Spatial adaptive wavelet thresholding for image denoising

The method of wavelet thresholding for removing noise, or denoising, has been researched extensively due to its effectiveness and simplicity. Much of the work has been concentrated on finding the best uniform threshold or best basis. However, not much has been done to make this method adaptive to spatially changing statistics which is typical of a large class of images. This work proposes a spatially adaptive wavelet thresholding method based on context modeling, a common technique used in image compression to adapt the coder to the non-stationarity of images. We model each coefficient as a random variable with the generalized Gaussian prior with unknown parameters. Context modeling is used to estimate the parameters for each coefficient, which are then used to adapt the thresholding strategy. Experimental results show that spatially adaptive wavelet thresholding yields significantly superior image quality and lower MSE than optimal uniform thresholding.

Finding gradient maximum and local maximum based on multi-scale Canny operator edge detection is really the optimization for two dimension multi-element function.The maximal function gradient or local maximum which was derived by the traditional analytics is approximate and local.A new multi-scale edge detection algorithm is proposed by a genetic optimum searching algorithm.To upgrade the genetic algorithm convergence about edge detection,an improved GA+SA+TABU is used in order to overcome the defects of local searching in the general genetic algorithm and upgrade the whole resolution.Alternating optimization tactics are utilized by combining the general algorithm and heuristic searching methods.The experimental results show that the proposed algorithm applied to IR target image segmentation could result in copious details,single edge and exact location.

The denoising of medical images degraded by various types of noises is a major problem in medical image processing. A very good tradeoff between detail preservation and noise removal is obtained while implementing with various successive filtering procedures. This paper presents a wavelet-based multiscale products thresholding and spatial based Bilateral Filter scheme for noise removal of magnetic resonance images. In the first stage, the noisy image is passed through Bilateral Filter (BF) and some amount of noise is reduced but the image becomes blurred, hence adaptive wavelet thresholding is applied with multiscale product rule in the second stage. Rather than using Wavelet Transform, Dyadic Wavelet Transform is employed. After apply the Dyadic Wavelet Transform, wavelet interscale dependencies is exploited by multiply the adjacent wavelet subbands to preserve edge structures while suppressing noise. An adaptive threshold is calculated and implemented on the multiscale products rather than on the wavelet coefficients. Hence to identify important features like edges, curves and textures the thresholding is applied to multiscale products. The proposed method for medical image denoising is performed through a recursive application of simple steps and demonstrated on a number of standard medical images. The results obtained with the proposed method outperform other methods both visually and in terms of objective quality peak-signal-to-noise ratio (PSNR).

Wavelet threshold denoising can easily cause edge blur, while total variational denoising is insufficient to achieve image smooth area denoising. Therefore, this paper proposes a new denoising method combining the total variational model and the adaptive wavelet threshold. Firstly, the original noisy image is processed by the total variational model, the improved Prewitt operator is used to extract the edge of the processed image, and the edge denoising image is obtained. Then, the original noisy image is divided into edge and non-edge images by the improved Prewitt operator, and the non-edge image is de-noised by the adaptive wavelet threshold method to obtain the non-edge denoising image. Finally, after denoising, the edge image and non-edge image are superimposed to achieve high-definition restoration of medical images. Experimental results show that the proposed method can effectively denoise medical images and preserve the original edges and details in the images.KeywordsImage denoisingTotal variational modelPrewitt operatorWavelet transform

Medical images are generally suffered from signal dependent noises i.e. speckle noise and broken edges. Most of the noises signals appear from machine and environment generally not contribute to the tissue differentiation. But, the noise generated due to above mentioned reason causes a grainy appearance on the image, hence image enhancement is required. This paper proposed a method based on image de-noising and edge enhancement of noisy multidimensional imaging data sets. For the intent of image denoising, Adaptive Multiscale Product Thresholding based on 2-D wavelet transform is used. In this method, contiguous wavelet sub bands are multiplied to improve edge structure while reducing noise. In multiscale products, boundaries can be successfully distinguished from noise. Adaptive threshold is designed and forced on multiscale products as an alternative of wavelet coefficients or recognize important features. For the edge enhancement, Canny Edge Detection Algorithm is used with scale multiplication technique. The proposed algorithm is implemented using MATLAB. Simulation results for image enhancement were presented both for standard test images and CT scan image. Simulation results shows that the planned technique better suppress the Poisson noise among several noises i.e. salt & pepper, speckle noise and random noise. From the qualitative analysis it is also evident that algorithm preserve edges while enhancement.

