Speckle Noise Reduction Research Articles

Ultrasound breast images denoising using generative adversarial networks (GANs)

INTRODUCTION: Ultrasound in conjunction with mammography imaging, plays a vital role in the early detection and diagnosis of breast cancer. However, speckle noise affects medical ultrasound images and degrades visual radiological interpretation. Speckle carries information about the interactions of the ultrasound pulse with the tissue microstructure, which generally causes several difficulties in identifying malignant and benign regions. The application of deep learning in image denoising has gained more attention in recent years. OBJECTIVES: The main objective of this work is to reduce speckle noise while preserving features and details in breast ultrasound images using GAN models. METHODS: We proposed two GANs models (Conditional GAN and Wasserstein GAN) for speckle-denoising public breast ultrasound databases: BUSI, DATASET A, AND UDIAT (DATASET B). The Conditional GAN model was trained using the Unet architecture, and the WGAN model was trained using the Resnet architecture. The image quality results in both algorithms were measured by Peak Signal to Noise Ratio (PSNR, 35–40 dB) and Structural Similarity Index (SSIM, 0.90–0.95) standard values. RESULTS: The experimental analysis clearly shows that the Conditional GAN model achieves better breast ultrasound despeckling performance over the datasets in terms of PSNR = 38.18 dB and SSIM = 0.96 with respect to the WGAN model (PSNR = 33.0068 dB and SSIM = 0.91) on the small ultrasound training datasets. CONCLUSIONS: The observed performance differences between CGAN and WGAN will help to better implement new tasks in a computer-aided detection/diagnosis (CAD) system. In future work, these data can be used as CAD input training for image classification, reducing overfitting and improving the performance and accuracy of deep convolutional algorithms.

Rigid alignment method for secondary analyses of optical coherence tomography volumes.

Optical coherence tomography (OCT) provides micron level resolution of retinal tissue and is widely used in ophthalmology. Millions of pre-existing OCT images are available from research and clinical databases. Analysis of this data often requires or can benefit significantly from image registration and reduction of speckle noise. One method of reducing noise is to align and average multiple OCT scans together. We propose to use surface feature information and whole volume information to create a novel and simple pipeline that can rigidly align, and average multiple previously acquired 3D OCT volumes from a commercially available OCT device. This pipeline significantly improves both image quality and visualization of clinically relevant image features over single, unaligned volumes from the commercial scanner.

Speckle noise reduction in US images: Enhanced filtering with optimal filter coefficients

Ultrasound (US) imaging is one of the prominent techniques, which is widely used for its advantages such as affordability, non-radiation, and non-invasiveness. However, the visual evaluation in this imaging model was hampered by speckle noise. De-specking is therefore seen as the most important task. The preservation of every minute detail in the US images is the main challenge of de-speckling. This paper proposes a new de-speckling model that includes two major phases: the preprocessing stage and the noise reduction phase. Preprocessing is carried out by the anisotropic filtering process. In the noise reduction phase, two levels of decomposition take place. In the first level of decomposition, Discrete Waveform Transform (DWT) is used. The next level of decomposition (2-level) intakes the output from the first level, the sub-bands. The 2-level uses improved Weighted Guided Image Filtering (WGIF), and Gradient Domain Guided Image Filtering (GDGIF) for surpassing the noise. To make the process even more specific on efficient de-speckling, the concept of optimization is incorporated and eventually tunes the filter coefficients by considering Structural Similarity Index (SSIM) and Peak Signal to Noise Ratio (PSNR). This optimal tuning process is carried out by a new Shark Solution Integrated Coot Optimization Algorithm (SSI-COA). Finally, the noise-free image is produced through the inverse DWT. The efficiency of SSI-COA is evaluated over the conventional methods under various measures. The cost values using SSI-COA are less than (around 0.0518) when compared to other models. Moreover, a PSNR of 0.053275 is attained using SSI-COA for best-case scenarios.

A Multi-Scale Fusion and Transformer Based Registration Guided Speckle Noise Reduction for OCT Images.

