Search

Evaluation of food microbiome is of major importance since it accounts for the product's organoleptic characteristics and their nutritional value. In this dataset, microbes present in olive samples (‘Kalamata’ variety) stored under modified atmosphere and throughout different time-points of the shelf life of the product are presented, originated after 16S and 18S rRNA sequencing. The different time-points analyzed were: T0 (immediately after packaging), T6 (six months of storage), T12 (12 months of storage) and T18 (six months after the end of shelf life). Sequencing was performed on a MiSeq platform with the MiSeq Reagent Kit v3 (600 cycles). The raw sequence data used for analysis are available in NCBI under the Sequence Read Archive (SRA), with BioProject ID PRJNA688686. Raw reads were analyzed using the QIIME2 pipeline, clustered into Operational Taxonomic Units (OTU) and aligned against SILVA 132 reference database. OTUs are presented in different taxonomic levels for each time-point. These data present valuable information on the microbial communities of table olives, a dynamic niche that affect the final product quality. The data presented are related to the research article “Insights into the evolution of Greek style table olives microbiome stored under modified atmosphere: biochemical implications on the product quality” [1] .

In this work, we propose a mechanism for knowledge transfer between Convolutional Neural Networks via the geometric regularization of local features produced by the activations of convolutional layers. We formulate appropriate loss functions, driving a “student” model to adapt such that its local features exhibit similar geometrical characteristics to those of an “instructor” model, at corresponding layers. The investigated functions, inspired by manifold-to-manifold distance measures, are designed to compare the neighboring information inside the feature space of the involved activations without any restrictions in the features’ dimensionality, thus enabling knowledge transfer between different architectures. Experimental evidence demonstrates that the proposed technique is effective in different settings, including knowledge-transfer to smaller models, transfer between different deep architectures and harnessing knowledge from external data, producing models with increased accuracy compared to a typical training. Furthermore, results indicate that the presented method can work synergistically with methods such as knowledge distillation, further increasing the accuracy of the trained models. Finally, experiments on training with limited data show that a combined regularization scheme can achieve the same generalization as a non-regularized training with 50% of the data in the CIFAR-10 classification task.

Offline writer independent (WI) signature verification is conceivably a challenging task in the domain of handwritten biometrics. This work addresses it by introducing a feature extraction scheme that relies on the detection of first order transitions between asymmetrical lattice arrangements of simple pixel structures. The experiments are conducted with a decision stumps committee, accompanied with boosting feature selection, by employing only unrelated or blind training and testing datasets, all derived from four widely used signature databases. In addition, a fifth signature dataset which contains disguised signatures, originating from a signature verification contest, is also used for testing purposes. The impact of the preprocessing stage per dataset is exploited by allowing various pruning levels of the raw binary signature. In addition to standard training protocols, we introduce the use of the applicability domain (AD) in WI signature verification and examine its effectiveness in providing reliable classifier predictions by quantifying subregions of the input distance space that have been sufficiently covered by training examples. The derived experimental results expressed by means of equal error rate, along with best average error rate, are considered to be very competitive to those provided from state of the art WI methods.

