as the input, comparing the results with those of hard classification and several classic SRM methods. The results showed that SGS performed well at reducing land cover fraction errors, reconstructing spatial details, removing unpleasant and unrealistic land cover artifacts, and eliminating false recognition.

show abstract