Search

Image denoising is a significant task in computer vision. Previous studies have mostly concentrated on removing noise with specific levels. The blind image denoising approach has recently gained more popularity due to its adaptability. Nonetheless, existing deep learning methods only train networks to learn the direct projection from noisy images to clean ones, which limits their denoising performance. This paper proposes a novel perspective for blind denoising by converting the static image denoising problem into a dynamic process inspired by the diffusion model. To achieve this, we present a new method that views a noisy image as a mid-state of a Gaussian diffusion process. Specifically, the image noise is separated into multiple sub-level noises through the diffusion process, and subsequently eliminated in a sequential manner. Furthermore, we propose a diffusion denoising network that comprises a Feature Extraction Module for extracting image features and a Diffusion Noise Estimation Module for estimating the sub-level noises. Our experiments demonstrate that our proposed method outperforms existing methods and achieves state-of-the-art results in blind additive white Gaussian noise and real-world image denoising. [ABSTRACT FROM AUTHOR]

Recently, convolutional neural networks (CNNs) and vision transformers (ViTs) have emerged as powerful tools for image restoration (IR). Nonetheless, they encountered some limitations due to their characteristics, such as CNNs sacrificing global reception and ViTs requiring large memory and graphics resources. To address these limitations and explore an alternative approach for improved IR performance, we propose two clustering-based frameworks for general IR tasks, which are style-guided context cluster U-Net (SCoC-UNet) and style-guided clustered point interaction U-Net (SCPI-UNet). The SCoC-UNet adopts a U-shaped architecture, comprising position embedding, Encoder, Decoder, and reconstruction block. Specifically, the input low-quality image is viewed as a set of unorganized points, each of which is first given location information by the continuous relative position embedding method. These points are then fed into a symmetric Encoder and Decoder which utilize style-guided context cluster (SCoC) blocks to extract potential context features and high-frequency information. Although SCoC-UNet has obtained decent performance for image restoration, its SCoC block can only capture connectivity at points within the same cluster, which may ignore long-range dependencies in different clusters. To address this issue, we further propose a SCPI-UNet based on SCoC-UNet, which leverages a style-guided clustered point interaction (SCPI) block in place of the SCoC block. The SCPI block utilizes a cross-attention mechanism to establish the connections of feature points between different clusters. Extensive experimental results demonstrate that the proposed SCoC-UNet and SCPI-UNet can handle several typical IR tasks (i.e., JPEG compression artifact reduction, image denoising, and super-resolution) and achieve superior quantitative and qualitative performance over some state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Effective aggregation of temporal information of consecutive frames is the core of achieving video super-resolution. Many scholars have utilized structures such as sliding windows and recurrences to gather the spatio-temporal information of frames. However, although the performances of constructed video super-resolution models are improving, the sizes of the models are also increasing, exacerbating the demand on the equipment. Thus, to reduce the stress on the device, we propose a novel lightweight recurrent grouping attention network. The parameters of this model are only 0.878 M, which is much lower than the current mainstream model for studying video super-resolution. We have designed a forward feature extraction module and a backward feature extraction module to collect temporal information between consecutive frames from two directions. Moreover, a new grouping mechanism is proposed to efficiently collect spatio-temporal information of the reference frame and its neighboring frames. The attention supplementation module is presented to further enhance the information gathering range of the model. The feature reconstruction module aims to aggregate information from different directions to reconstruct high-resolution features. Experiments demonstrate that our model achieves state-of-the-art performance on multiple datasets. [ABSTRACT FROM AUTHOR]

Recently, convolutional neural networks (CNNs) have been widely used in image denoising, and their performance has been enhanced through residual learning. However, previous research mostly focused on optimizing the network architecture of CNNs, ignoring the limitations of the commonly used residual learning. This paper identifies two of its limitations, which are the neglect of image information and the lack of effective consideration of image self-similarity. To solve these limitations, this paper proposes a compositional denoising network (CDN), which contains two sub-paths, the image information path (IIP) and the noise estimation path (NEP), respectively. IIP is trained via an image-to-image method to extract image information. For NEP, it utilizes image self-similarity from the perspective of training. This similarity-based training method constrains NEP to output similar estimated noise distributions for different image patches with a specific kind of noise. Finally, image information and noise distribution information are comprehensively considered for image denoising. Experimental results indicate that CDN outperforms other CNN-based methods in both synthetic and real-world image denoising, achieving state-of-the-art performance. [ABSTRACT FROM AUTHOR]

We use a positive parameter to develop a differentiable perturbed reconstruction model to solve the L1-L2 magnetic resonance image (MRI) reconstruction problem. We use Bregman iterative formulation to solve the differentiable perturbed L1-L2 model, and lagged diffusivity fixed-point iteration to solve the minimization problem in the Bregman iteration. Two Fourier transforms and an inverse Fourier transform are used to accelerate L1-L2 MRI reconstruction. Real MR images are used to test the method in numerical experiments. The results demonstrate that the proposed method is very efficient for L1-L2 MRI reconstruction. [ABSTRACT FROM AUTHOR]

We use a small positive parameter to change the total-variation function for unconstrained MR image reconstruction to a strictly convex perturbed function. Bregman iteration is applied to solve the modified total-variation MR image (TVMRI) reconstruction problem. A lagged diffusivity fixed-point algorithm is applied to solve the minimization problem in the Bregman iteration. We use the periodic boundary condition and a Fourier transform to accelerate TVMRI reconstruction. Real MR images are used to test the approach in numerical experiments. The experimental results demonstrate that the proposed method is very efficient for TVMRI reconstruction. [ABSTRACT FROM AUTHOR]