Spanish-style green olives are one of the main trade preparations in the international market. Nowadays, there is an increasing trend to use plastic packaging due to reduced weight, lower costs, flexibility and convenience. The purpose of this work was to investigate the effect of modified atmosphere packaging of Spanish-style green olives in multilaminated pouches on the microbiological and physicochemical characteristics of olives. Green pitted olives of cv. Conservolea and Halkidiki were packaged in high barrier multilaminated pouches (113 μm thickness; O2 permeability < 1.5 mL/m2/24 h) under modified atmospheres (30% CO2 / 70% N2) and stored at room temperature for 12 months. Microbiological [lactic acid bacteria (LAB), yeasts, enterobacteria] and physicochemical changes (pH, acidity, salt content, texture and color) were monitored during storage together with sensory evaluation. No enterobacteria could be detected in both olive varieties. The initial microbiota consisted of LAB (5.9-6.3 log CFU/g) and yeasts (4.8-5.3 log CFU/g). Yeasts started to decline and could not be detected after 60 days. In the end, LAB dominated in populations exceeding 5.4 and 6.1 log CFU/g for cv. Halkidiki and Conservolea, respectively. pH values increased gradually from 3.51 to 4.19, the mean acidity was ca. 0.2% and the salt content fluctuated between 3.0-3.6%. Color parameters L* (luminance), a* (greenness) and b* did not change during storage in both varieties. Texture analysis showed a decrease in olive hardness from 15-20 N to 8 N. Sensory analysis revealed that cv. Halkidiki olives were crisper and with increased fibrousness compared to cv. Conservolea. The survival of LAB combined with the probiotic potential of this microbial group creates new possibilities for the use of non-thermally preserved olives as a functional food. Acknowledgement: This research has been co-financed by the European Union and Greek national funds through the Operational Program Competitiveness, Entrepreneurship and Innovation, under the call RESEARCH – CREATE – INNOVATE (project code: T1EDK-04110)

Sulphide ores hosted in deeper parts of ophiolite complexes may be related to either primary magmatic processes or links to hydrothermal alteration and metal remobilization into hydrothermal systems. The Pindos ophiolite complex was selected for the present study because it hosts both Cyprus-type sulphides (Kondro Hill) and Fe&ndash, Cu&ndash, Co&ndash, Zn sulphides associated with magnetite (Perivoli-Tsoumes) within gabbro, close to its tectonic contact with serpentinized harzburgite, and thus offers the opportunity to delineate constraints controlling their origin. Massive Cyprus-type sulphides characterized by relatively high Zn, Se, Au, Mo, Hg, and Sb content are composed of pyrite, chalcopyrite, bornite, and in lesser amounts covellite, siegenite, sphalerite, selenide-clausthalite, telluride-melonite, and occasionally tennantite&ndash, tetrahedrite. Massive Fe&ndash, Zn-type sulphides associated with magnetite occur in a matrix of calcite and an unknown (Fe,Mg) silicate, resembling Mg&ndash, hisingerite within a deformed/metamorphosed ophiolite zone. The texture and mineralogical characteristics of this sulphide-magnetite ore suggest formation during a multistage evolution of the ophiolite complex. Sulphides (pyrrhotite, chalcopyrite, bornite, and sphalerite) associated with magnetite, at deeper parts of the Pindos (Tsoumes), exhibit relatively high Cu/(Cu + Ni) and Pt/(Pt + Pd), and low Ni/Co ratios, suggesting either no magmatic origin or a complete transformation of a preexisting magmatic assemblages. Differences recorded in the geochemical characteristics, such as higher Zn, Se, Mo, Au, Ag, Hg, and Sb and lower Ni contents in the Pindos compared to the Othrys sulphides, may reflect inheritance of a primary magmatic signature.

Handwritten signature is a common biometric trait, widely used for confirming the presence or the consent of a person. Offline Signature Verification (OSV) is the task of verifying the signer using static signature images captured after the finish of signing process, with many applications especially in the domain of forensics. Deep Convolutional Neural Networks (CNNs) can generate efficient feature representations, but their training is data-intensive. Since limited training data is an intrinsic problem of an OSV system’s development, this work focuses on addressing the problem of learning informative features by employing prior knowledge from a similar task in a domain with an abundance of training data. In particular, we demonstrate that an appropriate pre-training of a CNN model in the task of handwritten text-based writer identification task, can dramatically improve the efficiency of the CNN in the OSV task, enabling to obtain state-of-the-art performance with an order of magnitude less training signature samples. In the proposed scheme, after the pre-training of the CNN in writer identification task through specially processed handwritten text data, the learned features are tailored to the signature problem though a metric learning stage that utilizes contrastive loss to learn a mapping of the signatures’ features to a latent space that suits the OSV task. At the final stage, the proposed scheme utilizes Writer-Dependent (WD) classifiers learned on a few reference samples from each writer. Our system is tested on the three challenging signature datasets, CEDAR, MCYT-75 and GPDS300GRAY. The obtained accuracy in terms of Equal Error Rates (EER) is statistically equivalent to the popular SigNet CNN, despite a significantly smaller training set of signature images and no use of skilled forgeries signatures during training.

