A universal denoising framework with a new impulse detector and nonlocal means

Impulse noise detection is a critical issue when removing impulse noise and impulse/Gaussian mixed noise. In this paper, we propose a new detection mechanism for universal noise and a universal noise-filtering framework based on the nonlocal means (NL-means). The operation is carried out in two stages, i.e., detection followed by filtering. For detection, first, we propose the robust outlyingness ratio (ROR) for measuring how impulselike each pixel is, and then all the pixels are divided into four clusters according to the ROR values. Second, different decision rules are used to detect the impulse noise based on the absolute deviation to the median in each cluster. In order to make the detection results more accurate and more robust, the from-coarse-to-fine strategy and the iterative framework are used. In addition, the detection procedure consists of two stages, i.e., the coarse and fine detection stages. For filtering, the NL-means are extended to the impulse noise by introducing a reference image. Then, a universal denoising framework is proposed by combining the new detection mechanism with the NL-means (ROR-NLM). Finally, extensive simulation results show that the proposed noise detector is superior to most existing detectors, and the ROR-NLM produces excellent results and outperforms most existing filters for different noise models. Unlike most of the other impulse noise filters, the proposed ROR-NLM also achieves high peak signal-to-noise ratio and great image quality by efficiently removing impulse/Gaussian mixed noise.     

An effective 2-stage method for removing impulse noise in images

In this paper, a robust 2-stage impulse noise removal system is proposed to remove impulse noise from extremely corrupted images. The contributions are in two-fold. First, a neuro-fuzzy based impulse noise detector (NFIDET) is introduced to identify the noisy pixels. NFIDET is a powerful noise detector that can handle image corruption even up to 90% with zero miss and false detection rate with a simple neuro-fuzzy structure. This is the best result among the other impulse noise detectors in the literature. Second, this paper presents a new approach for weight calculation of adaptive weighted mean filter by using robust statistical model. An adaptive robust weighted mean (ARWM) filter removes a detected noisy pixel by adaptively determining filtering window size and replacing a noisy pixel with the weighted mean of the noise-free pixels in its window. A Geman–McClure robust estimation function is used to estimate the weights of the pixels. Simulation results also show that the proposed robust filter substantially outperforms many other existing algorithms in terms of image restoration.     

A Universal Impulse Noise Filter with an Impulse Detector and Nonlocal Means

In this paper, we propose an efficient filter for universal impulse noise removal. Operation is carried out in two stages: impulse detection followed by filtering. For detection, a robust local image statistic, called the extremum compression rank-order absolute difference (ECROAD), is designed to detect impulse noise in an image. For filtering, a universal impulse noise filter is proposed by combining the ECROAD statistic with the nonlocal means (NLM). The inherited switching behavior will preserve image details by selecting possible “noise pixels” for processing. Meanwhile, the joint impulsive weight is able to avoid the effect of impulsive components in restoring candidates. Simulation results show that the proposed filter produces excellent results and outperforms most existing filters for different impulse noise models.     

图像脉冲噪声的模糊检测与消除

论述了基于模糊规则的脉冲噪声滤波器。该滤波器由模糊脉冲噪声检测器、噪声消除器与模糊结合器构成。模糊脉冲噪声检测器用窗口风的中值与邻近像素信息来检测脉冲噪声,而脉冲消除器用最小值算法来计算噪声像素的估计值。与传统的脉冲噪声滤波器相比较,所设计的新滤波器具有良好的脉冲噪声抑制与图像细节边缘保护的性能。     

Switching degenerate diffusion PDE filter based on impulselike probability for universal impulse noise removal

In this paper, a switching degenerate diffusion partial differential equation filter (SDDPDE) is developed by introducing the switching operators for reducing all kinds of impulse noise, and especially for images having a mixture of salt-and-pepper impulse noise and random-valued impulse noise which is a shortage for most of the existing filtering models. Our SDDPDE consists of the coarse and fine filtering stages. In the coarse filtering stages, the switching operator depends on a simple noise detector. In the fine filtering stages, we introduce the notion of impulselike probability, and the switching operator depends on both a simple noise detector and impulselike probability. Our SDDPDE will denoise noise pixels detected by the coarse detector while further modify the so-called noise-free pixels according to impulselike probability. The main advantages of our SDDPDE over published approaches are its simplicity and universality. In addition, we demonstrate the performance of our SDDPDE via application to three standard test images, corrupted by salt-and-pepper impulse noise, random-valued impulse noise and mixed impulse noise with high-noise levels, and the comparison with the other well-known filters. Experimental results show that our SDDPDE achieves high peak signal-to-noise ratio and better visual effect.     

