Showing papers on "Mahalanobis distance published in 2000"

Methods of constructing core collections by stepwise clustering with three sampling strategies based on the genotypic values of crops.

Abstract: A genetic model with genotype×environment (GE) interactions for controlling systematical errors in the field can be used for predicting genotypic values by an adjusted unbiased prediction (AUP) method. Mahalanobis distance, calculated based on the genotypic values, is then applied to measure the genetic distance among accessions. The unweighted pair-group average, Ward’s and the complete linkage methods of hierarchical clustering combined with three sampling strategies are proposed to construct core collections in a procedure of stepwise clustering. A homogeneous test and t-tests are suggested for use in testing variances and means, respectively. The coincidence rate (CR%) for range and the variable rate (VR%) for the coefficient of variation are designed to evaluate the property of core collections. A worked example of constructing core collections in cotton with 21 traits was conducted. Random sampling can represent the genetic diversity structure of the initial collection. Preferred sampling can keep the accessions with special or valuable characteristics in the initial collection. Deviation sampling can retain the larger genetic variability of the initial collection. For better representation of the core collection, cluster methods should be combined with different sampling strategies. The core collections based on genotypic values retained larger genetic variability and had superior representatives than those based on phenotypic values.

FALCON: Feedback Adaptive Loop for Content-Based Retrieval

Abstract: Several methods currently exist that can perform relatively simple queries driven by relevance feedback on large multimedia databases. However, all these methods work only for vector spaces; that is, they require that objects be represented as vectors within feature spaces. Moreover, their implied query regions are typically convex. This research paper explains our solution. We propose a novel method that is designed to handle disjunctive queries within metric spaces. The user provides weights for positive examples; our system "learns" the implied concept and returns similar objects. Our method differs from existing relevance-feedback methods that base themselves upon Euclidean or Mahalanobis metrics, as it facilitates learning even disjunctive, concave models within vector spaces, as well as arbitrary metric spaces. Our main contributions are two-fold. Not only do we present a novel way to estimate the dissimilarity of an object to a set of desirable objects, but we support it with an algorithm that shows how to exploit metric indexing structures that support range queries to accelerate the search without incurring false dismissals. Our empirical results demonstrate that our method converges rapidly to excellent precision/recall, while outperforming sequential scanning by up to 200%.

Mahalanobis-Taguchi System

Abstract: From the Publisher: Learn how you can harness the power of an amazing new pattern-recognition and forecasting method from Dr. Genichi Taguchi, a world-renowned quality genius. 15 case studies from around the U.S. and Japan show how industry giants used the MTS effectively in their organizations. With this important and authoritative book, you can achieve the same success.

