Denoising Process Research Articles

Diffusion-model-based inverse problem processing for optically-measured sound field

This paper proposes a diffusion-model-based method for addressing inverse problems in optical sound-field imaging. Optical sound-field imaging, known for its high spatial resolution, measures sound by detecting small variations in the refractive index of air caused by sound but often suffers from unavoidable noise contamination. Therefore, we present a diffusion model-based approach for sound-field inverse problems, including denoising, noisy sound-field reconstruction and extrapolation. During inference, sound-field degradation is introduced into the inverse denoising process, with range-null space decomposition used as a solver to handle degradation, iteratively generating degraded sound-field information. Numerical experiments show that our method outperforms other deep-learning-based methods in denoising and reconstruction tasks, and obtains effective results in extrapolation task. The experimental results demonstrate the applicability of our model to the real world.

EI-ISOA-VMD: Adaptive denoising and detrending method for nuclear circulating water pump impeller

As a key component of the nuclear circulating water pump, it is difficult to extract effective information from low-quality impeller signal because of the interference noise and non-stationary trend, such as water flow impact, vibrations from unrelated components, and sensor noise. Traditional signal denoising methods, particularly those utilizing variational mode decomposition, often require manual parameter tuning, resulting in unsatisfactory outcome. To address these issues, this paper proposes a signal denoising and detrending method based on the evaluation index driven the improved seagull algorithm optimized for variational mode decomposition (EI-ISOA-VMD). During the optimization process, sinusoidal chaotic mapping is utilized for initializing the seagull population, while a tanh function is employed to design nonlinearly decreasing inertia weights. Furthermore, the Levy flight mechanism enhances the randomness of position updates and optimizes seagull individuals beyond the boundary range in order to form the improved seagull optimization algorithm. The denoising, detrending process, and evaluation index proposed in this paper are utilized as the fitness function for the improved seagull optimization algorithm to optimize key parameters of variational mode decomposition. During the denoising and detrending process, the trend component is initially selected using the mean ratio method. The useful, low noise and high noise intrinsic mode functions are screened by correlation coefficient and weighted permutation entropy. Subsequently, the low noise components are denoised by wavelet thresholding method, and the high noise components are discarded to finally reconstruct the signal. The experimental results demonstrate that the proposed method effectively eliminates noise and trend components in low-quality signal while preserving more valuable information.

SITF: A Self-Supervised Iterative Training Framework for Point Cloud Denoising

Despite existing supervised point cloud denoising methods having made great progress, they require paired ideal noisy-clean datasets for training which is expensive and impractical in real-world applications. Moreover, they may perform the denoising process multiple times with fixed network parameters for better denoising results at test time. To address above issues, this paper proposes a self-supervised iterative training framework (SITF) for point cloud denoising, which only requires single noisy point clouds and a noise model. Given an off-the-shelf denoising network and original noisy point clouds, firstly, an intermediate noisier-noisy dataset is created by adding additional noises from the known noise model to noisy point clouds (i.e. learning targets). Secondly, after training on the noisier-noisy dataset, the denoising network is employed to denoise the original noisy point clouds to obtain the learning targets for the next iteration. The above two steps are iteratively and alternatively performed to get a better and better trained denoising network. Furthermore, to get better learning targets for the next round, this paper also proposes a novel iterative denoising network (IDN) architecture of stacked source attention denoising modules. The IDN explicitly models the iterative denoising process internally within a single network via reforming the given denoising network. Experimental results show that existing supervised networks trained through the SITF can achieve competitive denoising results and even outperform supervised networks under high noise conditions. The source code can be found at: https://github.com/VCG-NJUST/SITF.

