Search

International audience; Super-resolution combines several low resolution images having different sampling into a high resolution image. L1-norm data fit minimization has been proposed to solve this problem in a robust way. The outlier rejection capability of this methods has been shown experimentally for super-resolution. However, existing approaches add a regularization term to perform the minimization while it may not be necessary. In this paper, we recall the link between robustness to outliers and the sparse recovery framework. We use a slightly weaker Null Space Property to characterize this capability. Then, we apply these results to super resolution and show both theoretically and experimentally that we can quantify the robustness to outliers with respect to the number of images.

Robust distances are mainly used for the purpose of detecting multivariate outliers. The precise definition of cut-off values for formal outlier testing assumes that the ''good'' part of the data comes from a multivariate normal population. Robust distances also provide valuable information on the units not declared to be outliers and, under mild regularity conditions, they can be used to test the postulated hypothesis of multivariate normality of the uncontaminated data. This approach is not influenced by nasty outliers and thus provides a robust alternative to classical tests for multivariate normality relying on Mahalanobis distances. One major advantage of the suggested procedure is that it takes into account the effect induced by trimming of outliers in several ways. First, it is shown that stochastic trimming is an important ingredient for the purpose of obtaining a reliable estimate of the number of ''good'' observations. Second, trimming must be allowed for in the empirical distribution of the robust distances when comparing them to their nominal distribution. Finally, alternative trimming rules can be exploited by controlling alternative error rates, such as the False Discovery Rate. Numerical evidence based on simulated and real data shows that the proposed method performs well in a variety of situations of practical interest. It is thus a valuable companion to the existing outlier detection tools for the robust analysis of complex multivariate data structures.

The availability of appropriate stimulus material is a key concern for an experimental approach to research on alcohol use disorders (AUDs). A large number of such stimuli are necessary to evoke relevant alcohol-related associations. We report the development of a large stimulus database consisting of 457 pictures of alcoholic beverages and 398 pictures of neutral objects. These stimuli were rated by 18 inpatients hospitalized due to severe AUD and 18 healthy controls along four dimensions: arousal, valence, alcohol-relatedness, and craving. Physical parameters of the pictures were assessed. After outlier removal, 831 stimuli that were characterized as either alcohol-related or neutral were retained in the final stimulus pool. Alcohol-related pictures (versus neutral pictures) evoked higher arousal, more craving and were judged to have higher alcohol-relatedness and a more negative valence. Group comparisons indicated that in patients, neutral pictures evoked more craving and had higher alcohol-relatedness than they did in controls. Physical parameters such as visual complexity, luminance, and color were extracted from these pictures, and extreme values were normalized to minimize mean differences between alcoholic and neutral stimuli. The pictures met the qualitative requirements for (neurophysiological) research. A data file containing rating values and physical parameters will be provided upon request.

Part 3: Pattern Recognition and Image Processing; International audience; In this paper a 2D Leap Frog Algorithm is applied to solve the so-called noisy Photometric Stereo problem. In 3-source Photometric Stereo (noiseless or noisy) an ideal unknown Lambertian surface is illuminated from distant light-source directions (their directions are assumed to be linearly independent). The subsequent goal, given three images is to reconstruct the illuminated object’s shape. Ultimately, in the presence of noise, this problem leads to a highly non-linear optimization task with the corresponding cost function having a large number of independent variables. One method to solve it is 2D Leap Frog Algorithm. During reconstruction, problem that commonly arises, renders the outliers generated in the retrieved shape. In this paper we implement 2D Leap Frog. In particular we focus on choosing snapshot size and on invoking two algorithms that can remove outliers from reconstructed shape. Performance of extended 2D Leap Frog is illustrated by examples chosen especially to demonstrate how this solution is applicable in computer vision. Remarkably, this optimization scheme can also be used for an arbitrary optimization problem depending on large number of variables.

Vision based navigation algorithms estimate the position and attitude of a sensor platform by tracking stationary features in the neighboring environment across multiple image frames captured with an on-board camera. The set of feature matches between two frames is used to compute camera motion using algorithms based on multi-view geometry. The presence of bad feature matches in the data can introduce significant errors in the computed values of the rotation and translation. In this article, we present a robust estimation method to remove the false matched features. Performance of the proposed method is illustrated with numerical examples and compared with conventional Ransac approach. Reduction in computational load is observed for typical datasets without loss of performance making the scheme attractive for real-time implementations.