A switching median filter with boundary discriminative noise detection for extremely corrupted images.

A novel switching median filter incorporating with a powerful impulse noise detection method, called the boundary discriminative noise detection (BDND), is proposed in this paper for effectively denoising extremely corrupted images. To determine whether the current pixel is corrupted, the proposed BDND algorithm first classifies the pixels of a localized window, centering on the current pixel, into three groups--lower intensity impulse noise, uncorrupted pixels, and higher intensity impulse noise. The center pixel will then be considered as "uncorrupted," provided that it belongs to the "uncorrupted" pixel group, or "corrupted." For that, two boundaries that discriminate these three groups require to be accurately determined for yielding a very high noise detection accuracy--in our case, achieving zero miss-detection rate while maintaining a fairly low false-alarm rate, even up to 70% noise corruption. Four noise models are considered for performance evaluation. Extensive simulation results conducted on both monochrome and color images under a wide range (from 10% to 90%) of noise corruption clearly show that our proposed switching median filter substantially outperforms all existing median-based filters, in terms of suppressing impulse noise while preserving image details, and yet, the proposed BDND is algorithmically simple, suitable for real-time implementation and application.     

Adaptive median filters: new algorithms and results

Based on two types of image models corrupted by impulse noise, we propose two new algorithms for adaptive median filters. They have variable window size for removal of impulses while preserving sharpness. The first one, called the ranked-order based adaptive median filter (RAMF), is based on a test for the presence of impulses in the center pixel itself followed by a test for the presence of residual impulses in the median filter output. The second one, called the impulse size based adaptive median filter (SAMF), is based on the detection of the size of the impulse noise. It is shown that the RAMF is superior to the nonlinear mean L(p) filter in removing positive and negative impulses while simultaneously preserving sharpness; the SAMF is superior to Lin's (1988) adaptive scheme because it is simpler with better performance in removing the high density impulsive noise as well as nonimpulsive noise and in preserving the fine details. Simulations on standard images confirm that these algorithms are superior to standard median filters.     

Cognition and removal of impulse noise with uncertainty

Uncertainties are the major inherent feature of impulse noise. This fact makes image denoising a difficult task. Understanding the uncertainties can improve the performance of image denoising. This paper presents a novel adaptive detail-preserving filter based on the cloud model (CM) to remove impulse noise. It is called the CM filter. First, an uncertainty-based detector identifies the pixels corrupted by impulse noise. Then, a weighted fuzzy mean filter is applied to remove the noise candidates. The experimental results show that, compared with the traditional switching filters, the CM filter makes a great improvement in image denoising. Even at a noise level as high as 95%, the CM filter still can restore the image with good detail preservation.     

Feedback-based object detection for multi-person pose estimation

In this paper, a novel method is proposed for increasing the performance through coupling of top-down models adjusting the object detector based on a new loss function. Generally, object detectors and keypoint estimators are sequentially used in real-time multi-person pose estimations; however, these two models are separately trained. Therefore, the results of the object detector are not optimized for the keypoint estimator. To solve this problem, we analyze the relationship between the two models and propose a feedback-based loss optimization in the object detector, based on the estimation results of the keypoint estimator. In addition, the resulting bounding box of the object detector is readjusted to improve the accuracy of the keypoint estimation model. The experimental results demonstrate that the proposed approach can perform real-time operations with a high frame rate similar to that of the baseline model. Moreover, it achieved an accuracy of 74.2 average precision (AP), which is higher than the state-of-the-arts model including the human detector used in the experiment.     

Difference-type noise detector for adaptive median filter

Noise detection-based median filters have been widely applied to impulse noise reduction. However, the number of pixels misclassified is obviously increased in high noise density. To overcome such drawback, a difference-type noise detector is proposed. In image filtering, a noise detection-based adaptive median algorithm is presented. Experimental results show that the proposed filter can well remove the impulse noise and preserve more details of original images.     