HPIDN: A Hierarchical prior-guided iterative denoising network with global–local fusion for enhancing low-dose CT images

Low-dose computed tomography (LDCT) is an emerging medical diagnostic tool that reduces radiation exposure but suffers from noise retention. Current CNN-based LDCT denoising algorithms struggle to capture comprehensive global representations, impacting diagnostic accuracy. To address this, we propose a novel Hierarchical Prior-guided Iterative Denoising Network (HPIDN) for LDCT images, consisting of two main modules: the Dynamic Feature Extraction and Fusion Module (DFEFM) and the Feature-domain Iterative Denoising Module (FIDM). DFEFM dynamically captures a comprehensive representation, encompassing detailed local features in intra-relationships and complex global features in inter-relationships. It effectively guides the multi-stage iterative denoising process. FIDM hierarchically fuses the prior with image features from DFEFM by using the dual-domain attention fusion sub-network (DAFSN), enhancing denoising robustness and adaptability. This yields higher-quality images with reduced noise artifacts. Extensive experiments on the Mayo and ELCAP Datasets demonstrate the superior performance of our method quantitatively and qualitatively, improving diagnostic accuracy of lung diseases.

OriFlexClust denoising method for shallow water ATL03 single-photon point clouds

Abstract Conventional clustering algorithms suffer from misclassification problems caused by uneven density and poor initial parameter choices. This paper introduces an adaptive denoising algorithm called OriFlexClust, based on directional angle continuity. Experimental results demonstrate that the OriFlexClust algorithm outperforms other clustering algorithms, effectively addressing the issues of parameter selection and uneven data density. It enhances denoising accuracy and processing speed. Therefore, the OriFlexClust algorithm proposed in this paper is more superior.

VCC-DiffNet: Visual Conditional Control Diffusion Network for Remote Sensing Image Captioning

Pioneering remote sensing image captioning (RSIC) works use autoregressive decoding for fluent and coherent sentences but suffer from high latency and high computation costs. In contrast, non-autoregressive approaches improve inference speed by predicting multiple tokens simultaneously, though at the cost of performance due to a lack of sequential dependencies. Recently, diffusion model-based non-autoregressive decoding has shown promise in natural image captioning with iterative refinement, but its effectiveness is limited by the intrinsic characteristics of remote sensing images, which complicate robust input construction and affect the description accuracy. To overcome these challenges, we propose an innovative diffusion model for RSIC, named the Visual Conditional Control Diffusion Network (VCC-DiffNet). Specifically, we propose a Refined Multi-scale Feature Extraction (RMFE) module to extract the discernible visual context features of RSIs as input of the diffusion model-based non-autoregressive decoder to conditionally control a multi-step denoising process. Furthermore, we propose an Interactive Enhanced Decoder (IE-Decoder) utilizing dual image–description interactions to generate descriptions finely aligned with the image content. Experiments conducted on four representative RSIC datasets demonstrate that our non-autoregressive VCC-DiffNet performs comparably to, or even better than, popular autoregressive baselines in classic metrics, achieving around an 8.22× speedup in Sydney-Captions, an 11.61× speedup in UCM-Captions, a 15.20× speedup in RSICD, and an 8.13× speedup in NWPU-Captions.

LDCT image denoising algorithm based on two-dimensional variational mode decomposition and dictionary learning

Low-dose X-CT scanning method effectively reduces radiation hazards, however, reducing the radiation dose will introduce noise and artifacts during the projection process, resulting in a decrease in the quality of the reconstructed image. To address this problem, we combined 2D variational modal decomposition and dictionary learning. We proposed a low-dose CT (LDCT) image denoising algorithm based on an improved K-SVD algorithm with image decomposition. The dictionary obtained by K-SVD training lacks consideration of image structure information. To address this problem, we employ the two-dimensional variational mode decomposition (2D-VMD) method to decompose the image into distinct modal components. Through the adaptive learning of dictionaries based on the characteristics of each modal component, independent denoising processing is applied to each component, avoiding the loss of structural and detailed information in the image. In addition, we introduce the regularized orthogonal matching pursuit algorithm (ROMP) and dictionary atom optimization method to improve the sparse representation ability of the dictionary and reduce the impact of noise atoms on denoising performance. The experiments show that the proposed method outperforms other denoising methods regarding peak signal-to-noise ratio and structural similarity. The proposed method maintains the denoised image details and structural information while removing LDCT image noise and artifacts. The image quality after denoising is significantly improved and facilitates more accurate detection and analysis of lesion areas.