Lately, off-line signature verification systems have been reintroduced based on the idea of modelling the signature images with various relations among their pixels. In this paper, a modified version of the partially ordered set feature extraction procedure is presented by enabling distant range interactions between preceded relations of pixel groups. In this way, the spatial diversity of correlation that exists among signature features can be exploited. The motive behind this approach is related to our belief that the particular idiosyncratic writing style characteristics of each individual will be present over the whole length of the signature. Experiments involve the well-known Center of Excellence for Document Analysis and Recognition (CEDAR) and Ministerio de Ciencia y Tecnologia (MCYT) datasets in two popular writer dependent training modes: In the first mode, genuine and simulated forgeries were utilized, while in the second one genuine and random forgeries only were utilized during the training stage of the classifier. In both cases, the testing phase is exploited with genuine, simulated and random forgeries while receiver operating characteristic along with decision oriented FAR, FRR and average error metrics are assessing the proposed feature extraction method. The results obtained, show that the long range correlation of grid features can be efficiently employed for off-line signature verification.

In this work, we present a novel approach for training Generative Adversarial Networks (GANs). Using the attention maps produced by a Teacher Network we are able to improve the quality of the generated images as well as perform weakly supervised object localization on the generated images. To this end, we generate images of HEp-2 cells captured with Indirect Imunofluoresence (IIF) and study the ability of our network to perform a weakly localization of the cell. First, we demonstrate that whilst GANs can learn the mapping between the input domain and the target distribution efficiently, the discriminator network is not able to detect the regions of interest. Second, we present a novel attention transfer mechanism which allows us to enforce the discriminator to put emphasis on the regions of interest via transfer learning. Third, we show that this leads to more realistic images, as the discriminator learns to put emphasis on the area of interest. Fourth, the proposed method allows one to generate both images as well as attention maps which can be useful for data annotation e.g in object detection.

A method for the construction of saliency maps is proposed.Features are extracted via local sparse coding on image patches.The overcomplete dictionary is trained using natural images.A bio-plausible scheme based on the Hamming distance is used to compare patch representations.The algorithm is efficient both in terms of computational cost and of detection performance. Modeling of visual saliency is an important domain of research in computer vision, given the significant role of attention mechanisms during neural processing of visual information. This work presents a new approach for the construction of image representations of salient locations, generally known as saliency maps. The developed method is based on an efficient comparison scheme for the local sparse representations deriving from non-overlapping image patches. The sparse coding stage is implemented via an overcomplete dictionary trained with a soft-competitive bio-inspired algorithm and the use of natural images. The resulting local sparse codes are pairwise compared using the Hamming distance as a gauge of their co-activation. The calculated distances are used to quantify the saliency strength for each individual patch, and then, the saliency values are non-linearly filtered to form the final map. The evaluation results obtained on four image databases, demonstrate the competitive performance of the proposed approach compared to several state-of-the-art saliency modeling algorithms. More importantly, the proposed scheme is simple, efficient, and robust under a variety of visual conditions. Thus, it appears as an ideal solution for a hardware implementation of a frontend saliency modeling module in a computer vision system.