The Mahalanobis-Taguchi System

Abstract: Preface Acknoledgements Chapter 1: Introduction 1.1 Lives Versus Machines 1.2 What Is MTS? 1.3 Where MTS Can Be Applied 1.4 An Important Issue Chapter 2: A Detailed Example of MTS 2.1 Clutch Disc Defects 2.2 Visual Inspection Method 2.3 Differential and Integral Characteristics 2.4 Raw Data Collection 2.5 Normalization 2.6 Matrix Construction 2.7 Mahalanobis Distance 2.8 Results of Recognition 2.9 Selection of Threshold Value 2.10 Selection of Characteristics 2.11 Confirmation Chapter 3: Business-Process Forecasting 3.1 A Problem in Business Systems 3.2 Preparation of the Database 3.3 Calculation of the Database 3.4 Mahalanobis Distance 3.5 Estimation of the Unpaid Amount and its Estimating Error by Using the SN Ratio 3.6 Rationalization of Item Selection 3.7 Design of a Business System Section One -- Health Care Chapter 4: Diagnosis of a Special Health Checking 4.1 Diagnosis of Liver Function 4.2 Selection of Characteristics 4.3 The Loss Function of Health Checkups (Determination of Threshold) Chapter 5: Application for Medical Treatment 5.1 Problems in Clinical Research 5.2 Mahalanobis Distance and Treatment Effect 5.3 The Study Using One Patient 5.4 Comparison Treatment Methods Section Two -- Mechanical Industry Chapter 6: Wafer Yield Prediction 6.1 Objective 6.2 Base Space 6.3 Relatinship Between Mahalanobis Distance and Yield 6.4 Selection of Characteristics Chapter 7: Inkject Quality Inspection 7.1 Introduction 7.2 Camera Inspection System 7.3 Mahalanobis Distance Results 7.4 Measurement System Cost Reduction 7.5 Conclusions Chapter 8: Prevention of Driving Accidents 8.1 Introduction 8.2 Measuring System 8.3 Base Space 8.4 Base Data Collection 8.5 Base Space Construction and Mahalanobis Distance Distribution 8.6 Mahalanaobis Distance Under Dangerous Situations 8.7 Evaluation of Functionality 8.8 Conclusion Section Three -- Electrical Industry Chapter 9: Solder Joint Appearance Inspection 9.1 Introduction 9.2 Data Collection and Mahalanobis Distance Calculation 9.3 Mahalanobis Distance Using Inspection-Logic Characteristics 9.4 Mahalanobis Distance Using Reflection Characteristics Chapter 10: Fire Alarm System Optimization 10.1 Introduction 10.2 Data Collection 10.3 Calculation of Mahalanobis Space 10.4 Calculation of Mahalanobis Distance 10.5 Selection of Sensors Section Four -- Chemical Industry Chapter 11: Diagnosis of Photographic Processing Solution 11.1 Introduction 11.2 Processing of Photo-Sensitive Materials 11.3 Selection of the Base Space 11.4 Mahalanobis Distance of Rejected Solutions and Their Photographic Quality 11.5 Factorial Effects 11.6 Discussions Chapter 12: Pattern Recognition for Infrared Absorption Spectrum Analysis 12.1 Introduction 12.2 Experiment 12.3 Selection of Characteristics 12.4 Results Section Five -- Space Industry Chapter 13: Fault Analysis 13.1 Introduction 13.2 Outline of the Fault Diagnosis Program (FDP) 13.3 Observer and Residual 13.4 Process of Fault Diagnosis by Using Residuals 13.5 Collection of Normal Data 13.6 The Standard Space of a Normal Group 13.7 Fault Detection by Mahalanobis Distance 13.8 Fault Identification by Mahalanobis Distance 13.9 Fault Identification by Estimation Errors Section Six -- Software Industry Chapter 14: Valuation of a Programmer's Capability 14.1 Introduction 14.2 Data Collection 14.3 The MTS Method 14.4 Analysis 14.5 Factor Reductin 14.6 Effective Factors 14.7 Results Chapter 15: Handwriting Recognition 15.1 Introduction 15.2 Extraction of Character Elements 15.3 Procedures of Character Recognition 15.4 Calculation of Mahalanobis Distance 15.5 Hand-Written Character Recognition Section Seven -- Government Chapter 16: U.S. Dollar Bill Inspection 16.1 Patterns of U.S. Dollar Bills 16.2 Characteristics of One-Dollar Bills 16.3 Differentiation of One-Dollar Bills Index

Factorization with Uncertainty

Abstract: Factorization using Singular Value Decomposition (SVD) is often used for recovering 3D shape and motion from feature correspondences across multiple views. SVD is powerful at finding the global solution to the associated least-square-error minimization problem. However, this is the correct error to minimize only when the x and y positional errors in the features are uncorrelated and identically distributed. But this is rarely the case in real data. Uncertainty in feature position depends on the underlying spatial intensity structure in the image, which has strong directionality to it. Hence, the proper measure to minimize is covariance-weighted squared-error (or the Mahalanobis distance). In this paper, we describe a new approach to covariance-weighted factorization, which can factor noisy feature correspondences with high degree of directional uncertainty into structure and motion. Our approach is based on transforming the raw-data into a covariance-weighted data space, where the components of noise in the different directions are uncorrelated and identically distributed. Applying SVD to the transformed data now minimizes a meaningful objective function. We empirically show that our new algorithm gives good results for varying degrees of directional uncertainty. In particular, we show that unlike other SVD-based factorization algorithms, our method does not degrade with increase in directionality of uncertainty, even in the extreme when only normal-flow data is available. It thus provides a unified approach for treating corner-like points together with points along linear structures in the image.