International Journal of Computer Vision, 95 (1), ISSN:0920-5691, ISSN:1573-1405

Outlier due to false peak error (FPE) usually emerges during displacement estimation of ultrasound elastography, for which the performance of the resulted elastogram is degraded evidently. In this paper, a new method to remove the displacement outliers is proposed. The method is carried out in two steps, i.e., outlier detection and replacement. For the first step, the median of local displacement difference is computed and compared with a threshold deduced according to the application parameters such as the maximum tissue strain, the signal-to-noise ratio of the displacement estimation (SNRe), and the dimensions of the local window. If the median is larger than the threshold, the displacement point is deemed an outlier. For the second step, the displacement outlier is replaced with the median of the displacement points in a local region. The results from the experiments with theoretical data contaminated by artificial outliers show that the proposed method is effective in removing the displacement outliers when the outlier points is not more than a reasonable ratio in all the displacement points. Meanwhile, the result of breast phantom experiment further verifies the validity of this method.

In this paper, a method for the automatic qualitative discrimination of liquid samples based on their absorption spectrum in the ultraviolet, visible and near-infrared regions is presented. An alternative implementation of conventional spectrum matching methodologies is proposed working towards the improvement of the response time of the discrimination system. The method takes advantage of not making assumptions on the probability density function of the data and it is also capable of automatic outlier removal. Preliminary discrimination results have been evaluated on the classification of different oil samples from seeds and olives. The system here proposed could be easily and efficiently implemented in hardware platforms, improving in this way the system performance. This work has been supported by the project TEC2007-67987-C02-01 and the measurement effort of Sergio Diego.

NASA has recently been building 3-dimensional models of the moon based on photos taken from orbiting satellites and the Apollo missions. One issue with the stereo reconstruction is the handling of "outliers", or areas with rapid and unexpected change in the data. Outliers may be introduced by issues such as shadows on the surface, areas with low amounts of surface detail, or flaws in the camera systems. These errors may result in elevation spikes which cause the model to differ significantly from accurate ground truth. We are seeking to remove outliers from reconstructions by using a pair of filters which target the characteristics of these outliers. The first filter will use edge detection to filter areas with low detail and the second filter will remove areas in the disparity map which differ too far from their surrounding neighbors.

In this paper, we address the problem of detecting outlier samples with highly different expression patterns in microarray data. Although outliers are not common, they appear even in widely used benchmark data sets and can negatively affect microarray data analysis. It is important to identify outliers in order to explore underlying experimental or biological problems and remove erroneous data. We propose an outlier detection method based on principal component analysis (PCA) and robust estimation of Mahalanobis distances that is fully automatic. We demonstrate that our outlier detection method identifies biologically significant outliers with high accuracy and that outlier removal improves the prediction accuracy of classifiers. Our outlier detection method is closely related to existing robust PCA methods, so we compare our outlier detection method to a prominent robust PCA method.

In the paper a new measure of distance between events/observations in the pattern space is proposed and experimentally evaluated with the use of k-NN classifier in the context of binary classification problems. The application of the proposed approach visibly improves the results compared to the case of training without postulated enhancements in terms of speed and accuracy. Numerical results are very promising and outperform the reference literature results of k-NN classifiers built with other distance measures.

Many interactivity functions exist in explorative map applications. This chapter provides a typology of these functions with the aim to contribute towards a standardization of the variety of interactive visualization tools. The typology contains 8 types and around 70 particular interactivity functions and is based on an evaluation of existing divisions and categorizations of interactions. The standardization implies a facilitation of both the concept of interfaces and the assessment of realized interfaces. Using the developed structure, two existing applications are evaluated showing the usefulness of the typology.

Techniques based on agglomerative hierarchical clustering constitute one of the most frequent approaches in unsupervised clustering. Some are based on the single linkage methodology, which has been shown to produce good results with sets of clusters of various sizes and shapes. However, the application of this type of algorithms in a wide variety of fields has posed a number of problems, such as the sensitivity to outliers and fluctuations in the density of data points. Additionally, these algorithms do not usually allow for automatic clustering. http://www.sciencedirect.com/science/article/B6TFP-4MYVG3H-1/1/714f82559ca43bbdbff8ad8a1e2d14ac

Copyright © 2015 Inderscience Enterprises Ltd. We propose a Monte Carlo simulation as a service (MCSaaS) which takes the benefits from two sides: the accuracy and reliability of typical Monte Carlo simulations and the fast performance of offering services in the Cloud. In the use of MCSaaS, we propose to remove outliers to enhance the improvement in accuracy. We propose three hypotheses and describe our rationale, architecture and steps involved for validation. We set up three major experiments. We confirm that firstly, MCSaaS with outlier removal reduces percentage of errors to 0.1%. Secondly, MCSaaS with outlier removal is expected to have slower performance than the one without removal but is kept within one second difference. Thirdly, MCSaaS in the Cloud has a significant performance improvement over a popular model on desktop. We demonstrate our approach can meet the demands for accuracy and performance.