Optical coherence tomography (OCT) images are inevitably affected by speckle noise because OCT is based on low-coherence interference. Multi-frame averaging is one of the effective methods to reduce speckle noise. Before averaging, the misalignment between images must be calibrated. In this paper, in order to reduce misalignment between images caused during the acquisition, a novel multi-scale fusion and Transformer based (MsFTMorph) method is proposed for deformable retinal OCT image registration. The proposed method captures global connectivity and locality with convolutional vision transformer and also incorporates a multi-resolution fusion strategy for learning the global affine transformation. Comparative experiments with other state-of-the-art registration methods demonstrate that the proposed method achieves higher registration accuracy. Guided by the registration, subsequent multi-frame averaging shows better results in speckle noise reduction. The noise is suppressed while the edges can be preserved. In addition, our proposed method has strong cross-domain generalization, which can be directly applied to images acquired by different scanners with different modes.

Semi-Supervised Speckle Noise Reduction in OCT Images With UNet and Swin-Uformer

Semi-Supervised Speckle Noise Reduction in OCT Images With UNet and Swin-Uformer

Evaluation of Speckle Noise Reduction Filters and Machine Learning Algorithms for Ultrasound Images

Evaluation of Speckle Noise Reduction Filters and Machine Learning Algorithms for Ultrasound Images

Comparative Evaluation of Filters for Speckle Noise Reduction in a Clinical Liver Ultrasound Image

Comparative Evaluation of Filters for Speckle Noise Reduction in a Clinical Liver Ultrasound Image

RepDNet: A re-parameterization despeckling network for autonomous underwater side-scan sonar imaging with prior-knowledge customized convolution

Side-scan sonar (SSS) is now a prevalent instrument for large-scale seafloor topography measurements, deployable on an autonomous underwater vehicle (AUV) to execute fully automated underwater acoustic scanning imaging along a predetermined trajectory. However, SSS images often suffer from speckle noise caused by mutual interference between echoes, and limited AUV computational resources further hinder noise suppression. Existing approaches for SSS image processing and speckle noise reduction rely heavily on complex network structures and fail to combine the benefits of deep learning and domain knowledge. To address the problem, RepDNet, a novel and effective despeckling convolutional neural network is proposed. RepDNet introduces two re-parameterized blocks: the Pixel Smoothing Block (PSB) and Edge Enhancement Block (EEB), preserving edge information while attenuating speckle noise. During training, PSB and EEB manifest as double-layered multi-branch structures, integrating first-order and second-order derivatives and smoothing functions. During inference, the branches are re-parameterized into a 3 × 3 convolution, enabling efficient inference without sacrificing accuracy. RepDNet comprises three computational operations: 3 × 3 convolution, element-wise summation and Rectified Linear Unit activation. Evaluations on benchmark datasets, a real SSS dataset and Data collected at Lake Mulan aestablish RepDNet as a well-balanced network, meeting the AUV computational constraints in terms of performance and latency.

Depolarized Holography with Polarization-Multiplexing Metasurface

The evolution of computer-generated holography (CGH) algorithms has prompted significant improvements in the performances of holographic displays. Nonetheless, they start to encounter a limited degree of freedom in CGH optimization and physical constraints stemming from the coherent nature of holograms. To surpass the physical limitations, we consider polarization as a new degree of freedom by utilizing a novel optical platform called metasurface. Polarization-multiplexing metasurfaces enable incoherent-like behavior in holographic displays due to the mutual incoherence of orthogonal polarization states. We leverage this unique characteristic of a metasurface by integrating it into a holographic display and exploiting polarization diversity to bring an additional degree of freedom for CGH algorithms. To minimize the speckle noise while maximizing the image quality, we devise a fully differentiable optimization pipeline by taking into account the metasurface proxy model, thereby jointly optimizing spatial light modulator phase patterns and geometric parameters of metasurface nanostructures. We evaluate the metasurface-enabled depolarized holography through simulations and experiments, demonstrating its ability to reduce speckle noise and enhance image quality.

Speckle-Reduced Optical Coherence Tomography Using a Tunable Quasi-Supercontinuum Source

Optical coherence tomography (OCT), which has long been used for fine-scale structure imaging with higher resolution, larger penetration depth, and more detailed information, is a fast-growing technique for biological tissue imaging. However, speckle is an inherent property in OCT, appearing as bright and dark granular patterns, and hinders the visibility of the fine-scale structure. For the first time, we demonstrated speckle-reduced high-resolution imaging using a tunable quasi-supercontinuum (SC) source. OCT images with uncorrelated speckle patterns could be obtained by several quasi-SC spectra and compounded to improve the image quality. We confirmed that the implementation of a quasi-SC source enables us to reduce speckle noise for better observation of fine-scale structure.

