Image Restoration with Generative Adversarial Networks (GANs)

Using Generative Adversarial Networks to restore image quality.

Image Restoration

Dataset generation

Crappified Images (Left), Original Images (Right) from Oxford Pets Dataset.

Pre-Training Generator and Critic

Image Restoration with Generative Adversarial Networks (GANs) 1
Results of pre-training generator network
Image Restoration with Generative Adversarial Networks (GANs) 2
Critic pre-training dataset with Originals (label as images) and Fakes (label as image_gen)

GAN training

Results

Image Restoration with Generative Adversarial Networks (GANs) 3
Image Restoration with Generative Adversarial Networks (GANs) 4

Conclusion

This article has been published from the source link without modifications to the text. Only the headline has been changed.

Source link