A robust background subtraction method for changing background

Abstract: Background subtraction is a useful and effective method for detecting moving objects in video images. Since this method assumes that image variations are caused only by moving objects (i.e., the background scene is assumed to be stationary), however, its applicability is limited. In this paper, we propose a background subtraction method that robustly handles various changes in the background. The method learns the chronological changes in the observed scene's background in terms of distributions of image vectors. The method operates the subtraction by evaluating the Mahalanobis distances between the averages of such image vectors and newly observed image vectors. The method we propose herein expresses actual changes in the background using a multi-dimensional image vector space. This enables the method to detect objects with the correct sensitivity. We also introduce an eigenspace to reduce the computational cost. We describe herein how approximate Mahalanobis distances are obtained in this eigenspace. In our experiments, we confirmed the proposed method's effectiveness for real world scenes.

Estimating the Mahalanobis distance from mixed continuous and discrete data.

Abstract: We present a method for estimating the Mahalanobis distance between two multivariate normal populations when a subset of the measurements is observed as ordered categorical responses. Asymptotic properties of the proposed estimator are developed. Two examples are discussed.

Texture classification using dominant wavelet packet energy features

Abstract: This paper proposes a high performance texture classification method using dominant energy features from wavelet packet decomposition. We decompose the texture images with a family of real orthonormal wavelet bases and compute the energy signatures using the wavelet packet coefficients. Then we select a few of the most dominant energy values as features and employ a Mahalanobis distance classifier to classify a set of distinct natural textures selected from the Brodatz album. In our experiments, the proposed method employed a reduced feature set and involved less computation in classification time while still archiving high accuracy rate (94.8%) for classifying twenty classes of natural texture images.

Eigensnakes for vessel segmentation in angiography

Abstract: We introduce a new deformable model, called eigensnake, for segmentation of elongated structures in a probabilistic framework. Instead of snake attraction by specific image features extracted independently of the snake, our eigensnake learns an optimal object description and searches for such image feature in the target image. This is achieved applying principal component analysis on image responses of a bank of Gaussian derivative filters. Therefore, attraction by eigensnakes is defined in terms of classification of image features. The potential energy for the snake is defined in terms of likelihood in the feature space and incorporated into a new energy minimising scheme. Hence, the snake deforms to minimise the mahalanobis distance in the feature space. A real application of segmenting and tracking coronary vessels in angiography is considered and the results are very encouraging.

Multivariate outliers and decompositions of mahalanobis distance

Abstract: Two decompositions of the Mahalanobis distance are considered. These decompositions help to explain some reasons for the outlyingness of multivari- ate observations. They also provide a graphical tool for identifying outliers including those that have a large influence on the multiple correlation coefficient, Illustrative examples are given.

Model Updating for the Identification of NIR Spectra from a Pharmaceutical Excipient

Abstract: The effect of model updating on the identification of a pharmaceutical excipient based on its near-infrared (NIR) spectra has been investigated. A pragmatic updating approach, consisting of adding stepwise newly available samples to the training set and rebuilding the classification model, was applied. Its performance is compared for three pattern recognition methods: the wavelength distance method, the Mahalanobis distance method, and the SIMCA (soft independent modeling of class analogy) residual variance method. For the wavelength distance method, the updating approach is straightforward. In the case of the multivariate classification methods, which are based on a certain number of significant principal components (PCs), the selection of the number of PCs included in the model must be performed with care, as this number has a major impact on the classification results.