In this work, a feature extraction method for offline signature verification is presented that harnesses the power of sparse representation in order to deliver state-of-the-art verification performance in several signature datasets like CEDAR, MCYT-75, GPDS and UTSIG. Beyond the accuracy improvements, several major parameters associated with sparse representation; such as selected configuration, dictionary size, sparsity level and positivity priors are investigated. Besides, it is evinced that 2nd order statistics of the sparse codes is a powerful pooling function for the formation of the global signature descriptor. Also, a thorough evaluation of the effects of preprocessing is introduced by an automated algorithm in order to select the optimum thinning level. Finally, a segmentation strategy which employs a special form of spatial pyramid tailored to the problem of sparse representation is presented along with the enhancing of the produced descriptor on meaningful areas of the signature as emerged from the BRISK key-point detection mechanism. The obtained state-of-the-art results on the most challenging signature datasets provide a strong indication towards the benefits of learned features, even in writer dependent (WD) scenarios with a unique model for each writer and only a few available reference samples of him/her., Comment: 10 pages

The paper presents a semi-supervised weather classification method based on 850-hPa isobaric level maps. A preprocessing step is employed, where isolines of geopotential height are extracted from weather map images via an image processing procedure. Α feature extraction stage follows where two techniques are applied. The first technique implements phase space reconstruction, and yields multidimensional delay distributions. The second technique is based on chain code representation of signals, from which histogram features are derived. Similarity measures are used to compare multidimensional data and the k-means algorithm is applied in the final stage. The method is applied over the area of Greece, and the resulting catalogues are compared to a subjective classification for this area. Numerical experiments with datasets derived from the European Meteorological Bulletin archives exhibit an up to 91 % accurate agreement with the subjective weather patterns.

In the present paper, a novel graph-based approach to the shape decomposition problem is addressed. The shape is appropriately transformed into a visibility graph enriched with local neighborhood information. A two-step diffusion process is then applied to the visibility graph that efficiently enhances the information provided, thus leading to a more robust and meaningful graph construction. Inspired by the notion of a clique as a strict cluster definition, the dominant sets algorithm is invoked, slightly modified to comport with the specific problem of defining shape parts. The cluster cohesiveness and a node participation vector are two important outputs of the proposed graph partitioning method. Opposed to most of the existing techniques, the final number of the clusters is determined automatically, by estimating the cluster cohesiveness on a random network generation process. Experimental results on several shape databases show the effectiveness of our framework for graph-based shape decomposition.

Classification of weather maps at various isobaric levels as a methodological tool is used in several problems related to meteorology, climatology, atmospheric pollution and to other fields for many years. Initially the classification was performed manually. The criteria used by the person performing the classification are features of isobars or isopleths of geopotential height, depending on the type of maps to be classified. Although manual classifications integrate the perceptual experience and other unquantifiable qualities of the meteorology specialists involved, these are typically subjective and time consuming. Furthermore, during the last years different approaches of automated methods for atmospheric circulation classification have been proposed, which present automated and so-called objective classifications. In this paper a new method of atmospheric circulation classification of isobaric maps is presented. The method is based on graph theory. It starts with an intelligent prototype selection using an over-partitioning mode of fuzzy c-means (FCM) algorithm, proceeds to a graph formulation for the entire dataset and produces the clusters based on the contemporary dominant sets clustering method. Graph theory is a novel mathematical approach, allowing a more efficient representation of spatially correlated data, compared to the classical Euclidian space representation approaches, used in conventional classification methods. The method has been applied to the classification of 850 hPa atmospheric circulation over the Eastern Mediterranean. The evaluation of the automated methods is performed by statistical indexes; results indicate that the classification is adequately comparable with other state-of-the-art automated map classification methods, for a variable number of clusters.

Online signature verification has been intensively investigated in several directions, such as the selected feature(s), similarity estimation and classification method. Local feature approaches combined with elastic distance metrics have the most successful performance so far. The Turning Angle Sequence (TAS) feature has not been extensively explored for signature verification, while the fusion of TASs of different scales, the Turning Angle Scale Space (TASS) is a new approach in this field. In this paper, we study the signatures TAS and TASS representations and their application to online signature verification. In the matching stage, a variation of the longest common sub-sequence matching technique has been employed. Experimental results using varying TAS(S) representation parameters on two publicly available signature databases, the SVC2004 and SUSIG, show the improved performance of the selected feature along with the chosen elastic distance measure on the equal error rate results of the online signature verification task.