Thresholding noise-free ordered mean filter based on Dempster-Shafer theory for image restoration

This work proposes a new decision-based filter, the thresholding noise-free ordered mean (TNOM) filter based on the Dempster-Shafer (D-S) evidence theory, to preserve more details of images than can other decision-based filters, while effectively suppressing impulse noise. The new filter mechanism is composed of an efficient D-S impulse detector and a noise filter that works by estimating the central noise-free ordered mean (CNOM) value. The D-S evidence theory provides a way to deal with the uncertainty in the evidence and information fusion. Pieces of evidence are extracted, and the mass functions defined using the local information in the filter window. Then, the decision rule is applied to determine whether noise exists, according to the final combined belief value. If a pixel is detected to be a corrupted pixel, then the proposed filter will be triggered to replace it. Otherwise, the pixel is kept unchanged. With respect to the noise suppression of noise on both fixed-valued and random-valued impulses without smearing the fine details in the image, extensive simulation results reveal that the proposed scheme significantly outperforms other decision-based filters.     

双迭代等距均值滤波的医学图像恢复北大核心CSCD

医学图像中往往有很多与脉冲噪声灰度相同的像素,因此含脉冲噪声的医学图像的恢复非常困难。为了获得比现有的脉冲噪声滤波器更好的噪声抑制和纹理结构保持效果,提出了一种双迭代等距均值滤波(dual iterative equidistant mean filter,DIEMF)的医学图像恢复方法。该方法采用等距离邻域进行噪声检测和去除;噪声检测器循环地利用邻域的非最值像素与中心像素之间的平均绝对差,以及利用多数原则,将噪声像素与无噪像素区分开来;噪声去除采用自适应和双迭代的方法,以等距邻域中无噪像素和先前恢复像素的平均值作为中心噪声像素的灰度估计值,充分利用最近的先前恢复的像素。实验结果表明,该方法在噪声抑制和纹理结构保持方面优于现有的方法,特别是对于低密度噪声,它比现有的滤波器具有显著的优越性。     

Multichannel image processing by using the Rank M-type L-filter

In this paper, we introduce the Vector Rank M-type L (VRML)-filter to remove impulsive noise from color images and video sequences. The proposed filter uses the Median M-type (MM) and Ansari-Bradley-Siegel-Tukey M-type (AM) estimators into L-filter to provide robustness to proposed filtering scheme. We also introduce the use of impulsive noise detectors to improve the properties of noise suppression and detail preservation in the proposed filtering scheme in the case of low and high densities of impulsive noise. Simulation results indicate that the proposed filter consistently outperforms other color image filters by balancing the trade-off between noise suppression, detail preservation, and color retention.     

一种新的图像去噪混合滤波方法

为了去除图像中混入的脉冲噪声和高斯噪声,提出了一种基于自适应中值滤波和模糊加权均值滤波的混合滤波方法.该方法首先进行噪声检测把受高斯型噪声污染的像素和受脉冲型噪声污染的像素区别开来,然后对受高斯噪声污染的像素采用模糊加权均值滤波算法,而对受脉冲噪声污染的像素则采用改进的中值滤波算法进行去噪.仿真结果证明,该方法更具有实用性和有效性.     

Adaptive detection in asynchronous code-division multiple-accesssystem in multipath fading channels

In code-division multiple-access systems transmitting data over time-varying multipath channels, both intersymbol interference (ISI) and multiple-access interference (MAI) arise. In this paper, we address interference suppression, multipath diversity and processing gain protection for multiuser detection with less noise enhancement by using a parallel cancelling scheme. The proposed detector consists of a RAKE filter, forward filter, and feedback filter with different functions for each filter. The RAKE filter increases the signal-to-noise ratio by taking the advantage of multipath and code diversities. The forward filter is proposed, in combination with the feedback filter, to remove the effects of MAI and ISI by parallel cancellation. In order to avoid performance deterioration due to unreliable initial estimation in the parallel cancellation, a cost function with proper weighting is introduced to improve the performance of the proposed detector. In the proposed design method, a recursive least square algorithm is employed to update the tap-coefficients of all filters for MAI and ISI cancellation. Finally, the performance of the proposed detector is analyzed and compared with other detectors     