Statistical procedures for task assignment and robot selection in assembly cells

Abstract: The purpose of this paper is to show how statistical procedures can be used to design robotic assembly cells. The proposed methodologyhas two stages. In the first stage, a fuzzy clustering algorithm is employed to group similar tasks together so that they can be assigned to robots while maintaining a balanced cell and achieving a desired production cycle time. In the second stage, a Mahalanobis distance procedure is used to select robots appropriate for the task groups. The proposed approach recognizes and exploits the flexibility of robots. It also recognizes that the manufacturer specifications of robots do not hold simultaneously under normal operating conditions. A numerical example is presented and a small experiment is conducted to test the procedures.

Multi-scale feature extraction and nested-subset classifier design for high accuracy handwritten character recognition

Abstract: Both efficient representation and robust classification are essential to high-performance cursive offline handwritten Chinese character recognition. A novel multi-scale feature extraction method is presented based on the information entropy theory. Feature detection and compression are thus combined into an integrated optimization process. A series of optimal feature-spaces are constructed at varying values of the scale parameter and the best one is obtained with the maximum LDA criterion over the scale interval. For more robust classification, we introduce a structure into the Mahalanobis distance classifier and strike the balance between machine capacity and the performance on the training data in light of the ideas of structural risk minimization. A high accuracy recognition system is developed based on the new methods and for the first time, 4 widely different databases ranging from regular to completely unconstrained with several structural distortions and stroke connections are fully tested. The accuracies of 99.S% on regular database and 88.4% on cursive one at the speed of over 40 characters/s are achieved.

The Use of Feature Selection Techniques in the Context of Artificial Neural Networks

Abstract: Feature selection is an important issue, especially for classification problems where artificial neural networks are involved. It is known that using large number of inputs can make the network overspecific and require significantly longer time to learn the characteristics of the training data. Such over-specificity also reduces the generalisation capabilities of a neural network, so the network may fail to classify new data outside the range of the training data. Although feature selection methods have been used in remote sensing studies for many years, their use in the context of artificial neural networks has not been fully investigated. This paper sets out some results of an investigation of feature selection techniques, specifically the separability indices, in the problem of determining the optimum network structure in terms of achieved accuracy. For this purpose, separability indices, including divergence, transformed divergence, Bhattacharyya distance and Jeffries-Matusita distance, and the Mahalanobis distance classifier (MDC) based on two accuracy measures are employed to determine the best eight-band combination out of a 24 band multitemporal dataset. Two search procedures, sequential forward selection and the genetic algorithm, have been used to search for the best band combinations using separability measures as evaluation functions.

Nautical Scene Segmentation Using Variable Size Image Windows and Feature Space Reclustering

Abstract: This paper describes the development of a system for the segmentation of small vessels and objects present in a maritime environment. The system assumes no a priori knowledge of the sea, but uses statistical analysis within variable size image windows to determine a characteristic vector that represents the current sea state. A space of characteristic vectors is searched and a main group of characteristic vectors and its centroid found automatically by using a new method of iterative reclustering. This method is an extension and improvement of the work described in [9]. A Mahalanobis distance measure from the centroid is calculated for each characteristic vector and is used to determine inhomogenities in the sea caused by the presence of a rigid object. The system has been tested using several input image sequences of static small objects such as buoys and small and large maritime vessels moving into and out of a harbour scene and the system successfully segmented these objects.

Supervised Mineral Classification with Semiautomatic Training and Validation Set Generation in Scanning Electron Microscope Energy Dispersive Spectroscopy Images of Thin Sections