ReF-DDPM: A novel DDPM-based data augmentation method for imbalanced rolling bearing fault diagnosis

Effective bearing fault diagnosis is crucial to ensure the safety and reliability of mechanical systems. Due to the complex and harsh working environment, mechanical data often comes from imbalanced datasets, which is a pressing problem in diagnosis applications. However, currently proposed data augmentation methods mainly based on generative adversarial networks, remain challenging in balancing the quality and diversity of the generation samples. To solve it, this paper proposes a new data enhancement method called the reparameterized residual denoising diffusion probability model (ReF-DDPM) and applies it to fault diagnosis. The proposed architecture includes a forward diffusion process and a reverse denoising process, where Gaussian noise and original samples are transformed by Markov chains. To improve the quality of generation samples, the noise prediction network is modified for better feature representation by enhancing intra-level and inter-level features. Furthermore, signal labels are added to the model as conditional information to direct the generation of relevant category samples during the sampling process. The study provides a new data augmentation method for bearing imbalanced data, and generation data can be further used for fault diagnosis tasks. Verification experiments demonstrate the effectiveness and good generalization of the method, and improve the accuracy of imbalance fault diagnosis.

A Side-Lobe Denoise Process for ISAR Imaging Applications: Combined Fast Clean and Spatial Focus Technique

A Side-Lobe Denoise Process for ISAR Imaging Applications: Combined Fast Clean and Spatial Focus Technique

Innovative Approach to Dam Deformation Analysis: Integration of VMD, Fractal Theory, and WOA-DELM

This paper introduces a novel and comprehensive model for the analysis of dam deformation trends, integrating the variational mode decomposition (VMD) method, fractal theory, and the whale optimization algorithm (WOA) to refine the deep extreme learning machine (DELM) model. This integration allows for a meticulous denoising process through VMD, effectively isolating pertinent signal characteristics from noise and measurement interference. Following this, fractal theory is utilized to conduct an in-depth qualitative analysis of the denoised data, capturing intricate patterns within the deformation trends. The model further evolves with the application of WOA to optimize the DELM model, thereby facilitating an integrated approach that merges qualitative insights with quantitative analysis. The efficacy of this advanced model is demonstrated through a case study, highlighting its capability to deliver accurate and reliable predictions that are in harmony with practical engineering scenarios. This research not only offers a robust framework for analyzing dam deformation trends but also sets a new standard in the field, providing a new solution for assessing structural integrity in hydrological engineering.

Deep Learning-Based De-blurring/Denoising of Indian Heritage Images

Using the deep learning algorithms, our method comes up with the process that impedes the conservation and restores the Indian heritage images, while dealing with the obstacles like blur, noise, and obscurity of the images. In this case, neural networks that are developed themselves are used to carry out the process of denoising and reconstruction of these images but also contain intricate object detection capabilities. These multi-faceted actions, therefore, help to preserve the historically relevant aspects and make the Indian cultural heritage more colorful and accessible to the outside world. In addition, not only are we protecting them but we are also transforming them into the best images that represent India's past. Through application of deep learning principles to indigenous materials, we come up with powerful educational tools that make a deep impact in the community, promoting a profound appreciation for India's cultural heritage. By means of careful experimentation and stringent testing, we establish that the specified methodology is indeed workable in conservation applications as it has been shown to be effective. This way, we make it clear that the cultural heritage of India is not only treasures of the ancient time but the living legacy, which is welcoming for everyone and contributes to the cultural diversity and pride.

Uncertainty-aware pedestrian trajectory prediction via distributional diffusion

Tremendous efforts have been put forth on predicting pedestrian trajectory with generative models to accommodate uncertainty and multi-modality in human behaviors. An individual’s inherent uncertainty, e.g., change of destination, can be masked by complex patterns resulting from the movements of interacting pedestrians. However, latent variable-based generative models often entangle such uncertainty with complexity, leading to limited either latent expressivity or predictive diversity. In this work, we propose to separately model these two factors by implicitly deriving a flexible latent representation to capture intricate pedestrian movements, while integrating predictive uncertainty of individuals with explicit bivariate Gaussian mixture densities over their future locations. More specifically, we present a model-agnostic uncertainty-aware pedestrian trajectory prediction framework, parameterizing sufficient statistics for the mixture of Gaussians that jointly comprise the multi-modal trajectories. We further estimate these parameters of interest by approximating a denoising process that progressively recovers pedestrian movements from noise. Unlike previous studies, we translate the predictive stochasticity to explicit distributions, allowing it to readily generate plausible future trajectories indicating individuals’ self-uncertainty. Moreover, our framework is compatible with different neural net architectures. We empirically show the performance gains over state-of-the-art even with lighter backbones, across most scenes on two public benchmarks.

