Normalized Similarity Measure Research Articles

Generalizations and Properties of Normalized Similarity Measures for Boolean Models

In this paper, we provide a closer look at some of the most popular normalized similarity/distance measures for Boolean models. This work includes the generalization of three classes of measures described as generalized Kulczynski, generalized Jaccard, and generalized Consonni and Todeschini measures, theoretical ordering of the similarity measures inside each class, as well as between classes, and positive and negative results regarding the metric properties of measures related to satisfying or not satisfying the triangle inequality axiom.

A New Method for 2D-Adapted Wavelet Construction: An Application in Mass-Type Anomalies Localization in Mammographic Images

This contribution presents a wavelet-based algorithm to detect patterns in images. A two-dimensional extension of the DST-II is introduced to construct adapted wavelets using the equation of the tensor product corresponding to the diagonal coefficients in the 2D discrete wavelet transform. A 1D filter was then estimated that meets finite energy conditions, vanished moments, orthogonality, and four new detection conditions. These allow, when performing the 2D transform, for the filter to detect the pattern by taking the diagonal coefficients with values of the normalized similarity measure, defined by Guido, as greater than 0.7, and α=0.1. The positions of these coefficients are used to estimate the position of the pattern in the original image. This strategy has been used successfully to detect artificial patterns and localize mass-like abnormalities in digital mammography images. In the case of the latter, high sensitivity and positive predictive value in detection were achieved but not high specificity or negative predictive value, contrary to what occurred in the 1D strategy. This means that the proposed detection algorithm presents a high number of false negatives, which can be explained by the complexity of detection in these types of images.

Tracking of a Fixed-Shape Moving Object Based on the Gradient Descent Method.

Tracking moving objects is one of the most promising yet the most challenging research areas pertaining to computer vision, pattern recognition and image processing. The challenges associated with object tracking range from problems pertaining to camera axis orientations to object occlusion. In addition, variations in remote scene environments add to the difficulties related to object tracking. All the mentioned challenges and problems pertaining to object tracking make the procedure computationally complex and time-consuming. In this paper, a stochastic gradient-based optimization technique has been used in conjunction with particle filters for object tracking. First, the object that needs to be tracked is detected using the Maximum Average Correlation Height (MACH) filter. The object of interest is detected based on the presence of a correlation peak and average similarity measure. The results of object detection are fed to the tracking routine. The gradient descent technique is employed for object tracking and is used to optimize the particle filters. The gradient descent technique allows particles to converge quickly, allowing less time for the object to be tracked. The results of the proposed algorithm are compared with similar state-of-the-art tracking algorithms on five datasets that include both artificial moving objects and humans to show that the gradient-based tracking algorithm provides better results, both in terms of accuracy and speed.

Supervoxel-Based Region Growing Segmentation for Point Cloud Data

Point cloud segmentation is a crucial fundamental step in 3D reconstruction, object recognition and scene understanding. This paper proposes a supervoxel-based point cloud segmentation algorithm in region growing principle to solve the issues of inaccurate boundaries and nonsmooth segments in the existing methods. To begin with, the input point cloud is voxelized and then pre-segmented into sparse supervoxels by flow constrained clustering, considering the spatial distance and local geometry between voxels. Afterwards, plane fitting is applied to the over-segmented supervoxels and seeds for region growing are selected with respect to the fitting residuals. Starting from pruned seed patches, adjacent supervoxels are merged in region growing style to form the final segments, according to the normalized similarity measure that integrates the smoothness and shape constraints of supervoxels. We determine the values of parameters via experimental tests, and the final results show that, by voxelizing and pre-segmenting the point clouds, the proposed algorithm is robust to noises and can obtain smooth segmentation regions with accurate boundaries in high efficiency.