Voltage transformers are an important component of power systems. However, transformers work in complex environments such as high noise and strong magnetic interference for a long time, resulting in problems such as multiple types of faults, difficulty in extracting fault characteristics, and low accuracy in fault diagnosis. Based on this, this article proposes a fault diagnosis method based on adaptive local iterative filtering (ALIF)-fuzzy dispersion entropy (FDE)-adaptive wavelet threshold filtering and improved graph neural networks (GNNs). Firstly, ALIF is used to perform modal decomposition on the original signal, and FDE is employed to eliminate noise-dominant modes and extract optimal voltage sample feature vectors. Secondly, an adaptive wavelet threshold function based on FDE features is proposed to suppress residual noise. By dynamically adjusting parameters, it synchronously suppresses excessive high-frequency smoothing and pseudo-Gibbs phenomenon, achieving noise suppression and feature fidelity of the signal. Finally, a multi-layer bridged high-order GNN integrating dynamic attention mechanisms and contrastive normalization is constructed. This network enhances the ability to capture long-range dependencies in voltage signals through hierarchical feature aggregation. Experimental results show that the fault diagnosis algorithm proposed in this paper exhibits high noise robustness and diagnostic accuracy. Compared to fixed threshold functions and traditional GNNs, it improves the signal-to-noise ratio by 6.4%, reduces the mean squared error by 42.5%, increases the diagnostic accuracy by 2.61%, and enhances the F1 score by 3.60%, achieving high-precision fault diagnosis for voltage transformers.

Accurate segmentation of fingernail images is essential for biomedical applications like dermatological diagnostics and nail disease assessments. This study compares traditional methods (Sobel and Canny edge detectors) with an improved method using adaptive thresholding and morphological closing for fingernail image segmentation. The methodology includes data collection, preprocessing, edge detection, segmentation, and evaluation. A dataset of 500 fingernail images (free of nail polish) was captured using a digital camera. Preprocessing involves grayscale conversion to simplify analysis and Gaussian smoothing to reduce noise while preserving key features. For segmentation, thresholding and K-means clustering isolate the fingernail from the background. Evaluation combines qualitative and quantitative analyses. Qualitative results demonstrate that the improved method consistently outperforms traditional techniques under diverse conditions. Quantitative evaluation, based on accuracy, recall, F1 score, and Intersection over Union (IoU), further supports these findings. The Sobel method achieves 0.80 accuracy, 0.77 recall, 0.87 F1 score, and 0.77 IoU. The Canny method achieves 0.82 accuracy, 0.78 recall, 0.88 F1 score, and 0.78 IoU. In contrast, the improved method achieves 0.97 accuracy, 0.98 recall, 0.99 F1 score, and 0.98 IoU. The results clearly show that the improved method, using adaptive thresholding and morphological closing, provides superior segmentation performance. Additionally, the approach remains computationally efficient, making it suitable for real-time applications in medical diagnostics.