An underwater image enhancement by reducing speckle noise using modified anisotropic diffusion filter

Underwater images are usually suffering from the issues of quality degradation, such as low contrast due to blurring details, color deviations, non-uniform lighting, and noise. Since last few decades, many researches are undergoing for restoration and enhancement for degraded underwater images. In this paper, we proposed a novel algorithm using modified anisotropic diffusion filter with dynamic color balancing strategy. This proposed algorithm performs based on an employing effective noise reduction as well as edge preserving technique with dynamic color correction to make uniform lighting and minimize the speckle noise. Furthermore, reanalyze the contributions and limitations of existing underwater image restoration and enhancement methods. Finally, in this research provided the detailed objective evaluations and compared with the various underwater scenarios for above said challenges also made subjective studies, which shows that our proposed method will improve the quality of the image and significantly enhanced the image.

Multiplane HiLo microscopy with speckle illumination and non-local means denoising.

HiLo microscopy synthesizes an optically sectioned image from two images, one obtained with uniform and another with patterned illumination, such as laser speckle. Speckle-based HiLo has the advantage of being robust to aberrations but is susceptible to residual speckle noise that is difficult to control. We present a computational method to reduce this residual noise without undermining resolution. In addition, we improve the versatility of HiLo microscopy by enabling simultaneous multiplane imaging (here nine planes). Our goal is to perform fast, high-contrast, multiplane imaging with a conventional camera-based fluorescence microscope. Multiplane HiLo imaging is achieved with the use of a single camera and z-splitter prism. Speckle noise reduction is based on the application of a non-local means (NLM) denoising method to perform ensemble averaging of speckle grains. We demonstrate the capabilities of multiplane HiLo with NLM denoising both with synthesized data and by imaging cardiac and brain activity in zebrafish larvae at 40Hz frame rates. Multiplane HiLo microscopy aided by NLM denoising provides a simple tool for fast optically sectioned volumetric imaging that can be of general utility for fluorescence imaging applications.

A New ground open water detection scheme using Sentinel-1 SAR images

ABSTRACT The detection of groundwater is essential not only for scientific research but also for agricultural purposes. This research aims to improve the accuracy and reliability of detecting ground standing water in cropland during the spring/early summer season in eastern South Dakota, USA, by reducing misclassification between water and vegetation. The study utilizes Sentinel-1 synthetic aperture radar (SAR) data. We selected and surveyed 159 ground sites, comprising 78 water sites and 81 nonwater sites, located between Brookings and Sioux Falls, SD, USA. The proposed scheme consists of three steps: 1) developing a modified speckle filter to reduce speckle noise while preserving image details, 2) characterizing the data for water and nonwater sites and providing parameter estimation using the Method of log-cumulants (MoLC) with generalized Gamma distribution (GΓD), and 3) applying Markov random field (MRF) for SAR data classification. The developed scheme demonstrates good performance with a site-based overall detection accuracy of 93.7%. In addition, it provides a computationally efficient solution for water detection, which can be applied in various applications such as crop insurance, precision agriculture and drought monitoring.

Quantitative Measurements of Red Blood Cell Indices Using Spectroscopic Differential Phase-Contrast Microscopy.

Red blood cell (RBC) indices serve as clinically important parameters for diagnosing various blood-related diseases. Conventional hematology analyzers provide the highly accurate detection of RBC indices but require large blood volumes (>1 mL), and the results are bulk mean values averaged over a large number of RBCs. Moreover, they do not provide quantitative information related to the morphological and chemical alteration of RBCs at the single-cell level. Recently, quantitative phase imaging (QPI) methods have been introduced as viable detection platforms for RBC indices. However, coherent QPI methods are built on complex optical setups and suffer from coherent speckle noise, which limits their detection accuracy and precision. Here, we present spectroscopic differential phase-contrast (sDPC) microscopy as a platform for measuring RBC indices. sDPC is a computational microscope that produces color-dependent phase images with higher spatial resolution and reduced speckle noise compared to coherent QPIs. Using these spectroscopic phase images and computational algorithms, RBC indices can be extracted with high accuracy. We experimentally demonstrate that sDPC enables the high-accuracy measurement of the mean corpuscular hemoglobin concentration, mean corpuscular volume, mean corpuscular hemoglobin, red cell distribution width, hematocrit, hemoglobin concentration, and RBC count with errors smaller than 7% as compared to a clinical hematology analyzer based on flow cytometry (XN-2000; Sysmex, Kobe, Japan). We further validate the clinical utility of the sDPC method by measuring and comparing the RBC indices of the control and anemic groups against those obtained using the clinical hematology analyzer.

