Blind Universal Bayesian Image Denoising with Gaussian Noise Level Learning

Progressive Feature Fusion Attention Dense Network for Speckle Noise Removal in OCT Images

Although deep learning for Big Data analytics has achieved promising results in the field of optical coherence tomography (OCT) image denoising, the low recognition rate caused by complex noise distribution and a large number of redundant features is still a challenge faced by deep learning-based denoising methods. Moreover, the network with large depth will bring high computational complexity. To this end, we propose a progressive feature fusion attention dense network (PFFADN) for speckle noise removal in OCT images. We arrange densely connected dense blocks in the deep convolution network, and sequentially connect the shallow convolution feature map with the deep one extracted from each dense block to form a residual block. We add attention mechanism to the network to extract the key features and suppress the irrelevant ones. We fuse the output feature maps from all dense blocks and input them to the reconstruction output layer. We compare PFFADN with the state-of-the-art denoising algorithms on retinal OCT images. Experiments show that our method has better improvement in denoising performance.

Image denoising and segmentation model construction based on IWOA-PCNN

The research suggests a method to improve the present pulse coupled neural network (PCNN), which has a complex structure and unsatisfactory performance in image denoising and image segmentation. Then, a multi strategy collaborative improvement whale optimization algorithm (WOA) is proposed, and an improved whale optimization algorithm (IWOA) is constructed. IWOA is used to find the optimal parameter values of PCNN to optimize PCNN. By combining the aforementioned components, the IWOA-PCNN model had the best image denoising performance, and the produced images were crisper and preserve more information. IWOA-PCNN processed pictures have an average PSNR of 35.87 and an average MSE of 0.24. The average processing time for photos with noise is typically 24.80 s, which is 7.30 s and 7.76 s faster than the WTGAN and IGA-NLM models, respectively. Additionally, the average NU value measures 0.947, and the average D value exceeds 1000. The aforementioned findings demonstrate that the suggested method can successfully enhance the PCNN, improving its capability for image denoising and image segmentation. This can, in part, encourage the use and advancement of the PCNN.

Bayesian Image Super-Resolution With Deep Modeling of Image Statistics

Modeling statistics of image priors is useful for image super-resolution, but little attention has been paid from the massive works of deep learning-based methods. In this work, we propose a Bayesian image restoration framework, where natural image statistics are modeled with the combination of smoothness and sparsity priors. Concretely, first we consider an ideal image as the sum of a smoothness component and a sparsity residual, and model real image degradation including blurring, downscaling, and noise corruption. Then, we develop a variational Bayesian approach to infer their posteriors. Finally, we implement the variational approach for single image super-resolution (SISR) using deep neural networks, and propose an unsupervised training strategy. The experiments on three image restoration tasks, i.e., ideal SISR, realistic SISR, and real-world SISR, demonstrate that our method has superior model generalizability against varying noise levels and degradation kernels and is effective in unsupervised SISR. The code and resulting models are released via https://zmiclab.github.io/projects.html.

TRNR: Task-Driven Image Rain and Noise Removal with a Few Images Based on Patch Analysis

The recent success of learning-based image rain and noise removal can be attributed primarily to well-designed neural network architectures and large labeled datasets. However, we discover that current image rain and noise removal methods result in low utilization of images. To alleviate the reliance of deep models on large labeled datasets, we propose the task-driven image rain and noise removal (TRNR) based on a patch analysis strategy. The patch analysis strategy samples image patches with various spatial and statistical properties for training and can increase image utilization. Furthermore, the patch analysis strategy encourages us to introduce the N-frequency-K-shot learning task for the task-driven approach TRNR. TRNR allows neural networks to learn from numerous N-frequency-K-shot learning tasks, rather than from a large amount of data. To verify the effectiveness of TRNR, we build a Multi-Scale Residual Network (MSResNet) for both image rain removal and Gaussian noise removal. Specifically, we train MSResNet for image rain removal and noise removal with a few images (for example, 20.0% train-set of Rain100H). Experimental results demonstrate that TRNR enables MSResNet to learn more effectively when data is scarce. TRNR has also been shown in experiments to improve the performance of existing methods. Furthermore, MSResNet trained with a few images using TRNR outperforms most recent deep learning methods trained data-driven on large labeled datasets. These experimental results have confirmed the effectiveness and superiority of the proposed TRNR. The source code is available on https://github.com/Schizophreni/MSResNet-TRNR.

Transformer-based progressive residual network for single image dehazing

Introduction

The seriously degraded fogging image affects the further visual tasks. How to obtain a fog-free image is not only challenging, but also important in computer vision. Recently, the vision transformer (ViT) architecture has achieved very efficient performance in several vision areas.

Methods

In this paper, we propose a new transformer-based progressive residual network. Different from the existing single-stage ViT architecture, we recursively call the progressive residual network with the introduction of swin transformer. Specifically, our progressive residual network consists of three main components: the recurrent block, the transformer codecs and the supervise fusion module. First, the recursive block learns the features of the input image, while connecting the original image features of the original iteration. Then, the encoder introduces the swin transformer block to encode the feature representation of the decomposed block, and continuously reduces the feature mapping resolution to extract remote context features. The decoder recursively selects and fuses image features by combining attention mechanism and dense residual blocks. In addition, we add a channel attention mechanism between codecs to focus on the importance of different features.