Bidirectional image denoising with blurred image feature

Image denoising remains a classical yet still challenging problem, because the noise can destroy details and cause severe information loss. In recent years, various well-designed CNN-based methods have been extensively applied in image denoising because of the strong learning ability. However, most of them share an unidirectional procedure, which directly learns a mapping from noisy input to a clean image without focusing on the over-smoothed state of the denoising process, limiting the richness of extracted features. Different from previous works, we propose a blurred image feature guided CNN (BFCNN) network that implements a novel blurring-adjusting strategy to address the complex denoising problem via two stages. In stage 1, we build a blurring module (BM) to capture over-smoothed features from noisy observations and generate the blurred image restoration, which is a less informative version of the clean image. Furthermore, a multi-level concatenating module (CM) and an adjusting module (AM) are then designed to recover more detailed information in stage 2. These two modules are jointly designed to restore a properly-smoothed image from the over-smoothed blurred image and the given under-smoothed noisy image. Comparing to the traditional denoising process, the proposed blurring-adjusting strategy produces a precise denoised image more efficiently by converting the unidirectional denoising process into a bidirectional denoising process. To our knowledge, this is the first study that utilizes the over-smoothed image to address the denoising problem. Extensive experiments demonstrate the superiority of our BFCNN with more accurate reconstruction quality and achieve competitive quantitative results among current CNN-based methods. This research will release the code upon acceptance.

Spatial virtual recovery design technology of ancient buildings based on 3D laser scanning technology

Due to the persistent advancements in science and technology, the implementation of digital technology in the preservation of cultural heritage is progressively expanding. One of the focal points of research lies in the spatial virtual restoration design technology of ancient edifices, which is based on three-dimensional laser scanning technology. The study initially outlines the process of obtaining point cloud data via 3D laser scanning technology and subsequently executes denoising, splicing, and simplification processing on the point cloud. Subsequently, the pre-processed data undergo 3D mesh model construction. Finally, an effective repair technique is implemented to address the void phenomenon present during the modeling process. To construct the implicit surface, the RBF method is employed and new vertices are adjusted, resulting in a more accurate spatial virtual restoration of ancient buildings (ABs). The study found that using the radial basis function based void repair method resulted in a mean void repair accuracy of 92.38% in the west wall, an improvement of 3.08% compared to the Liepa-based method. In addition, this method achieved the highest accuracy of 98.94% on the north wall, improving by 6.37% compared to the Poisson grid editing algorithm. Meanwhile, the RBF-based method repaired cavities in the west wall model with an average runtime of only 20.613 s, resulting in a 19.16 s reduction compared to the Liepa-based method. In addition, the method’s average repair time for the north wall was only 5.364 s, a decrease of 13.28 s compared to the Poisson grid editing algorithm. This shows that combining 3D laser scanning technology and cavity repair technology can acquire high-quality point cloud data of historic structures, create precise 3D models, and achieve spatial virtual restoration. This offers a new and efficient technological approach to preserving and passing down ABs.