E-CLONALG: A classifier based on Clonal Selection Algorithm

This paper proposes an improved version of CLONALG, Clone Selection Algorithm based on Artificial Immune System(AIS), that matches with the conventional classifiers in terms of accuracy tested on the same data sets. Clonal Selection Algorithm is an artificial immune system model. Instead of randomly selecting antibodies, it is proposed to take k memory pools consisting of all the learning cases. Also, an array averaged over the pools is created and is considered for cloning. Instead of using the best clone and calculating the similarity measure and comparing with the original cell, here, k best clones were selected, the average similarity measure was evaluated and noise was filtered. This process enhances the accuracy from 76.9 percentage to 94.2 percentage, ahead of the conventional classification methods.

The Unifying Frameworks of Information Measures

Information measures are capable of providing us with fundamental methodologies to analyze uncertainty and unveiling the substantive characteristics of random variables. In this paper, we address the issues of different types of entropies through q-generalized Kolmogorov-Nagumo averages, which lead to the propositions of the survival Rényi entropy and survival Tsallis entropy. Therefore, we make an inventory of eight types of entropies and then classify them into two categories: the density entropy that is defined on density functions and survival entropy that is defined on survival functions. This study demonstrates that, for each type of the density entropy, there exists a kind of the survival entropy corresponding to it. Furthermore, the similarity measures and normalized similarity measures are, respectively, proposed for each type of entropies. Generally, functionals of different types of information-theoretic metrics are equally diverse, while, simultaneously, they also exhibit some unifying features in all their manifestations. We present the unifying frameworks for entropies, similarity measures, and normalized similarity measures, which helps us deal with the available information measures as a whole and move from one functional to another in harmony with various applications.

Unsupervised Classification of PolSAR Data Using a Scattering Similarity Measure Derived From a Geodesic Distance

In this letter, we propose a novel technique for obtaining scattering components from Polarimetric Synthetic Aperture Radar (PolSAR) data using the geodesic distance on the unit sphere. This geodesic distance is obtained between an elementary target and the observed Kennaugh matrix, and it is further utilized to compute a similarity measure between scattering mechanisms. The normalized similarity measure for each elementary target is then modulated with the total scattering power (Span). This measure is used to categorize pixels into three categories i.e. odd-bounce, double-bounce and volume, depending on which of the above scattering mechanisms dominate. Then the maximum likelihood classifier of [J.-S. Lee, M. R. Grunes, E. Pottier, and L. Ferro-Famil, Unsupervised terrain classification preserving polarimetric scattering characteristics, IEEE Trans. Geos. Rem. Sens., vol. 42, no. 4, pp. 722731, April 2004.] based on the complex Wishart distribution is iteratively used for each category. Dominant scattering mechanisms are thus preserved in this classification scheme. We show results for L-band AIRSAR and ALOS-2 datasets acquired over San Francisco and Mumbai, respectively. The scattering mechanisms are better preserved using the proposed methodology than the unsupervised classification results using the Freeman-Durden scattering powers on an orientation angle (OA) corrected PolSAR image. Furthermore, (1) the scattering similarity is a completely non-negative quantity unlike the negative powers that might occur in double- bounce and odd-bounce scattering component under Freeman Durden decomposition (FDD), and (2) the methodology can be extended to more canonical targets as well as for bistatic scattering.

FaDA: Fast Document Aligner using Word Embedding

Abstract FaDA is a free/open-source tool for aligning multilingual documents. It employs a novel crosslingual information retrieval (CLIR)-based document-alignment algorithm involving the distances between embedded word vectors in combination with the word overlap between the source-language and the target-language documents. In this approach, we initially construct a pseudo-query from a source-language document. We then represent the target-language documents and the pseudo-query as word vectors to find the average similarity measure between them. This word vector-based similarity measure is then combined with the term overlap-based similarity. Our initial experiments show that s standard Statistical Machine Translation (SMT)- based approach is outperformed by our CLIR-based approach in finding the correct alignment pairs. In addition to this, subsequent experiments with the word vector-based method show further improvements in the performance of the system.