Results and discussion

The experimental results show that the performance of this method outperforms state-of-the-art handcrafted and learning-based methods.

WINNet: Wavelet-Inspired Invertible Network for Image Denoising

Image denoising aims to restore a clean image from an observed noisy one. Model-based image denoising approaches can achieve good generalization ability over different noise levels and are with high interpretability. Learning-based approaches are able to achieve better results, but usually with weaker generalization ability and interpretability. In this paper, we propose a wavelet-inspired invertible network (WINNet) to combine the merits of the wavelet-based approaches and learning-based approaches. The proposed WINNet consists of K -scale of lifting inspired invertible neural networks (LINNs) and sparsity-driven denoising networks together with a noise estimation network. The network architecture of LINNs is inspired by the lifting scheme in wavelets. LINNs are used to learn a non-linear redundant transform with perfect reconstruction property to facilitate noise removal. The denoising network implements a sparse coding process for denoising. The noise estimation network estimates the noise level from the input image which will be used to adaptively adjust the soft-thresholds in LINNs. The forward transform of LINNs produces a redundant multi-scale representation for denoising. The denoised image is reconstructed using the inverse transform of LINNs with the denoised detail channels and the original coarse channel. The simulation results show that the proposed WINNet method is highly interpretable and has strong generalization ability to unseen noise levels. It also achieves competitive results in the non-blind/blind image denoising and in image deblurring.

BIGPrior: Toward Decoupling Learned Prior Hallucination and Data Fidelity in Image Restoration

Classic image-restoration algorithms use a variety of priors, either implicitly or explicitly. Their priors are hand-designed and their corresponding weights are heuristically assigned. Hence, deep learning methods often produce superior image restoration quality. Deep networks are, however, capable of inducing strong and hardly predictable hallucinations. Networks implicitly learn to be jointly faithful to the observed data while learning an image prior; and the separation of original data and hallucinated data downstream is then not possible. This limits their wide-spread adoption in image restoration. Furthermore, it is often the hallucinated part that is victim to degradation-model overfitting. We present an approach with decoupled network-prior based hallucination and data fidelity terms. We refer to our framework as the Bayesian Integration of a Generative Prior (BIGPrior). Our method is rooted in a Bayesian framework and tightly connected to classic restoration methods. In fact, it can be viewed as a generalization of a large family of classic restoration algorithms. We use network inversion to extract image prior information from a generative network. We show that, on image colorization, inpainting and denoising, our framework consistently improves the inversion results. Our method, though partly reliant on the quality of the generative network inversion, is competitive with state-of-the-art supervised and task-specific restoration methods. It also provides an additional metric that sets forth the degree of prior reliance per pixel relative to data fidelity.

AVLSM: Adaptive Variational Level Set Model for Image Segmentation in the Presence of Severe Intensity Inhomogeneity and High Noise

Intensity inhomogeneity and noise are two common issues in images but inevitably lead to significant challenges for image segmentation and is particularly pronounced when the two issues simultaneously appear in one image. As a result, most existing level set models yield poor performance when applied to this images. To this end, this paper proposes a novel hybrid level set model, named adaptive variational level set model (AVLSM) by integrating an adaptive scale bias field correction term and a denoising term into one level set framework, which can simultaneously correct the severe inhomogeneous intensity and denoise in segmentation. Specifically, an adaptive scale bias field correction term is first defined to correct the severe inhomogeneous intensity by adaptively adjusting the scale according to the degree of intensity inhomogeneity while segmentation. More importantly, the proposed adaptive scale truncation function in the term is model-agnostic, which can be applied to most off-the-shelf models and improves their performance for image segmentation with severe intensity inhomogeneity. Then, a denoising energy term is constructed based on the variational model, which can remove not only common additive noise but also multiplicative noise often occurred in medical image during segmentation. Finally, by integrating the two proposed energy terms into a variational level set framework, the AVLSM is proposed. The experimental results on synthetic and real images demonstrate the superiority of AVLSM over most state-of-the-art level set models in terms of accuracy, robustness and running time.

Bi-Directional Filter for the Removal of Lines and Cracks in Images

In this paper, a method for the removal of noisy lines and cracks corrupted by different noise types is explored, using a cascade of filtering cycles based on the principle of symmetry among neighboring pixels. Each filtering cycle includes a filter in two perpendicular directions, one horizontal and the other vertical. Any pixel, to be deemed original, should have a number of symmetric pixels within its neighboring pixels greater than the number specified by the condition set for each direction in all the filters. Since the conditions of each filter increase gradually from one cycle to the next, it becomes more difficult for a noisy pixel to satisfy the filter conditions in each filtering cycle, while an original pixel can easily satisfy the conditions in all the filtering cycles. The reason is that a noisy pixel has a random value and therefore faces difficulty in finding a sufficient number of symmetric pixels in each direction, while an original one has a value correlated with the values of its neighboring pixels. Extensive simulation experiments prove that the proposed method efficiently detects and restores different noisy lines and cracks of different shape and thickness. Also, it retains the image details and outperforms other well-known algorithms, both objectively and subjectively. More specifically, the proposed algorithm achieves restoration performance better than the other known methods by ≥0.81dB in all simulation experiments.