Integrating IoT and Novel Approaches to Enhance Electromagnetic Image Quality using Modern Anisotropic Diffusion and Speckle Noise Reduction Techniques

Electromagnetic imaging is becoming more important in many sectors, and this requires high-quality pictures for reliable analysis. This study makes use of the complementary relationship between IoT and current image processing methods to improve the quality of electromagnetic images. The research presents a new framework for connecting Internet of Things sensors to imaging equipment, allowing for instantaneous input and adjustment. At the same time, the suggested system makes use of sophisticated anisotropic diffusion algorithms to bring out key details and hide noise in electromagnetic pictures. In addition, a cutting-edge technique for reducing speckle noise is used to combat this persistent issue in electromagnetic imaging. The effectiveness of the suggested system was determined via a comparison to standard imaging techniques. There was a noticeable improvement in visual sharpness, contrast, and overall clarity without any loss of information, as shown by the results. Incorporating IoT sensors also facilitated faster calibration and real-time modifications, which opened up new possibilities for use in contexts with a high degree of variation. In fields where electromagnetic imaging plays a crucial role, such as medicine, remote sensing, and aerospace, the ramifications of this study are far-reaching. Our research demonstrates how the Internet of Things (IoT) and cutting-edge image processing have the potential to dramatically improve the functionality and versatility of electromagnetic imaging systems.

Deep-Learning-Based Multitask Ultrasound Beamforming

In this paper, we present a new method for multitask learning applied to ultrasound beamforming. Beamforming is a critical component in the ultrasound image formation pipeline. Ultrasound images are constructed using sensor readings from multiple transducer elements, with each element typically capturing multiple acquisitions per frame. Hence, the beamformer is crucial for framerate performance and overall image quality. Furthermore, post-processing, such as image denoising, is usually applied to the beamformed image to achieve high clarity for diagnosis. This work shows a fully convolutional neural network that can learn different tasks by applying a new weight normalization scheme. We adapt our model to both high frame rate requirements by fitting weight normalization parameters for the sub-sampling task and image denoising by optimizing the normalization parameters for the speckle reduction task. Our model outperforms single-angle delay and sum on pixel-level measures for speckle noise reduction, subsampling, and single-angle reconstruction.

Speckle reduction in digital holography based on cosine similarity and polynomial interpolation

Speckle noise caused by coherent light often compromises the quality of digital holography reconstructions. To address the limitations of our previously proposed denoising method based on cosine similarity, we introduce a new technique in this paper to further enhance the image denoising performance. Our approach uses image information in four directions surrounding the target pixel and applies appropriate polynomial interpolation to accurately fit the pixel. Experimental results, including both qualitative and quantitative analyses, show that this approach can effectively reduce speckle noise while preserving the original image details. Across various metrics, this method based on cosine similarity and polynomial interpolation offers clear advantages over general image-processing techniques.

Deep Learning based Enhanced Adaptive Despeckling for Ultrasound Images

Considering the inadequacies observed in traditional medical ultrasound image de-speckling techniques, this research presents an innovative solution: a feedforward convolutional neural network (CNN) model coupled with an adaptive multi-exposure fusion framework. The study initiates by curating a specialized ultrasound image training dataset.It then proposes a multi-exposure fusion framework with adaptive enhancement factors to improve image quality for more efficient feature extraction. The proposed method trains a speckle model through the neural network and achieves the extraction of a de-speckled image. Experimental results unequivocally demonstrate the method's unparalleled effectiveness in speckle noise reduction within medical ultrasound images while concurrently preserving intricate image details, thus exemplifying its potential in clinical applications.

An effective adaptive fuzzy filter for speckle noise reduction

An effective adaptive fuzzy filter for speckle noise reduction

Enhanced Speckle Noise Reduction in Breast Cancer Ultrasound Imagery Using a Hybrid Deep Learning Model

Enhanced Speckle Noise Reduction in Breast Cancer Ultrasound Imagery Using a Hybrid Deep Learning Model