Comparison between the widely used adversarial training scheme in image restoration and the PNGAN framework. Image degradation and restoration are two important research fields in computational photography. However, too much work is dedicated to handling the image restoration problem and the image degradation remains under-studied. This is the most significant difference between the PNGAN and previous image restoration works. As depicted in Figure 1, we compare the common adversarial training scheme in image restoration and our PNGAN framework.

Existing deep learning real denoising methods require a large amount of noisy-clean image pairs for supervision. Nonetheless, capturing a real noisy-clean dataset is an unacceptable expensive and cumbersome procedure.

Existing deep learning real denoising methods require a large amount of noisy-clean image pairs for supervision. Nonetheless, capturing a real noisy-clean dataset is an unacceptable expensive and cumbersome procedure. To alleviate this problem, this work investigates how to generate realistic noisy images. Firstly, we formulate a simple yet reasonable noise model that treats each real noisy pixel as a random variable. This model splits the noisy image generation problem into two sub-problems: image domain alignment and noise domain alignment.