We present a practice-oriented, i.e. fast and robust, estimator for strong signal-dependent noise in medical low-dose X-ray images. Structure estimation by median filtering has shown to be superior to linear binomial filtering. Falsifications due to remaining structure in the estimated noise image are significantly reduced by iterative outlier removal.

Biological markers are useful tools for the diagnosis and prognosis of disease. Many different methods are currently used to extract markers from multiple data sources, including gene expression microarrays. This paper investigates the effect of outlier removal on the performance of one such biomarker selection method, Support Vector Machines (SVM). A simple method of outlier removal is employed as a preprocessing step before the data is used for SVM analysis. Both linear and radial basis kernels are used as well as four different normalization techniques. Results show that outlier removal increases the number of highly predictive genes as well as the number of poorly predicting genes. This result thus supports the use of outlier removal prior to biological marker identification via SVM analysis.

Three problems arise in handling numerical values in databases: bad data, missing data, and sloppy data. The effects of bad data are mitigated by using statistical subterfuges such as robust statistics or outlier removal. Missing data are replaced by creating a substitute through interpolation or by using statistics appropriate to unbalanced designs. Sloppy, semiquantitative data are relegated to innocuous positions by using nonparametric, rank, or attribute statistics. These techniques are illustrated by the telephone directory, a database of carcinogenicity test results, and a database of precision parameters derived from method performance (collaborative) studies.

This paper explores the fusion of inertial information with vision for 3D reconstruction. A method is proposed for vertical line segment detection and subsequent local geometric map building. Visual and inertial sensing are two sensory modalities that can be explored to give robust solutions on image segmentation and recovery of 3D structure from images, increasing the capabilities of autonomous vehicles and enlarging the application potential of vision systems. From the inertial sensors, a camera stereo rig, and a few system parameters we can recover the 3D parameters of the ground plane and vertical lines. The homography between stereo images of ground points can be found. By detecting the vertical line segments in each image, and using the homography of ground points for the foot of each segment, the lines can be matched and reconstructed in 3D. The mobile robot then maps the detected vertical line segments in a world map as it moves. To build this map an outlier removal method is implemented and a statistical approach used, so that a simplified metric map can be obtained for robot navigation. http://www.sciencedirect.com/science/article/B6V16-4817F09-5/1/8e3dab0308057a8b8472882aeb69950f

We study the problem of finding an outlier-free subset of a set of points (or a probability distribution) in n-dimensional Euclidean space. A point x is defined to be a β-outlier if there exists some direction w in which its squared distance from the mean along w is greater than β times the average squared distance from the mean along w [1]. Our main theorem is that for any e>0, there exists a (1-e) fraction of the original distribution that has no O(\frac{n}{e}(b+log \frac{n}{e))-outliers, improving on the previous bound of O(n^7b/e). This bound is shown to be nearly the best possible. The theorem is constructive, and results in a \frac{1}{1-e} approximation to the following optimization problem: given a distribution μ (i.e. the ability to sample from it), and a parameter e>0, find the minimum β for which there exists a subset of probability at least (1-e) with no β-outliers.

The objective of this project was to conduct an aflatoxin proficiency test program in government, academia, and industry laboratories in Thailand. Aflatoxin-free corn and peanuts and corn and peanuts naturally contaminated with aflatoxins diluted to approximately 25 μg/kg were analyzed. Homogeneity of prepared, naturally contaminated test samples was checked on multiple replicates. The test was conducted according to the ISO/IUPAC/AOAC INTERNATIONAL Harmonized Protocol with z scores indicating laboratory performance. The participants used 3 methods: enzyme-linked immunosorbent assay, thin-layer chromatography, and the minicolumn. Of 19 laboratories that reported results for aflatoxins in naturally contaminated corn, 13 (68%) performed satisfactorily, on the basis of the mean obtained by an expert laboratory, a calculated target value for standard deviation, and the z score. Of 21 laboratories that reported results for aflatoxins in naturally contaminated peanuts, 10 (48%) performed satisfactorily. For aflatoxin-free corn, 6 laboratories reported finding aflatoxins at ≥10 ng/g, chiefly by the minicolumn method; for aflatoxin-free peanuts, 1 laboratory reported finding aflatoxins at >10 ng/g. Subsequently, a workshop of lectures and laboratory sessions was conducted to improve performance. A new and simple successive outlier removal procedure applied to the same data removed the same laboratories as did the use of z scores.