Intra-hemispheric intrinsic connectivity asymmetry and its relationships with handedness and language Lateralization

Asymmetry in intra-hemispheric intrinsic connectivity, and its association with handedness and hemispheric dominance for language, were investigated in a sample of 290 healthy volunteers enriched in left-handers (52.7%). From the resting-state FMRI data of each participant, we derived an intra-hemispheric intrinsic connectivity asymmetry (HICA) matrix as the difference between the left and right intra-hemispheric matrices of intrinsic correlation computed for each pair of the AICHA atlas ROIs. We defined a similarity measure between the HICA matrices of two individuals as the correlation coefficient of their corresponding elements, and computed for each individual an index of intra-hemispheric intrinsic connectivity asymmetry as the average similarity measure of his HICA matrix to those of the other subjects of the sample (HICAs). Gaussian-mixture modeling of the age-corrected HICAs sample distribution revealed that two types of HICA patterns were present, one (Typical_HICA) including 92.4% of the participants while the other (Atypical_HICA) included only 7.6% of them, mostly left-handers. In addition, we investigated the relationship between asymmetry in intra-hemispheric intrinsic connectivity and language hemispheric dominance, including a potential effect of handedness on this relationship, thanks to an FMRI acquisition during language production from which an hemispheric functional lateralization index for language (HFLI) and a type of hemispheric dominance for language, namely leftward, ambilateral, or rightward, were derived for each individual. There was a significant association between the types of language hemispheric dominance and of intra-hemispheric intrinsic connectivity asymmetry, occurrence of Atypical_HICAs individuals being very high in the group of individuals rightward-lateralized for language (80%), reduced in the ambilateral group (19%) and rare in individuals leftward-lateralized for language (less than 3%). Quantitatively, we found a significant positive linear relationship between the HICAs and HFLI indices, with an effect of handedness on the intercept but not on the slope of this relationship. These findings demonstrate that handedness and hemispheric dominance for language are significantly but independently associated with the asymmetry of intra-hemispheric intrinsic connectivity. These findings suggest that asymmetry in intra-hemispheric connectivity is a variable phenotype shaped in part by hemispheric lateralization for language, but possibly also depending on other lateralized functions.

Crystal lattice identification on the basis of assessing unit cell parameters

The paper is devoted to the problem of crystal lattice identification in three- dimensional space. Two identification methods based on assessing the parameters of unit cells have been developed to solve this problem. The first method is based on the calculation and comparison of the six basic cell parameters: the three lengths of the cell sides and the three magnitudes of angles between the sides. The second method is based on the calculation and comparison of Wigner-Seitz cell volumes using the Monte-Carlo method. Both methods include normalized similarity measures for the parameters calculated: the edge similarity and the angle similarity measure for Bravais unit cells and the volume similarity measure for Wigner-Seitz cells. A number of computational experiments for crystal lattice identification were carried out using these methods. In the course of the experiments, first, the dependence of the accuracy of defining the volume of a Wigner-Seitz cell on the number of the points included and, second, the probability of accurate Bravais lattice identification when comparing different sets of the parameters assessed have been specified. As a result of the experiments carried out, the accuracy of identification for four systems of Bravais lattices exceeded 95 per cent.

Crystal lattice identification by coordinates of their nodes in three dimensional space

The problem of crystal lattice identification is considered. Two matching methods were developed to solve this problem: the method based on estimation of the main Bravais cell parameters and the method based on estimation of the Wigner-Seitz cell volume. Both methods include the normalized similarity measures: edge similarity measure and angle similarity measure for Bravais cells and volume similarity measure for Wigner-Seitz cells. The experiments for crystal lattice identification were carried out using these methods and the analysis of derived results is given. The results demonstrate that the best way to identify lattice system type is simultaneous application of all introduced similarity measures.