A simple whole sky imaging system, based on a commercial digital camera with a fish-eye lens and a hemispheric dome, is used for the automatic estimation of total cloud coverage and classification. For the first time, a multi color criterion is applied on sky images, in order to improve the accuracy in detection of broken and overcast clouds under large solar zenith angles. The performance of the cloud detection algorithm is successfully compared with ground based weather observations. A simple method is presented for the detection of raindrops standing on the perimeter of hemispheric dome. Based on previous works on cloud classification, an improved k-Nearest-Neighbor algorithm is presented. The successful detection percentage of the classifier ranges between 78% and 95% for seven cloud types.

This work presents a novel face recognition approach that utilizes the whole manifold structure of the face's local spectral distribution. Local spectral features are extracted using Gabor wavelets, encoding at every image pixel the visual appearance of the surrounding patch by the vector of filter responses. The above representation provides a robust and discriminative statistical image description in the spatial frequency domain transform space. Parameterized by angle and scale, the manifold structure of the produced multidimensional point set contains both local and holistic information about the face image. In order to reduce redundancy and code efficiently the formed multivariate distribution, a neural vector quantizer is employed. The ensemble of the selected code vectors constitutes the spectral signature of a face image in the high-dimensional face space. The similarity between two face images is assessed by comparing the corresponding representative samples of the two distributions directly in the frequency space using the multivariate Wald-Wolfowitz test, a non-parametric statistical test dealing with the multivariate ''Two-Sample Problem''. Its operation is based on the construction of the minimal spanning tree, which is an effective tool for preserving and utilizing the manifold structure of the data set. The new representation is both holistic, considering the features' distribution as a whole, while at the same time utilizes local information extraction. Experimental results on four benchmark face databases demonstrate the favorable properties of the proposed methodology over traditional approaches particularly in the ''single image case''.

There is growing interest in studying the association of functional connectivity patterns with particular cognitive tasks. The ability of graphs to encapsulate relational data has been exploited in many related studies, where functional networks (sketched by different neural synchrony estimators) are characterized by a rich repertoire of graph-related metrics. We introduce commute times (CTs) as an alternative way to capture the true interplay between the nodes of a functional connectivity graph (FCG). CT is a measure of the time taken for a random walk to setout and return between a pair of nodes on a graph. Its computation is considered here as a robust and accurate integration, over the FCG, of the individual pairwise measurements of functional coupling. To demonstrate the benefits from our approach, we attempted the characterization of time evolving connectivity patterns derived from EEG signals recorded while the subject was engaged in an eye-movement task. With respect to standard ways, which are currently employed to characterize connectivity, an improved detection of event-related dynamical changes is noticeable. CTs appear to be a promising technique for deriving temporal fingerprints of the brain's dynamic functional organization.

In this study, a conceptually simple, yet flexible and extendable strategy to contrast two different color images is introduced. The proposed approach is based on the multivariate Wald-Wolfowitz test, a nonparametric test that assesses the commonality between two different sets of multivariate observations. It provides an aggregate gauge of the match between color images, taking into consideration all the (selected) low-level characteristics, while alleviating correspondence issues. We show that a powerful measure of similarity between two color images can emerge from the statistical comparison of their representations in a properly formed feature space. For the sake of simplicity, the RGB-space is selected as the feature space, while we are experimenting with different ways to represent the images within this space. By altering the feature-extraction implementation, complementary ways to portray the image content appear. The reported results, from the application on a diverse collection of images, clearly demonstrate the effectiveness of our method, its superiority over previous methods, and suggest that even further improvements can be achieved along the same line of research. It is not only the unifying character that makes our strategy appealing, but also the fact that the retrieval performance does not increase continuously with the amount of details in the image representation. The latter sets an upper limit to the computational demands and reminds of performance plateaus reached by novel approaches in information retrieval.