Rapid 2D 23Na MRI of the calf using a denoising convolutional neural network

Purpose23Na MRI can be used to quantify in-vivo tissue sodium concentration (TSC), but the inherently low 23Na signal leads to long scan times and/or noisy or low-resolution images. Reconstruction algorithms such as compressed sensing (CS) have been proposed to mitigate low signal-to-noise ratio (SNR); although, these can result in unnatural images, suboptimal denoising and long processing times. Recently, machine learning has been increasingly used to denoise 1H MRI acquisitions; however, this approach typically requires large volumes of high-quality training data, which is not readily available for 23Na MRI. Here, we propose using 1H data to train a denoising convolutional neural network (CNN), which we subsequently demonstrate on prospective 23Na images of the calf. Methods1893 1H fat-saturated transverse slices of the knee from the open-source fastMRI dataset were used to train denoising CNNs for different levels of noise. Synthetic low SNR images were generated by adding gaussian noise to the high-quality 1H k-space data before reconstruction to create paired training data. For prospective testing, 23Na images of the calf were acquired in 10 healthy volunteers with a total of 150 averages over ten minutes, which were used as a reference throughout the study. From this data, images with fewer averages were retrospectively reconstructed using a non-uniform fast Fourier transform (NUFFT) as well as CS, with the NUFFT images subsequently denoised using the trained CNN. ResultsCNNs were successfully applied to 23Na images reconstructed with 50, 40 and 30 averages. Muscle and skin apparent TSC quantification from CNN-denoised images were equivalent to those from CS images, with <0.9 mM bias compared to reference values. Estimated SNR was significantly higher in CNN-denoised images compared to NUFFT, CS and reference images. Quantitative edge sharpness was equivalent for all images. For subjective image quality ranking, CNN-denoised images ranked equally best with reference images and significantly better than NUFFT and CS images. ConclusionDenoising CNNs trained on 1H data can be successfully applied to 23Na images of the calf; thus, allowing scan time to be reduced from ten minutes to two minutes with little impact on image quality or apparent TSC quantification accuracy.

Improving the spatial resolution of solar images using super-resolution diffusion generative adversarial networks

High-spatial-resolution solar images contribute to the study of small-scale structures on the Sun. The Helioseismic and Magnetic Imager (HMI) conducts continuous full-disk observations of the Sun at a fixed cadence, accumulating a wealth of observational data. However, the spatial resolution of HMI images is not sufficient to analyze the small-scale structures of solar activity. We present a new super-resolution (SR) method based on generative adversarial networks (GANs) and denoising diffusion probabilistic models (DDPMs) that can increase the spatial resolution of HMI images by a factor four. We propose a method called super-resolution diffusion GANs (SDGAN), which combines GANs and DDPMs for the SR reconstruction of HMI images. SDGAN progressively maps low-resolution (LR) images to high-resolution (HR) images through a conditional denoising process. It employs conditional GANs to simulate the denoising distribution and optimizes model results using nonsaturating adversarial loss and perceptual loss. This approach enables fast and high-quality reconstruction of solar images. We used high-spatial-resolution images from the Goode Solar Telescope (GST) as HR images and created a data set consisting of paired images from HMI and GST. We then used this data set to train SDGAN for the purpose of reconstructing HMI images with four times the original spatial resolution. The experimental results demonstrate that SDGAN can obtain high-quality HMI reconstructed images with just four denoising steps.

OIF-Net: An Optical Flow Registration-Based PET/MR Cross-Modal Interactive Fusion Network for Low-Count Brain PET Image Denoising.

The short frames of low-count positron emission tomography (PET) images generally cause high levels of statistical noise. Thus, improving the quality of low-count images by using image postprocessing algorithms to achieve better clinical diagnoses has attracted widespread attention in the medical imaging community. Most existing deep learning-based low-count PET image enhancement methods have achieved satisfying results, however, few of them focus on denoising low-count PET images with the magnetic resonance (MR) image modality as guidance. The prior context features contained in MR images can provide abundant and complementary information for single low-count PET image denoising, especially in ultralow-count (2.5%) cases. To this end, we propose a novel two-stream dual PET/MR cross-modal interactive fusion network with an optical flow pre-alignment module, namely, OIF-Net. Specifically, the learnable optical flow registration module enables the spatial manipulation of MR imaging inputs within the network without any extra training supervision. Registered MR images fundamentally solve the problem of feature misalignment in the multimodal fusion stage, which greatly benefits the subsequent denoising process. In addition, we design a spatial-channel feature enhancement module (SC-FEM) that considers the interactive impacts of multiple modalities and provides additional information flexibility in both the spatial and channel dimensions. Furthermore, instead of simply concatenating two extracted features from these two modalities as an intermediate fusion method, the proposed cross-modal feature fusion module (CM-FFM) adopts cross-attention at multiple feature levels and greatly improves the two modalities' feature fusion procedure. Extensive experimental assessments conducted on real clinical datasets, as well as an independent clinical testing dataset, demonstrate that the proposed OIF-Net outperforms the state-of-the-art methods.

