A Generative Adversarial Network-Based Image Denoiser Controlling Heterogeneous Losses

Abstract

We propose a novel generative adversarial network (GAN)-based image denoising method that utilizes heterogeneous losses. In order to improve the restoration quality of the structural information of the generator, the heterogeneous losses, including the structural loss in addition to the conventional mean squared error (MSE)-based loss, are used to train the generator. To maximize the improvements brought on by the heterogeneous losses, the strength of the structural loss is adaptively adjusted by the discriminator for each input patch. In addition, a depth wise separable convolution-based module that utilizes the dilated convolution and symmetric skip connection is used for the proposed GAN so as to reduce the computational complexity while providing improved denoising quality compared to the convolutional neural network (CNN) denoiser. The experiments showed that the proposed method improved visual information fidelity and feature similarity index values by up to 0.027 and 0.008, respectively, compared to the existing CNN denoiser.

Multiscale residual fusion network for image denoising

“…(1) As mentioned before, its basic function is to enlarge the receptive field of networks [31,32]. Thus, early works directly replace traditional convolutions by dilated ones [10,33]. Several modified networks were proposed by introducing dilated convolutions in the framework of DnCNN [27,34].…”

Efficient Iterative Regularization Method for Total Variation-Based Image Restoration

“…If we ignore the specific form of ϕ(x), superior denoisers such as non-local means filter [21], block-matching and 3D filtering (BM3D) [22], bilateral filter [23], and adversarial Gaussian denoiser [24], can be adopted for solving this denoising subproblem. Moreover, with the rapid development of deep learning technology in image denoising, super-resolution reconstruction, object detection and control [25,26], deep learning methods [27][28][29][30] using clean-noisy image pairs have been widely exploited in the design of denoisers. Multi-layer perceptron was adopted for image restoration in [27] while various convolutional neural network (CNN) and generative adversarial networks methods have been used to design specific denoisers [28][29][30].…”

“…Moreover, with the rapid development of deep learning technology in image denoising, super-resolution reconstruction, object detection and control [25,26], deep learning methods [27][28][29][30] using clean-noisy image pairs have been widely exploited in the design of denoisers. Multi-layer perceptron was adopted for image restoration in [27] while various convolutional neural network (CNN) and generative adversarial networks methods have been used to design specific denoisers [28][29][30]. It is well known that neural network methods are limited in computing speed and high requirements of hardware, with no universal adaptation for the applications that require simplicity and rapidity.…”