Doppler radar for monitoring vital signals is an emerging tool, and how to remove the noise during the detection process and reconstruct the accurate respiration and heartbeat signals are hot issues in current research. In this paper, a novel radar vital signal separation and de-noising technique based on improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN), sample entropy (SampEn), and wavelet threshold is proposed. First, the noisy radar signal was decomposed into a series of intrinsic mode functions (IMFs) using ICEEMDAN. Then, each IMF was analyzed using SampEn to find out the first few IMFs containing noise, and these IMFs were de-noised using the wavelet threshold. Finally, in order to extract accurate vital signals, spectrum analysis and Kullback–Leible (KL) divergence calculations were performed on all IMFs, and appropriate IMFs were selected to reconstruct respiration and heartbeat signals. Moreover, as far as we know, there is almost no previous research on radar vital signal de-noising based on the proposed technique. The effectiveness of the algorithm was verified using simulated and measured experiments. The results show that the proposed algorithm could effectively reduce the noise and was superior to the existing de-noising technologies, which is beneficial for extracting more accurate vital signals.

A modular approach on an adaptive thresholding method for segmentation of cell regions in bioelectric images with complex lighting environments and background conditions is presented in this paper. Preprocessing steps involve low-pass filtering of the image and local contrast enhancement. This image is then adaptively thresholded which produces a binary image. The binary image consists of cell regions and the edges of a metal electrode that show up as bright spots. A local region based approach is used to distinguish between cell regions and the metal electrode tip that cause bright spots. Regional properties such as area are used to separate the cell regions from the non-cell regions. Special emphasis is given on the detection of twins and triplet cells with the help of watershed transformation, which might have been lost if form-factor alone were to be used as the geometrical descriptor to sepa rate the cell and the non-cell regions. Keywords: bioelectric images, image segmentation, adaptive thresholding, cell segmentation, local contrast enhancement, watershed transformation, template matching

Application of new advanced CNN structure with adaptive thresholds to color edge detection

Accuracy and precision of segmentation and quantification of wrist bone microarchitecture using photon-counting computed tomography ex vivo.

Numerous methods have been introduced to segment the thyroid tumors in iodine-124 (124I) Positron Emission Tomography/Computer Tomography (PET/CT) imaging. An adaptive threshold-based is preferable to a manual and fixed threshold-based approach for forecasting accurate and precise tumor volume in real-patient studies, as it was reported to be more resistant to noise and resolution. Manual segmentation was reported to be time-consuming, subjective, subject to inter- and intra-observer variability and prone to segmentation errors [1]. This study aimed to determine the optimal adaptive threshold for thyroid tumor segmentation in 124I PET/CT imaging.
 
 The objective of this study was firstly to determine the segmented volume using the standardized uptake value threshold method. Secondly, to compare the adaptive threshold and recovery coefficient (RC) value between the time of flight (TOF) and non-time of flight (non-TOF) PET/CT reconstructions in 124I phantom studies. Finally, to determine the accuracy and RC for 124I PET/CT studies. This study was done by placing NEMA 2012/IEC 2008 PET IQ Phantom and filled with 1 kBq/ml and 20 kBq/ml in its background and sphere each to model tumor background ratio (TBR) 20:1 of 124I PET/CT imaging. The phantom was scanned using TOF and non-TOF PET at 3, 4 and 5 minutes per bed position and reconstructed using 600, 1200 and 2000 beta penalization factors. Image registration, contouring and segmentation were performed using MIM Encore Software. A set of different adaptive thresholds (Tadaptive: 10%-50%) was applied to each sphere in the TBR image of the phantom using the tri-dimensional automatic segmentation tool. The optimal Tadaptive for each of the parameter assessed was defined based on the lowest goodness value calculated. The RC was calculated using mean value and known activity. Finally, the RC against the percentage of maximum was plotted to analyse the descending order of beta penalization factors for time per bed position.
 
 Figure 1 shows the image of 124I contoured VOIs on the spheres. The results showed that segmented volumes were not affected by the time per bed position in 124I PET/CT imaging. Table 1 shows the descending order beta penalization factors for time per bed position. Beta penalization factors of 600 to 2000 resulted in either underestimation or overestimation of the measured volume. Comparing TOF to non-TOF PET/CT 124I imaging, consistent and reproducible data were observed for the TOF PET for the three beta penalization factors. Tadaptive of 28.0% consistently gave the lowest goodness value shown by the closest R2 value to one, thus can produce lowest error if it is being used to reconstruct images.
 As a conclusion, 28% Tadaptive is suggested as the optimal threshold for 124I TOF PET/CT imaging due to consistent data obtained for all beta penalization factors assessed in this study. The inconsistent data obtained by non-TOF PET thus requires further optimization.