Non-standard similarity/dissimilarity analysis of DNA sequences

Similarity/dissimilarity analysis of DNA sequences is performed using 3D-dynamic representation. The sequences are represented by material points in a 3D-space. Descriptors related to such 3D-dynamic graphs are calculated. A new normalized similarity measure is introduced for a comparison of the sequences. The method is applied to β-globin (HBB) genes of different species. Different methods are compared. The multidimensionality of similarity space of complex objects is discussed (different results for each of the methods). It is also shown that it is easier to find the properties which are consistent with the observation of the evolution of the species for the second exons of β-globin gene rather than for the first ones.

Automatic segmentation of hippocampal subfields based on multi-atlas image segmentation techniques

The volume of hippocampal subfields is closely related with early diagnosis of Alzheimer’s disease. Due to the anatomical complexity of hippocampal subfields, automatic segmentation merely on the content of MR images is extremely difficult. We presented a method which combines multi-atlas image segmentation with extreme learning machine based bias detection and correction technique to achieve a fully automatic segmentation of hippocampal subfields. Symmetric diffeomorphic registration driven by symmetric mutual information energy was implemented in atlas registration, which allows multi-modal image registration and accelerates execution time. An exponential function based label fusion strategy was proposed for the normalized similarity measure case in segmentation combination, which yields better combination accuracy. The test results show that this method is effective, especially for the larger subfields with an overlap of more than 80%, which is competitive with the current methods and is of potential clinical significance.

Semantic Approach for Classification of Web Services Using Unsupervised Normalized Similarity Measure

Automatic or semiautomatic categorization of web services facilitates the relevant service retrieval as well as it helps the administrators in attaining globally consistent classification decisions that are independent of the administrator’s knowledge of application domain, organization of taxonomies and service characteristics. Lack of automatic mechanisms to help service publishers in the classification task, irrelevant and huge number of services returned by the UDDI, and lack of standard mechanisms that helps in the discovery of desired web services are some of the issues that need focus. In this paper, an automatic approach for service categorization is proposed that uses a lexical semantic network constructed from the web snippets as a knowledge base for semantic similarity calculation between the service profiles and the categories. Our approach involves mapping of service profiles to a category based dimension vector by using the notion of semantic similarity and aims at alleviating the administrator’s job by automatically providing them with a set of categories ranked based on the degree of semantic similarity. Empirical evaluation on a set of OWL-S services shows that the proposed approach helps in better decisions for relevant classification of services by giving an ordered set of categories based on the similarity scores.

Sea surface temperatures in cooler climate stages bear more similarity with atmospheric CO2 forcing

The interglacial Marine Isotope Stage (MIS) 11 received special attention due to its remarkable resemblance with present‐day climate. Based on synchronicity of marine, ice sheet and terrestrial proxy responses, warm episodes with intervening cool phase(s) at MIS 11 are qualitatively established. Here we quantitatively evaluate 15 climate proxies during 368–552 kyr intervals adopting a novel long‐range cross‐correlation approach and information theory based similarity measures. We also estimate the information flow rate and dominant flow direction between these climate variables using transfer entropy and the related directionality index. Our results unequivocally establish that atmospheric CO2 (pCO2) is the driving signal while all other proxies used in this study are the responses. The climate forcing greenhouse gas, the atmospheric CO2 (pCO2) and the response signals like sea surface temperature (SST) and carbon isotope composition of total organic carbon (δ13CTOC) proxies are strongly correlated (∼1 or −1) without significant observable time lag (less than 1 kyr). Various substages of MIS 11 are recognizable in the SST data alone based on normalized similarity measures. Additionally, eight more proxies from lacustrine sediments are identified as primary. During the cooler substages these proxies bear more similarity with ambient atmospheric pCO2. Thus, the information theory–based similarity measures suggest that atmospheric CO2 fluctuations are best captured by at least 9 climate proxies during cooler interglacial events. Based on the results related to interglacial MIS 11 and 13 obtained in this study, an important implication relevant to anthropogenic CO2input to the present‐day atmosphere can be distilled. It is that sensitive and better‐coupled response proxies such as SST and MAT, which already show an increasing trend, are likely to behave in a more dissimilar manner in future. That is, they tend to behave more independently in the near future (∼0.75 kyr).