Abstract: This paper addresses the problem of classifying minerals common in siliciclastic and carbonate rocks Twelve chemical elements are mapped from thin sections by energy dispersive spectroscopy in a scanning electron microscope (SEM) Extensions to traditional multivariate statistical methods are applied to perform the classification First, training and validation sets are grown from one or a few seed points by a method that ensures spatial and spectral closeness of observations Spectral closeness is obtained by excluding observations that have high Mahalanobis distances to the training class mean Spatial closeness is obtained by requesting connectivity Second, class consistency is controlled by forcing each class into 5‐10 subclasses and checking the separability of these subclasses by means of canonical discriminant analysis Third, class separability is checked by means of the Jeffreys‐Matusita distance and the posterior probability of a class mean being classified as another class Fourth, the actual classification is carried out based on four supervised classifiers all assuming multinormal distributions: simple quadratic, a contextual quadratic, and two hierarchical quadratic classifiers Overall weighted misclassification rates for all quadratic classifiers are very low for both the training (025‐033%) and validation sets (065‐113%) Finally, the number of rejected observations in routine runs is checked to control the performance of the SEM image acquisition and the classification Although the contextual classifier performs marginally best on the validation set, the simple quadratic classifier is chosen in routine classifications because of the lower processing time required The method is presently used as a routine petrographical analysis method at Norsk Hydro Research Centre The data can be approximated by a Poisson distribution Accordingly, the square root of the data has constant variance and a linear classifier can be used Near orthogonal input data, enable the use of a minimum distance classifier Results from both linear and quadratic minimum distance classifications are described briefly

Texture segmentation of a 3D seismic section with wavelet transform and Gabor filters

Abstract: A supervised segmentation of a 3D seismic section is carried out using wavelet packet transform with the Daubechies-2 orthogonal base and multichannel Gabor filtering. Certain features are computed on the wavelet expansion and on the Gabor-filtered signal, and used by a Mahalanobis classifier to recognize and subsequently segment the seismic section. The Gabor filters of the multichannel scheme were designed by considering the minimal classification error in the recognition of geologically well understood zones, taken as patterns. As a result of the segmentation, zones of different internal stratification are identified in the seismic section. This recognition is based on the comparison of the 3D seismic data with the reference patterns extracted from the representative areas, characterized by different textures. A sandy channel is detected in both cases to a given depth. Applied to these data, the methods distinguish clearly between different layers and efficiently separate zones of different internal stratification.

A comparative study of filter based texture operators using Mahalanobis distance

Abstract: Texture feature extraction operators, which comprise linear filtering, eventually followed by post-processing, are considered. The filters used are Laws' masks (1980), filters derived from well-known discrete transforms, and Gabor filters. The post-processing step comprises nonlinear point operations and/or local statistics computation. The performance is measured by means of the Mahalanobis distance between clusters of feature vectors derived from different textures. The results show that post-processing improves considerably the performance of filter based texture operators.

Two-stage computational cost reduction algorithm based on Mahalanobis distance approximations

Abstract: For many pattern recognition methods, high recognition accuracy is obtained at very high expense of computational cost. In this paper, a new algorithm that reduces the computational cost for calculating discriminant function is proposed. This algorithm consists of two stages which are feature vector. Division and dimensional reduction. The processing of feature division is based on characteristic of covariance matrix. The dimensional reduction in the second stage is done by an approximation of the Mahalanobis distance. Compared with the well-known dimensional reduction method of K-L expansion, experimental results show the proposed algorithm not only reduces the computational cost but also improves the recognition accuracy.

Method for diagnosing equipment

Abstract: PROBLEM TO BE SOLVED: To determine equipment failure without depending on the experience and intuition of experts. SOLUTION: The operation variate of equipment under normal conditions is measured (S1). From the data of the measurement, variate data under anomalous conditions is estimated (S2). An AR model (autoregressive linear prediction model) is formed on each case under normal conditions and anomalous conditions (S3 and S4). Expressions D12 and D22 of Mahalanobis distance are obtained on the basis of the AR models (S6 and S7). The above-mentioned steps are preparations. At the time of diagnosing equipment, variate data is actually measured to form AR models. Whether it is normal or anomalous is determined on the basis of the closeness of the models.

Improving the direction-of-arrival resolution via double code filtering in WCDMA

Abstract: This paper presents a new method for improving the azimuthal resolution at the base station of the WCDMA system employing antenna array. The method uses two covariance matrices resulting from double code filtering. The first matrix is extracted by the traditional de-spreading and results in noisy signal estimate. The other matrix is extracted by an orthogonal re-spread method. The orthogonal re-spread (OR) is a new extraction method for the spatial correlation matrix of interference and noise only. The method uses one of the remaining orthogonal variable spreading factor code (OVSF) that is orthogonal to all traffic and control channels to spread energy form the desired user while attempting to maintain the same level of multiple access interference. In the presented method information obtained from OR correlation matrix is used to suppress interference and improve the signal resolution via the whitening (Mahalanobis) transform. Since the proposed method operates as additional transform on the sample covariance matrix, it can be used in conjunction with all DoA estimation methods.