In this paper, we present a novel image denoising method based on adaptive thresholding and k-means clustering. In this method, we adopt the adaptive thresholding technique as an alternative to the traditional hard-thresholding of the block-matching and 3D filtering (BM3D) method. This technique has a high capacity to adapt and change according to the amount of the noise. More precisely, in our method the soft-thresholding is applied to the areas with heavy noise, on the contrary the hard-thresholding is applied to the areas with slight noise. Based on the adaptation and stability of the adaptive thresholding, we can achieve optimal noise reduction and maintain the high spatial frequency detail (e.g. sharp edges). Owing to the capacity of k-means clustering in terms of finding the relevant candidate-blocks, we adopt this clustering at the last estimate to partition the denoised image into several regions and identify the boundaries between these regions. Applying k-means clustering will allow us to force the block matching to search within the region of the reference block, which in turn will lead to minimize the risk of finding poor matching. The main reason of applying the K-means clustering method on the denoised image and not on the noised image is specifically due to the flaw of accuracy in detecting edges in the noisy image. Experimental results demonstrate that the new algorithm consistently outperforms other reference methods in terms of visual quality, Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index (SSIM). Furthermore, in the proposed algorithm the time consumption of the image denoising is less than that in the other reference algorithms.

For the strong noise gear fault vibration signal is relatively weak, and the transmission path is complex and variable, in the case of composite faults, the modulation of different fault characteristics of the frequency, coupling, resulting in the actual acquisition of the fault characteristics are difficult to extract and separate. Aiming at fault feature extraction and separation, an adaptive threshold denoising fault detection method based on Maximum correlated kurtosis deconvolution (MCKD) and Empirical wavelet transform (EWT) is proposed. Firstly, envelope entropy and information entropy are used as fitness functions, and the parameters of the MCKD algorithm are optimized by the improved particle swarm algorithm, then the empirical wavelet decomposition is carried out on the signals, and finally adaptive wavelet threshold denoising is carried out on the decomposed Intrinsic mode functions (IMFs) components. The results of experimental data analysis show that compared with the feature extraction methods such as spatial scale threshold EWT-MCKD and Complete Ensemble Empirical Mode Decomposition (CEEMDAN)-MCKD, the proposed method is more suitable for the diagnosis of gear composite faults in a strong background noise environment, the noise interference is effectively suppressed, and the extraction effect of gear composite fault features is more obvious.

In response to the issues of single-scale information loss and large model parameter size during the sampling process in U-Net and its variants for medical image segmentation, this paper proposes a multi-scale medical image segmentation method based on pixel encoding and spatial attention. Firstly, by redesigning the input strategy of the Transformer structure, a pixel encoding module is introduced to enable the model to extract global semantic information from multi-scale image features, obtaining richer feature information. Additionally, deformable convolutions are incorporated into the Transformer module to accelerate convergence speed and improve module performance. Secondly, a spatial attention module with residual connections is introduced to allow the model to focus on the foreground information of the fused feature maps. Finally, through ablation experiments, the network is lightweighted to enhance segmentation accuracy and accelerate model convergence. The proposed algorithm achieves satisfactory results on the Synapse dataset, an official public dataset for multi-organ segmentation provided by the International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI), with Dice similarity coefficient (DSC) and 95% Hausdorff distance (HD95) scores of 77.65 and 18.34, respectively. The experimental results demonstrate that the proposed algorithm can enhance multi-organ segmentation performance, potentially filling the gap in multi-scale medical image segmentation algorithms, and providing assistance for professional physicians in diagnosis.