In this paper a novel Fuzzy Rule Based Dissimilarity Function is presented, to determine the hierarchical merging sequence in a region based segmentation scheme. The proposed technique, based on distinct region features and fuzzy logic principles, is designed to cope with the problems inherent in the segmentation task that the traditional merging cost functions cannot overcome. It combines the global (color) and local (spatial) information of the image to compare two adjacent regions in the rgb space. The validity of the approach has been subjectively and objectively verified for several types of color images such as head and shoulders, natural and texture images.

In this work a prototype video-based system for assessing the state of patients with Multiple sclerosis is proposed. In particular we introduce an automated system for capturing and analyzing gait sequences from patients performing the well known 2-minute walking test. The contribution of this work is twofold. First we provide a computerized approach for performing the 2-minute walk test and showing that there is a great correlation between the estimated by the system walking distance and the distance measured by the physicians (Pearson's Rho= 0.7292, p

Background. P63 immunostaining has been considered as potential prognostic factor in laryngeal cancer. Considering that P63 is mainly nuclear stain, a possible correlation between the texture of P63-stained nuclei and the tumor’s grade could be of value to diagnosis, since this may be related to biologic information imprinted as texture on P63 expressed nuclei.Objective. To investigate the association between P63 stained nuclei and histologic grade in laryngeal tumor lesions.Methods. Biopsy specimens from laryngeal tumour lesions of 55 patients diagnosed with laryngeal squamous cell carcinomas were immunohistochemically (IHC) stained for P63 expression. Four images were digitized from each patient’s IHC specimens. P63 positively expressed nuclei were identified, the percentage of P63 expressed nuclei was computed, and 118 textural, morphological, shape, and architectural features were calculated from each one of the 55 laryngeal lesions. Data were split into the low grade (21 grade I lesions) and high grade (34 grade II and grade III lesions) classes for statistical analysis.Results. With advancing grade, P63 expression decreased, P63 stained nuclei appeared of lower image intensity, more inhomogeneous, of higher local contrast, contained smaller randomly distributed dissimilar structures and had irregular shape.Conclusion. P63 expressed nuclei contain important information related to histologic grade.

Polymer optical fibers (POFs) are of great interest the last decade, replacing glass fibers, in high-speed optical communication networks for short and medium distance applications. The characterization of POF as a transmission medium is of great importance when designing such optical networks. Based on a generalized impulse response model of multimode media, we developed a method for time bandpass response evaluation of a graded index polymer optical fiber (GIPOF). This model can be used to derive the response (in both time and frequency domains) of any arbitrary light source with given optical pulse characteristics.

A new offline handwritten signature modeling is introduced that confluences disciplines from grid feature extraction and information theory. The proposed scheme advances further a previously reported feature extraction technique which exploits pixel transitions along the signature trace over predetermined two pixel paths. In this new work the feature components, partitioned in groups, are considered as events of a grid based discrete space probabilistic source. Based on the 16-ary FCB2 feature, a set of 87 orthogonal event schemes, organized in tetrads, is identified. Next an entropy rule is drawn in order to declare the most appropriate tetrad scheme for representing a writer's signature. When skilled forgery is encountered verification results derived on both the GPDS300 dataset and a proprietary one, indicate enhanced EER rates compared to other approaches, including the previous reference of FCB2 as well.