Detecting and Dealing with Outliers in Univariate and Multivariate Contexts.

Abstract: Because multivariate statistics are increasing in popularity with social science researchers, the challenge of detecting multivariate outliers warrants attention. Outliers are defined as cases which, in regression analyses, generally lie more than three standard deviations from Yhat and therefore distort statistics. There are, however, some outliers that do not distort statistics when they are on the mean of Yhat lines. In univariate analyses, finding outliers can be accomplished using Casewise Diagnostics in the Statistical Package for the Social Sciences (SPSS) version 9.0, which as a three standard deviation default that can be changed easily by the researcher. In bivariate and multivariate analyses, finding outliers more than three standard deviations from Yhat is not as easy. Casewise Diagnostics will detect outliers of "Y"; however, in multivariate analyses, statistics can be distorted by a case lying within the arbitrary three standard deviations because it is said to be exerting so much influence or leverage on the regression line that, in fact, the regression line is distorted. There are two popular ways of detecting this leverage, through distance and influence calculations. The most popular statistic for detecting outliers using distance calculations is Mahalanobis. Several other ways of detecting leverage in multivariate cases are available in SPSS 9.0. Once a researcher has identified a case as being a possible outlier, then the choices are to find out if there has been an error in recording the data, or if the outlier is truly an outlier. it can be argued that there are always going to be outliers in the population as a whole, and this is an argument for keeping the score, because it reflects something natural about the general population. If the researcher decides to drop the case, then the researcher should report it and offer reasons why. (Contains 10 figures, 3 tables, and 20 references.) (Author/SLD) Reproductions supplied by EDRS are the best that can be made from the original document.

New Fast Algorithms for Error Rate-Based Stepwise Variable Selection in Discriminant Analysis

Abstract: Variable selection is an important technique for reducing the dimensionality in multivariate predictive discriminant analysis and classification. In the past, direct evaluation of the subsets by means of a classifier has been computationally too expensive, rendering necessary the use of heuristic measures of class separation, such as Wilk's $\Lambda$ or the Mahalanobis distance between class means. We present new fast algorithms for stepwise variable selection based on quadratic and linear classifiers with time complexities which, to within a constant, are the same as those applying measures of class separation. Comparing the new algorithms to previous implementations of classifier-based variable selection, we show that dramatic speed-ups are achieved.

Method and apparatus for evaluating image

Abstract: The image evaluation method and apparatus check for image abnormalities. In the method and apparatus, a Mahalanobis space is set preliminarily using a specified characteristic quantity extracted from specified image data. Then, Mahalanobis distance in the Mahalanobis space using image data read from an image to be evaluated. The calculated Mahalanobis distance is compared with a specified threshold value to check for the image abnormalities in the image to be evaluated. The method and apparatus provide a quantitative criterion in checking for abnormal image and enable image evaluation to be performed in a simple and rapid manner to realize efficient inspection.

Method and device for information processing and program providing medium

Abstract: In estimating a state variable of a target on the basis of a previously estimated state variable and an observation value, a prediction value of a next state variable is found on the basis of the previously estimated state variable by a prediction process pre, and the state variable is updated by a first update process ed on the basis of the resultant prediction value and the observation value so as to find an update value. By a second update process opt, whether the update value meets the boundary condition or not is checked, and when the update value does not meet the boundary condition, a new update value is found which minimizes the Mahalanobis distance from the prediction value in the state space and the square sum or the absolute value sum of the Mahalanobis distance from the observation value in the observation space under the boundary condition. Thus, a parameter of motion information or the like of the target can be optimally estimated under the boundary condition or constraint.