Large-scale laboratory- and method-performance studies involving more than about 30 laboratories may be evaluated by calculating the HORRAT ratio for each test sample (HORRAT = [experimentally found among-laboratories relative standard deviation] divided by [relative standard deviation calculated from the Horwitz formula]). The chemical analytical method is deemed acceptable per se if HORRAT ͌ 1.0 (± 0.5). If HORRAT is ≳2.0, the most extreme values are removed successively until an "acceptable" ratio is obtained. The laboratories responsible for the extreme values that are removed should examine their technique and procedures. If ≳15% of the values have to be removed, the instructions and the methods should be examined. This suggested computation procedure is simple and does not require statistical outlier tables. Proposed action limits may be adjusted according to experience. Data supporting U.S. Environmental Protection Agency method 245.1 for mercury in waters (manual cold-vapor atomic absorption spectrometry), supplemented by subsequent laboratory-performance data, were reexamined in this manner. Method-performance parameters (means and among-laboratories relative standard deviations) were comparable with results from the original statistical analysis that used a robust biweight procedure for outlier removal. The precision of the current controlled performance is better by a factor of 4 than that of estimates resulting from the original method-performance study, at the expense of rejecting more experimental values as outliers.

A new substrate for the quantitative analysis of salts dissolved in liquids with Laser-induced Breakdown Spectroscopy (LIBS) is introduced for the first time. A steel surface scratched with HB pencil lead is introduced as a very efficient and sensitive substrate for quantitative analysis of dissolved salts in liquids. In this work we demonstrate the analytical quality of this system with the analysis of the crystalline deposits formed by the dried aqueous solutions of salts. We focused on analytical parameters such as sensitivity and linearity for the salt cations in each case. Four salts were studied (Sr(NO3)2, LiSO4, RbCl and BaCl), at nine different concentrations each. To improve linearity and lower the overall error in the calibration curves, we introduce a novel outlier removal method that takes into account the homogeneity of the dry deposits on the analytical surface.

Two lyophilized control sera were distributed through seven national external quality assessment schemes in six European countries—Belgium, Switzerland, France, The Netherlands, Sweden and the United Kingdom—participated in the study. The results for 17 routine analytes were obtained from almost 5000 laboratories for the two sera. The organizers of the schemes were asked to process the results according to a common outlier removal procedure, and submit method-related data if available. The two sera were also distributed through the external/internal scheme of The Netherlands, and the within-laboratory standard deviations calculated in this scheme have been used in a scaling procedure for the external mean values and between-laboratory standard deviations of the participating countries. The results show remarkable agreement in the national mean values for practically all analytes, but considerable differences in the between-laboratory variation. Data from comparable method groups was obtained for 12 analytes from Belgium, France, The Netherlands and the UK. Though revealing some specific differences between methods and countries, the method-related data are generally in agreement with the all-method data. In this study reference method values were only available for cholesterol. The high degree of agreement found suggests, however, that mutual recognition of all-method mean values in national schemes could be acceptable, especially for analytes for which reliable reference methods are not available. The major element of variation is between-laboratory rather than between-country.

Detecting changes in longitudinal data is important in medical research. However, the existence of measurement outliers can cause an unexpected increase in the false alarm rate in claiming changes. To reduce the outliers, a new method has been developed. In this scheme, two measures are initially taken and, if they are closer than a specified threshold, the average of the two is considered to be the estimate of the true mean; otherwise a third measurement is taken, and the mean of the closest pair is considered to be the estimate. It is shown that this method has considerable sample size advantage over naive repeated measurements. Moreover, this scheme is robust for outlier error distribution. Evidence on outlier removal in dental attachment probing is used as an example.

We consider the problem of removing c points from a set S of n points so that the remaining point set is optimal in some sense. Definitions of optimality we consider include having minimum diameter, having minimum area (perimeter) bounding box, having minimum area (perimeter) convex hull. For constant values of c, all our algorithms run in O(nlogn) time.