A new family of data dependent nonlinear filters (DDNL) is presented. Coefficients are computed locally and the absolute distance between signal samples is used to determine a rank order dependent weighting function. The new filters have very good performance regarding noise suppression and impulse rejection and at the same time provide better edge preservation than the averager and other linear smoothers. >

A robust methodology is presented in this chapter that tries to tackle the texture classification problem. The multivariate WW-test and the KWass distance are used in order to measure how close two distributions are. Their generic character stems from the fact that, by altering the character of texture image characteristics, we can modify the flavour of formulated queries. To avoid problems associated with histograms as empirical estimates of the distribution (e.g., the binning effect), particularly in high-dimensional spaces, dissimilarity between texture distributions is computed using distributional-based multivariate analysis. These multidimensional measures can adequately operate even with a small number of distributional samples and is well suited for texture matching. Individual texture samples were extracted from the images of the OUTex and the Photometric texture databases, by partitioning the image into regions of almost homogenous texture content. The intrinsic dimensionality of the texture regions is computed by means of image decomposition implementing some of the well-established techniques. Depending on the technique, the joint spectral distribution is sampled in the form of multivariate vectors. The efficiency in textural feature extraction of the different methods, as well as the competence of the above measures in distributional texture image representations, was tested with quite satisfactory results, which yield future ideas for research and application. In addition, a neural-network based vector quantizer was adopted in order to further boost the classification accuracy of the introduced methodologies. Each texture image distribution was summarized by a vector quantization scheme in order to select a restricted number of prototype code vectors, thus resulting in a sampled representation of the original spectral distribution. These code vectors played the role of a spectral signature for each texture image, capturing its basic structure and providing a sparse-compact representation. As a scheduled extension of our work, the application of the introduced methodology to color or multispectral images for texture classification can be straightforwardly implemented, by placing other subband's texture information to higher dimensions in the feature space. However, special care has to be taken to the number of extracted feature vectors in the case of multispectral images, due to the "curse of dimensionality" (Costa & Hero, 2004). In addition, more has to be done in order to overcome the problem that arises from the weakness to capture texture patterns that have different plain scales. This might be accomplished by utilizing adaptive scalable blocks, by means of ? each time ? different sized block patches inside the input images. In this way, the texture information can be efficiently captured in different scales, making use of possible regions of interest. Moreover, texture segmentation algorithms could be incorporated towards this solution, among the plethora that is available in the literature the recent years (Jain & Farrokhnia, 1991; Manjunath & Chellappa, 1991). Finally, the use of alternative ways to extract texture information based on high-level features (i.e., semantic-based texture attributes), so as to correlate better with the human perceptual inspection, is of crucial importance (Liu et al., 2007), as well as the potential application of combining other primitives in the feature extraction methodology, such as color and shape.

In the present study, an efficient strategy for retrieving texture images from large texture databases is introduced and studied within a distributional-statistical framework. Our approach incorporates the multivariate Wald-Wolfowitz test (WW-test), a non-parametric statistical test that measures the similarity between two different sets of multivariate data, which is utilized here for comparing texture distributions. By summarizing the texture information using standard feature extraction methodologies, the similarity measure provides a comprehensive estimate of the match between different images based on graph theory. The proposed "distributional metric" is shown to handle efficiently the texture space dimensionality and the limited sample size drawn from a given image. The experimental results, from the application on a typical texture database, clearly demonstrate the effectiveness of our approach over other texture distribution (dis)similarity metrics. In addition, its performance is used to evaluate several approaches for texture representation. Even though the classification results are obtained on grayscale images, a direct extension to color-based ones can be straightforward.