Towards Detailed Text-to-Motion Synthesis via Basic-to-Advanced Hierarchical Diffusion Model

Text-guided motion synthesis aims to generate 3D human motion that not only precisely reflects the textual description but reveals the motion details as much as possible. Pioneering methods explore the diffusion model for text-to-motion synthesis and obtain significant superiority. However, these methods conduct diffusion processes either on the raw data distribution or the low-dimensional latent space, which typically suffer from the problem of modality inconsistency or detail-scarce. To tackle this problem, we propose a novel Basic-to-Advanced Hierarchical Diffusion Model, named B2A-HDM, to collaboratively exploit low-dimensional and high-dimensional diffusion models for high quality detailed motion synthesis. Specifically, the basic diffusion model in low-dimensional latent space provides the intermediate denoising result that to be consistent with the textual description, while the advanced diffusion model in high-dimensional latent space focuses on the following detail-enhancing denoising process. Besides, we introduce a multi-denoiser framework for the advanced diffusion model to ease the learning of high-dimensional model and fully explore the generative potential of the diffusion model. Quantitative and qualitative experiment results on two text-to-motion benchmarks (HumanML3D and KIT-ML) demonstrate that B2A-HDM can outperform existing state-of-the-art methods in terms of fidelity, modality consistency, and diversity.

Terrain Diffusion Network: Climatic-Aware Terrain Generation with Geological Sketch Guidance

Sketch-based terrain generation seeks to create realistic landscapes for virtual environments in various applications such as computer games, animation and virtual reality. Recently, deep learning based terrain generation has emerged, notably the ones based on generative adversarial networks (GAN). However, these methods often struggle to fulfill the requirements of flexible user control and maintain generative diversity for realistic terrain. Therefore, we propose a novel diffusion-based method, namely terrain diffusion network (TDN), which actively incorporates user guidance for enhanced controllability, taking into account terrain features like rivers, ridges, basins, and peaks. Instead of adhering to a conventional monolithic denoising process, which often compromises the fidelity of terrain details or the alignment with user control, a multi-level denoising scheme is proposed to generate more realistic terrains by taking into account fine-grained details, particularly those related to climatic patterns influenced by erosion and tectonic activities. Specifically, three terrain synthesisers are designed for structural, intermediate, and fine-grained level denoising purposes, which allow each synthesiser concentrate on a distinct terrain aspect. Moreover, to maximise the efficiency of our TDN, we further introduce terrain and sketch latent spaces for the synthesizers with pre-trained terrain autoencoders. Comprehensive experiments on a new dataset constructed from NASA Topology Images clearly demonstrate the effectiveness of our proposed method, achieving the state-of-the-art performance. Our code is available at https://github.com/TDNResearch/TDN.

SkipDiff: Adaptive Skip Diffusion Model for High-Fidelity Perceptual Image Super-resolution

It is well-known that image quality assessment usually meets with the problem of perception-distortion (p-d) tradeoff. The existing deep image super-resolution (SR) methods either focus on high fidelity with pixel-level objectives or high perception with generative models. The emergence of diffusion model paves a fresh way for image restoration, which has the potential to offer a brand-new solution for p-d trade-off. We experimentally observed that the perceptual quality and distortion change in an opposite direction with the increase of sampling steps. In light of this property, we propose an adaptive skip diffusion model (SkipDiff), which aims to achieve high-fidelity perceptual image SR with fewer sampling steps. Specifically, it decouples the sampling procedure into coarse skip approximation and fine skip refinement stages. A coarse-grained skip diffusion is first performed as a high-fidelity prior to obtaining a latent approximation of the full diffusion. Then, a fine-grained skip diffusion is followed to further refine the latent sample for promoting perception, where the fine time steps are adaptively learned by deep reinforcement learning. Meanwhile, this approach also enables faster sampling of diffusion model through skipping the intermediate denoising process to shorten the effective steps of the computation. Extensive experimental results show that our SkipDiff achieves superior perceptual quality with plausible reconstruction accuracy and a faster sampling speed.