Perception of temporally filtered X-ray fluoroscopy images

For noisy X-ray fluoroscopy image sequences we quantitatively evaluated image quality after digital temporal filtering to reduce noise. Using an experimental paradigm called a reference/test adaptive forced-choice method we compared detectability of stationary low-contrast disks in filtered and unfiltered, computer-generated image sequences. In the first experiment, a low-pass first-order recursive filter used in X-ray fluoroscopy was found to be much less effective at enhancing detectability than predicted from the reduction of display noise variance, a common measurement of filter effectiveness. Detectability was reasonably predicted by a nonprewhitening human-observer model (NPW-HVS) that included an independently determined human temporal-contrast-sensitivity function. In another experiment, designed to test models over a range of temporal frequencies, we used paired high-pass and low-pass temporal filters that both reduced noise variance by 25%. The high-pass filter was artificially applied to the noise only and greatly improved detectability, while the low-pass filter had little effect. The human-observer model quantitatively described the measurements, but classical prewhitening and nonprewhitening signal detectors did not. As compared to the nonprewhitening, spatio-temporal matched filter, human-observer efficiency was low and variable at 2.1%, 2.9%, and 0.06% for 60 frames of unfiltered low-pass and high-pass noise, respectively. As compared to this detector, humans were not very effective at combining information across frames. On the other hand, signal to noise ratios (SNR's) from the human-observer model were comparable to human performance, and efficiencies were reasonably constant at 40%, 52%, and 32%, respectively. We conclude that it is imperative to include human-observer models and experiments in the analysis of noise-reduction filtering of noisy image sequences, such as X-ray fluoroscopy.     

A simple neuro-fuzzy method for improving the performances of impulse noise filters for digital images

A novel method for improving the performances of impulse noise filters is presented. The method enhances the performance of an impulse noise filter in two ways: increases its noise-suppression ability and decreases its distortion effects. The method is based on a simple 2-input 1-output neuro-fuzzy system. The internal parameters of the system are tuned by training. Training of the system is easily accomplished by using a simple computer-generated artificial image. The proposed method can easily be used with any impulse noise removal operator. The application of the method is completely independent of the noise removal operator and it has no influence on the filtering behavior of the operator. Experimental results show that the proposed method may efficiently be used with any type of impulse noise removal operator to significantly improve its filtering performance.     

基于直方图的自适应图像去噪滤波器

对于那些明显偏离高斯型白噪声的加性噪声,如拖尾脉冲噪声,高斯脉冲噪声等,已有方法的滤噪性能会严重退化.为此,该文提出了一种去除脉冲噪声的新方法.该方法首先由被污染图像估计出原图像的直方图.然后应用模糊集理论,利用加权策略得到了一个符合图像灰度分布统计规律的模糊隶属度函数,以此隶属度函数构建一个加权平均滤波器. 新方法有效地利用了原图像的先验知识,能够根据图像区域特性差异及脉冲噪声强弱自适应地采用不同的滤波尺度.文章比较了传统滤波器、已有的模糊滤波器和本文方法的结果.实验表明本文方法具有更好的效果.     

Nonparametric multiuser detection in non-Gaussian channels

Existing multiuser detection techniques in wireless systems are based on the assumption that some information on the parameters of the probability density function (pdf) of ambient noise is available. Such information may not be available in all cases, particularly for non-Gaussian and impulsive noises, or may change depending on circumstances. In this paper, we present a technique for multiuser detection that does not require any a priori knowledge about the noise parameters. This method is based on using pseudo norms for linear nonparametric regression. Analytical and simulation results show that the proposed method offers an improved, or at least comparable, performance over existing robust techniques in the absence of any information on the nature of noise in the environment. The increased computational complexity is marginal compared to existing parametric detectors. In addition, the proposed nonparametric detector is portable in the sense that it does not need to be tuned for different noise models without any considerable degradation of performance. We also show that in non-Gaussian noise, the performance of blind adaptive nonparametric multiuser detectors is better than that of robust multiuser detectors.     

Generalized feature extraction using expansion matching

A novel generalized feature extraction method based on the expansion matching (EXM) method and on the Karhunen-Loeve transform (KLT) is presented. The method provides an efficient way to locate complex features of interest like corners and junctions with reduced number of filtering operations. The EXM method is used to design optimal detectors for a set of model elementary features. The KL representation of these model EXM detectors is used to filter the image and detect candidate interest points from the energy peaks of the eigen coefficients. The KL coefficients at these candidate points are then used to efficiently reconstruct the response and differentiate real junctions and corners from arbitrary features in the image. The method is robust to additive noise and is able to successfully extract, classify, and find the myriad compositions of corner and junction features formed by combinations of two or more edges or lines. This method differs from previous works in several aspects. First, it treats the features not as distinct entities, but as combinations of elementary features. Second, it employs an optimal set of elementary feature detectors based on the EM approach. Third, the method incorporates a significant reduction in computational complexity by representing a large set of EXM filters by a relatively small number of eigen filters derived by the KL transform of the basic EXM filter set. This is a novel application of the KL transform, which is usually employed to represent signals and not impulse responses as in our present work.