The problem of image similarity has become a challenging task in the field of computer vision through the last two decades. The assessment of (dis)similarity between color (or multichannel, in general) images or parts of images has been studied on several image processing application domains such as image indexing and retrieval, classification and unsupervised segmentation (Rubner et al., 2001). The basic operations that need to be carried out in order to estimate the similarity between two color images are three-fold (Stricker & Orengo, 1995): first, choose an appropriate color space for image representation; then, extract a signature for each image (using, commonly, low-level features) to construct a theoretically valid distribution; finally, establish pairwise comparisons based on these signatures. Each signature constitutes the content description of a corresponding image. It is summarized based on pixel attributes and provides a representation of the image in a multidimensional feature space. There, a proper (dis)similarity measure is defined in order to act as a general rule for comparing any given pair of images. In these directions, several (dis)similarity measures have been developed and used as empirical estimates of the distribution of image features, confirming that distribution-based measures exhibit excellent performance in all areas (Rubner et al., 2001). In the context of visual image similarity, we make use of a nonparametric test from the field of multivariate statistics that deals with the “Multivariate Two-Sample Problem”, originally presented by Friedman and Rafsky (1979). The specific test is a multivariate extension of the classical Wald-Wolfowitz test (WW-test) and compares two different samples of vectorial observations (i.e. two sets of points in

This paper proposes a methodology that incorporates principles from cluster analysis and graph representation to achieve efficient image segmentation results. More specifically, a feature-based, inter-region dissimilarity relation is considered here in order to determine the dissimilarity matrix in a graph-based segmentation scheme. The calculation of the dissimilarity function between adjacent elementary image regions is based on the proximity of each region's feature vector to the main clusters that are formed by the image samples in the feature space. In contrast to typical segmentation approaches of the literature, the global feature space information is included in the spatial graph representation that was derived from the initial Watershed partitioning. A region grouping process is applied next to form the final segmentation results. The proposed approach was also compared to approaches that use feature-based, or spatial information exclusively, to indicate its effectiveness.

Autoimmune diseases are strictly connected with the presence of autoantibodies in patient serum. Detection of Antinucleolar Antibodies (ANAs) in patient serum is performed using a laboratory technique named Indirect Immunofluorescence (IIF) followed by manual evaluation on the acquired slides from specialized personnel. In this procedure, several limitations appear and several automatic techniques have been proposed for the task of ANA detection. In this work we present a method achieving state-of-the-art performance on a publicly available dataset. More precisely, two powerful and rotation invariant descriptors are incorporated into a two stage classification scheme where the feature vectors are represented and fused in the dissimilarity space. Then, in a second level dissimilarity vectors are classified using a linear SVM classifier. Evaluation on the HEp-2 cell contest dataset yields a 70.16% performance on cell-level classification. Furthermore we provide results in Image Level Classification where a 78.57% classification rate was achieved.

The ensemble average of pattern shift visual evoked potentials (PSVEPs) signals is seriously affected by random latency variations encountered in each individual sweep which is modeled as a continuous signal with linear segments and well-shaped triangular peaks. This effect is causing the smoothed peaks of the averaged PSVEP waveforms. It is the authors objective to restore the degraded peaks and provide accurate information about their exact location. The method used is based on nonlinear filtering of the FIR-median hybrid (FMH) type and is recommended as a postfiltering process to the well-known averaging methods of recovering PSVEP signals from noise by time-locking to stimuli. The new technique, tested in signals from clinical observations, has proven very promising. >

In this work, a feature extraction method is presented for handwritten signature verification. The proposed algorithm models the handwritten elements of a signature trace by probabilistically counting the distribution of fixed two- and three-step pixel paths, conditioned that they are confined within predetermined Chebyshev distances of two and three, respectively. This representation correlates the pixel transitions along the signature trace, with the writing style of an individual. Various partitions of the signature image into a group of sub-images were applied in order to define the overall dimensionality of the feature. In order to evaluate the classification efficiency of the introduced method, a number of verification strategies are implemented by making use of two internationally accepted and one domestic datasets. In all schemes, similarity scores and hard margin support vector machines (SVMs) are combined or evaluated as separate entities. Additionally, zoning the extracted feature vector into combinations of tetrads and heptads, which in turn are fed into the afore-mentioned classification schemes, is exploited. Results, derived from random or simple imitations as well as simulated (skilled) forgery indicate that the proposed method achieves noticeably low equal error rates and it is expected to provide a powerful discriminative representation of the handwritten signature.