Local Order Statistics Research Articles

Infrared cluster target detection based on Facet kernel and adaptive distance regularization level set

The acquisition of high signal-to-noise ratio (SNR) detection results for targets in infrared images is of great significance for early warning detection and precision guidance. To address the issue of detecting cluster targets in complex backgrounds, a detection method combining the Facet kernel and adaptive distance regularization level set evolution (ADRLSE) is proposed. Firstly, local order-statistics and mean filters are applied to remove high-brightness noise and smooth the image. Subsequently, the Facet kernel filter is applied to enhance the target while suppress the background. Candidate target pixels are obtained through adaptive threshold segmentation. Additionally, based on the edge characteristics of cluster targets, a novel edge detection function is proposed that adaptively determines the parameters using the contour volume (CV) and peak signal-to-noise ratio (PSNR) as references. The enhanced mapping of the target region is obtained by extracting the cluster target contour using ADRLSE. The ADRLSE map is weighted using the target map obtained from the Facet kernel to further enhance the target, and the target is detected through thresholding operations. Finally, the separation result of a single target within a cluster is obtained through a novel adaptive k-means clustering algorithm. Simulation experiments demonstrate that our method has high detection accuracy for cluster targets and can accomplish detection tasks such as cluster contour extraction and single target division.

An optimized non-local LMMSE approach for speckle noise reduction of medical ultrasound images

In this paper, a modified linear-minimum-mean-square-error (LMMSE)-based estimator is presented to reduce speckle noise in ultrasound (US) medical images. In order to significantly improve the performance of the LMMSE estimator, we consider the data redundancy feature which naturally exists in the US images. Since the classical LMMSE method, due to the use of local statistics, cannot perform well in areas where the intensity variation is high, we exploit the similarity between pixels to resolve this problem. In this regards, by using characteristics of the second order local statistics and Pearson distance, an optimum set of similar pixels is selected to be used in the proposed LMMSE-based estimator. Therefore, a good balance between maintaining small details and reducing speckle noise in different regions of US images can be achieved. Quantitative and qualitative results on synthetic and real US data demonstrate that the proposed method yields competitive results in despeckling process compared to the state-of-the-art methods.

Representative Consensus from Limited‐Size Ensembles

AbstractCharacterizing the uncertainty and extracting reliable visual information from ensemble data have been persistent challenges in various disciplines, specifically in simulation sciences. Many ensemble analysis and visualization techniques take a probabilistic approach to this problem with the assumption that the ensemble size is large enough to extract reliable statistical or probabilistic summaries. However, many real‐life ensembles are rather limited in size, with only a handful of members, due to various restrictions such as storage, computational power, or sampling limitations. As a result, probabilistic inference is subject to imprecision and can potentially result in untrustworthy information in the presence of a limited sample‐size ensemble. In this case, a more reliable approach is to fuse the information present in an ensemble with a limited number of members with minimal assumptions. In this paper, we propose a technique to construct a representative consensus that is particularly suited for ensembles of a relatively small size. The proposed technique casts the problem as an ordering problem in which at each point in the domain, the ensemble members are ranked based on the local neighborhood. This local approach allows us to provide shape and irregularity sensitivity. The local order statistics will then be fused to construct a global consensus using a Bayesian approach to ensure spatial coherency of the local information. We demonstrate the utility of the proposed technique using a synthetic and two real‐life examples.

An iterative denoising system based on Wiener filtering with application to biomedical images

Biomedical image denoising systems are important for accurate clinical diagnosis. The purpose of this study is to present a simple and effective iterative multistep image denoising system based on Wiener filtering (WF) where the denoised image from one stage is the input to the next stage. The denoising process stops when a particular condition measured by image energy is adaptively achieved. The proposed iterative system is tested on real clinical images and performance is measured by the well-known peak-signal-to-noise-ratio (PSNR) statistic. Experimental results showed that the proposed iterative system outperforms conventional image denoising algorithms; including wavelet packet (WP), fourth order partial differential equation (FOPDE), nonlocal Euclidean means (NLEM), first order local statistics (FOLS), and single Wiener filter used as baseline model. The experimental results demonstrate that the proposed approach can remove noise automatically and effectively while edges and texture characteristics are preserved.

Fast implementation of the Niblack binarization algorithm for microscope image segmentation

A fast way to implement the Niblack binarization algorithm is described. It uses not only the integral image for the local mean values calculation, but also the second order integral image for the local variance calculation. Following the proposed approach the time of segmentation has been significantly reduced providing the possibility of its use in practice. The generalization of integral image representation, called `k-order integral image' could be used for fast calculation of higher order local statistics. An example of algorithm for the segmentation of cells and Chlamydial inclusions on microscope images, containing the steps for color deconvolution and fast adaptive local binarization is presented.

Multiscale inference about a density

We introduce a multiscale test statistic based on local order statistics and spacings that provides simultaneous confidence statements for the existence and location of local increases and decreases of a density or a failure rate. The procedure provides guaranteed finite-sample significance levels, is easy to implement and possesses certain asymptotic optimality and adaptivity properties.

Three types of distributed CFAR detection based on local test statistic

Three types of distributed constant false-alarm rate (CFAR) detection based on local test statistic (LTS) are developed under the condition that all the samples in different sensors are independent and identically distributed (IID). They are called R (ratio), S (substract) and P (plus) types. For a two-sensor network and local order statistics CFAR (OS-CFAR) algorithm, we derive their closed-form expressions for false alarm and detection probabilities of detecting a Swerling II target in Gaussian noise. The numerical results show that /spl rho/, the ratio between the noise power levels in two local sensors, has very strong effects on the performances of the S and P type schemes and the centralized detection. They can't hold to be CFAR when there isn't the precise prior knowledge of /spl rho/. Compared with the centralized detection, the distributed detection scheme is more robust for the effect of /spl rho/. Fortunately, the R type scheme is not affected by the /spl rho/ at all. It implies that the local CFAR processing can immunize the distributed detection against the /spl rho/.

An adaptive order-statistics filter for SAR images

Abstract An algorithm based on local order statistics is proposed for adaptive reduction of speckle noise in synthetic aperture radar (SAR) images. A selective smoothing is obtained by replacing a pixel value belonging to either of the tails of the local histogram by its percentile, whose area is adaptively defined by a Gaussian function of the Local Variation Coefficient. The filter can fit the actual noise level and preserves structures, textures, and point targets, as well as the local mean without introducing any blur on the edges, mostly due to its closure property. Comparisons with algorithms suitable for speckle smoothing are performed on true SAR images and show selective signal-to-noise ratio (SNR) enhancements.

Texture classification by local rank correlation

A new approach to texture classification based on local rank correlation is proposed here. Its performance is compared with Laws' method which uses local convolution with feature masks. In the experiments, texture samples are classified based on their distribution of local statistics, either rank correlations or convolutions. The new method achieves generally optimal classification rates. It appears to be more robust because local order statistics are unaffected by local sample differences due to monotonic shifts of texture gray values and are less sensitive to noise.