Although image denoising methods based on a convolutional neural network (CNN) have achieved the state-of-the-art in Gaussian noise reduction, their performances are still very limited on real photographs even beyond the reach of traditional methods, such as block-matching and 3-D filtering and weighted nuclear norm minimization. We use a denoising benchmark to train a generative adversarial network for noise modeling and produce more data indistinguishable from the original ones, which can be regarded as a data augmentation scheme. Then, we utilize this extended dataset, including real images and synthetic images to train a CNN-based denoiser named DRNet for real photograph denoising. In the design of DRNet, we introduce a noise estimation module to improve the robustness of the single learning framework for handling unknown noise levels and a pair of reversible downsampling and upsampling operators to enlarge the receptive field. Experiments on real-world noisy images are conducted to evaluate our algorithm, and the results show that DRNet is effective for real photographs in comparison with other methods, especially in balancing the noise removal and the structure preservation.
The loss function plays an important role in model training for the single-image super-resolution task. Most convolutional neural network-based models adopt conventional pixel-wise loss functions to make impressive advances in peak signal-to-noise ratio and structural similarity index. However, these losses tend to find the average of plausible solutions, which lead to overly smoothed SR results with a lower visual perception. We propose a loss function combining the statistics loss with semantic priors and the quality assessment loss to produce an HR image with high visual quality while maintaining natural image statistics, as perceived by human observers. Our statistics loss measures the similarity of deep feature distributions in different semantic blocks and contributes to the maintenance of natural internal statistics in image restoration. Additionally, a no-reference quality metric that focuses on several aspects of human perceptual preferences for lighting, tone, and sharpness is introduced in our loss function to provide a more visually compelling approximation of human visual perception for perceptual image super-resolution. Experiments prove that our loss function can effectively guide the network to generate images of high-perceptual quality while considering the structural distortion for single-image super-resolution.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.