MEASURING INCLUSION, SIMILARITY AND COMPATIBILITY BETWEEN ATANASSOV'S INTUITIONISTIC FUZZY SETS

This paper establishes an axiomatic definition of inclusion measures between Atanassov's intuitionistic fuzzy (A-IF for short) sets. Some kinds of A-IF inclusion measures are constructed by different A-IF operators especially by A-IF implicator, and some new methods for measuring the degree of similarity between A-IF sets are proposed. Moreover, the similarity measure obtained from an A-IF inclusion measure satisfies properties of normal similarity measure. We then define a compatibility measure by a predicates logical idea and construct several functions to measure compatibility for an intuitionistic t -norm.

Hybrid monotonic inclusion measure and its use in measuring similarity and distance between fuzzy sets

In fuzzy set theory, an inclusion measure indicates the degree to which a given fuzzy set is contained in another fuzzy set. In this paper, we introduce a new definition of inclusion measure: hybrid monotonic inclusion measure ( S inclusion measure for short). Some inclusion measures defined by other authors and us are proved to be S inclusion measures. In addition, as the rational generalization of crisp inclusion relation, we prove most of the S inclusion measures possess some kind of T -transitivity and some properties postulated by Sinha and Dougherty for a reasonable inclusion measure. Similarity measure and distance between two fuzzy sets derived from S inclusion measure possess properties of normal similarity measure and distance. Furthermore, some similarity measures possess T -transitivity and distributivity.

Assignment Tests for Variety Identification Compared to Genetic Similarity‐Based Methods Using Experimental Datasets from Different Marker Systems in Sugar Beet

High genetic variation within sugar beet (Beta vulgaris L.) varieties hampers reliable classification procedures independent of the type of marker technique applied. Datasets on amplified fragment length polymorphisms, sequence tagged microsatellite sites, and cleaved amplified polymorphic sites markers in eight sugar beet varieties were subjected to supervised classifiers, methods in which individual assignments are made to predefined classes, and unsupervised classifiers, defined afterward on the similarity in marker composition from sampled individuals. Major issues addressed are (i) which classification method gives the most consistent results when three marker techniques are compared, and (ii) given different classification techniques available, for which marker technique is the output generated least constrained by the way data analysis is performed. Assignment tests showed a higher consistency across classifications independent from the marker technique. A good allocation to the proper variety was obtained, together with a reliable allocation pattern among the other varieties. Both aspects deal with the variation within a variety and the distance to other varieties. Assignment data were transformed into an average similarity measure, similarity by assignment (Sax,y), which is a new genetic distance measure with interesting properties.

Performance of a wavelet matched filter with optimized dilation designed using simulated annealing algorithm

We outline a method, using simulated annealing algorithm, to optimize the dilation factor of the wavelet function, from a number of images. The optimized dilation factor has been used to obtain wavelet matched filters (WMFs) for different fingerprint images. A single WMF instead of a bank of filters can be used for identifying each image. The filter performance was tested in terms of discrimination ratio, peak-to-correlation plane energy ratio, peak-to-sidelobe ratio and average similarity measure with digitally degraded fingerprints. The performance has been compared with that of a classical matched filter and a phase only filter. The filter performance has also been tested for noisy images.

Improving the false alarm capabilities of the maximum average correlation height correlation filter

We show that the maximum average correlation height (MACH) correlation filter overemphasizes the importance of the average training image leading to poor discrimination of the desired class images from clutter images. To overcome this, two new metrics termed the all image correlation height (AICH) and the modified average similarity measure (MASM) are introduced and optimized in a new correlation de- sign. The resulting filter exhibits improved clutter rejection performance while retaining the attractive distortion tolerance feature of the MACH filter. Simulation results based on a simulated synthetic aperture radar (SAR) image database are presented to illustrate the new filter's proper- ties. © 2000 Society of Photo-Optical Instrumentation Engineers. (S0091-3286(00)01805-5)