Separation of ground and nonground measurements is an essential task in the analysis of light detection and ranging (LiDAR) point clouds; however, it is challenge to implement a LiDAR filtering algorithm that integrates the mathematical definition of various landforms. In this letter, we propose a novel LiDAR filtering algorithm that adapts to the irregular structure and 3-D geometry of LiDAR point clouds. We exploit weighted graph representations to analyze the 3-D point cloud on its original domain. Then, we consider airborne LiDAR data as an irregular elevation signal residing on graph vertices. Based on a spectral graph approach, we introduce a new filtering algorithm that distinguishes ground and nonground points in terms of their spectral characteristics. Our complete filtering framework consists of outlier removal, iterative graph signal filtering, and erosion steps. Experimental results indicate that the proposed framework achieves a good accuracy on the scenes with data gaps and classifies the nonground points on bridges and complex shapes satisfactorily, while those are usually not handled well by the state-of-the-art filtering methods.

A novel approach is presented in order to reject correspondence outliers between frames using the parallax-based rigidity constraint for epipolar geometry estimation. In this approach, the invariance of 3-D relative projective structure of a stationary scene over different views is exploited to eliminate outliers, mostly due to independently moving objects of a typical scene. The proposed approach is compared against a well-known RANSAC-based algorithm by the help of a test-bed. The results showed that the speed-up, gained by utilization of the proposed technique as a preprocessing step before RANSAC-based approach, decreases the execution time of the overall outlier rejection, significantly.

Background: Laparoscopy or minimally invasive surgery is a surgical procedure in which laparoscope and other surgical instruments are inserted inside body via a few small incisions. Laparoscope is used to look inside the patient's body and records displayed images. Temporal segmentation of laparoscopic videos has many applications like detecting laparoscopic anomalies and interrupts. It is prerequisite of laparoscopic action recognition for tagging laparoscopic video clips, training to the surgeons and fast retrieval of tagged laparoscopic video clips. Temporal segmentation of videos is is done with the aim of generating homogeneous segments. Â Methods: In this paper, a novel approach for minimally-invasive video segmentation (MIVS) is proposed. In MIVS, several data sets are extracted from laparoscopic videos for increasing the confidence and reducing error of estimation. Each extracted data set is segmented individually with Genetic Algorithm several times after outlier removal. Each time, a different cost function is used as objective function of GA. The correlation coefficient is measured between objective values of individuals of each GA execution and their associated performance measures including detection rate, recognition rate and accuracy. Cost functions having negative correlation with all mentioned performance measures are selected as cost function of the next step segmentation which segments several data sets simultaneously exploiting Multi-objective GA. Results: MIVS is tested on laparoscopic videos of Varicocelle and UPJO surgeries collected from HASHEMINEZHAD Kidney Center. Experimental results show that MIVS can segment laparoscopic videos with accuracy of 94.89%. Conclusions: MIVS outperforms previous presented segmentation methods in segmenting minimally-invasive surgical videos.

Background Metabolite profiles can be used for identifying molecular signatures and mechanisms underlying diseases since they reflect the outcome of complex upstream genomic, transcriptomic, proteomic and environmental events. The scarcity of publicly accessible large scale metabolome datasets related to human disease has been a major obstacle for assessing the potential of metabolites as biomarkers as well as understanding the molecular events underlying disease-related metabolic changes. The availability of metabolite and gene expression profiles for the NCI-60 cell lines offers the possibility of identifying significant metabolome and transcriptome features and discovering unique molecular processes related to different cancer types. Methods We utilized a combination of analytical methods in the R statistical package to evaluate metabolic features associated with cancer cell lines from different tissue origins, identify metabolite-gene correlations and detect outliers cell lines based on metabolome and transcriptome data. Statistical analysis results are integrated with metabolic pathway annotations as well as COSMIC and Tumorscape databases to explore associated molecular mechanisms. Results Our analysis reveals that although the NCI-60 metabolome dataset is quite noisy comparing with microarray-based transcriptome data, it does contain tissue origin specific signatures. We also identified biologically meaningful gene-metabolite associations. Most remarkably, several abnormal gene-metabolite relationships identified by our approach can be directly linked to known gene mutations and copy number variations in the corresponding cell lines. Conclusions Our results suggest that integrative metabolome and transcriptome analysis is a powerful method for understanding molecular machinery underlying various pathophysiological processes. We expect the availability of large scale metabolome data in the coming years will significantly promote the discovery of novel biomarkers, which will in turn improve the understanding of molecular mechanism underlying diseases.