Search

Abstract: Cluster validity indexes have been used to evaluate the fitness of partitions produced by clustering algorithms. This paper presents a new validity index for fuzzy clustering called a partition coefficient and exponential separation (PCAES) index. It uses the factors from a normalized partition coefficient and an exponential separation measure for each cluster and then pools these two factors to create the PCAES validity index. Considerations involving the compactness and separation measures for each cluster provide different cluster validity merits. In this paper, we also discuss the problem that the validity indexes face in a noisy environment. The efficiency of the proposed PCAES index is compared with several popular validity indexes. More information about these indexes is acquired in series of numerical comparisons and also three real data sets of Iris, Glass and Vowel. The results of comparative study show that the proposed PCAES index has high ability in producing a good cluster number estimate and in addition, it provides a new point of view for cluster validity in a noisy environment. [Copyright &y& Elsevier]

The expectation of higher accuracy in recognition systems brings the problem of higher complexity. In this paper we introduce a novel Incremental Similarity (IS) that maintains high accuracy while preserving low complexity. We apply IS to on-line and incremental learning tasks, where the need of low complexity is of significant need. Using IS enables the system to directly compute with the samples themselves and update only few parameters in an incremental manner. We empirically prove its efficiency on several evolving models and show that by using IS they achieve competitive results and outperform the baseline models. We also consider the problem of incremental learning used to handle fast growing datasets. We present a very detailed comparison for not only evolving models, but also for the well-known batch models, showing the robustness of our proposal. We perform the evaluation on various classification problems to show the wide application of evolving models and our proposed IS. [ABSTRACT FROM AUTHOR]

This paper considers the general problem of image classification without using any prior knowledge about image classes. We study variants of a method based on supervised learning whose common steps are the extraction of random subwindows described by raw pixel intensity values and the use of ensemble of extremely randomized trees to directly classify images or to learn image features. The influence of method parameters and variants is thoroughly evaluated so as to provide baselines and guidelines for future studies. Detailed results are provided on 80 publicly available datasets that depict very diverse types of images (more than 3800 image classes and over 1.5 million images). [ABSTRACT FROM AUTHOR]

By definition, local image features provide a compact representation of the image in which most of the image information is preserved. This capability offered by local features has been overlooked, despite being relevant in many application scenarios. In this paper, we analyze and discuss the performance of feature-driven Maximally Stable Extremal Regions (MSER) in terms of the coverage of informative image parts (completeness). This type of features results from an MSER extraction on saliency maps in which features related to objects boundaries or even symmetry axes are highlighted. These maps are intended to be suitable domains for MSER detection, allowing this detector to provide a better coverage of informative image parts. Our experimental results, which were based on a large-scale evaluation, show that feature-driven MSER have relatively high completeness values and provide more complete sets than a traditional MSER detection even when sets of similar cardinality are considered. [ABSTRACT FROM AUTHOR]

Abstract: In this paper, a planar shape recognition system is proposed. This proposal is based on a global incremental scheme which combines two learning mechanisms: the Incremental Non-parametric Discriminant Analysis and the mode analysis method. At the feature selection stage, a novel adaptive curvature estimator for shape characterization is presented. This method describes the planar shape using an affine-invariant triangle-area representation obtained from its closed contour. Contrary to previous approaches, the triangle side lengths at each contour point are adapted to the local variations of the shape, removing noise from the contour without missing relevant points. In order to reduce the dimensionality of the shape descriptor, an Incremental Non-parametric Discriminant Analysis is conducted to seek directions for efficient discrimination (incremental eigenspace learning). At the classification stage, the incremental mode analysis is employed to classify feature vectors into a set of spherically-shaped groups (incremental prototype learning). The classification is conducted based on the k-nearest neighbor approach whose prototypes are updated by the mode analysis method. This scheme enables a classifier to learn incrementally, on-line, and in one-pass. Experimental results show that the proposed shape recognition system is well suited for shape indexing and retrieval. [Copyright &y& Elsevier]

Abstract: This paper deals with the regions matching problem of a pair of color stereoscopic images. The proposed approach is carried out in two main steps. The first step selects the candidate matches, which meet the relative position and the overlapping constraints. The second one finds, among these candidates, the correct matches according to a cost function which we define. To guarantee a great similarity between corresponding regions, we propose a new color-based cost function to disambiguate false matches. The cost function take into account the local property at the vicinity of region boundaries. The method is tested on synthesized and real stereo images and the results are satisfactory. [Copyright &y& Elsevier]

Abstract: This paper addresses the sparse view matching problem where the camera parameters lie within ranges depending on the sensor used. An approach, based on homographic transformation, is proposed. The operation is split into two phases. The first phase deals with matches on the ground surface, which is considered planar. The second phase detects matches on arbitrary planes. This is done by detecting junctions of different shapes and reconstructing planes inferred by them. Two versions of that approach are suggested based on the sum of absolute differences and the variance normalized correlation techniques. The first technique is computationally inexpensive while the later is more robust to changes in lighting condition between views. Experiments show that our approach outperforms non-homographic correlation. [Copyright &y& Elsevier]

Abstract: The singular points, core and delta, are widely used in fingerprint classification. However a true pair of core and delta that are close to one another is often ignored. In this paper, we define a new type of singular point denoted by S CD for representing a pair of core and delta. A new algorithm based on the distribution of Gaussian–Hermite moments is used to detect S CD. With core, delta and S CD, the accuracy of fingerprint classification is improved, especially for tented arches. The proposed method has been tested on the NIST-4. We can improve the accuracy of algorithm (Zhang and Yan, 2004) [Zhang, Q., Yan, H., 2004. Fingerprint classification based on extraction and analysis of singularities and pseudo ridges. Pattern Recognit. 37, 2233–2243] by 26.7% for identifying tented arch, and the classification accuracy can be improved by 4.3% for five-class problem. [Copyright &y& Elsevier]

Abstract: This work addresses the problem of 3D models retrieval and recognition using two-dimensional shape representation of 3D objects. However, the human perception of shapes is based on visual parts of objects, where a single significant visual part is sufficient to recognize the whole object. In this paper we present a shape similarity system based on the correspondence of visual 2D parts. These parts are obtained by a shape segmentation approach using the Curvature Scale Space (CSS) descriptor in order to solve scale problems. We propose to combine this partial search method with a probabilistic approach. Finally, we propose a 3D search engine based on 3D models characteristic views and a probabilistic Bayesian voting approach. [Copyright &y& Elsevier]

Abstract: The paper presents a sample-size adaptive SOM (SA-SOM) algorithm for color quantization of images to adapt to the variations of network parameters and training sample size. The sweep size of neighborhood function is modulated by the size of the training data. In addition, the minimax distortion principle which is modulated by training sample size is used to search winning neuron. Based on the SA-SOM, we use the sampling ratio of training data, rather than the conventional weight change between adjacent sweeps, as a stop criterion, to significantly speed up the learning process. The experimental results show that the SA-SOM achieves much better PSNR quality, and smaller PSNR variation under various combinations of network parameters. [Copyright &y& Elsevier]

The visual detection and recognition of objects is facilitated by context. This paper studies two types of learning methods for realizing context-based object detection in paintings. The first method is called the gradient method; it learns to transform the spatial context into a gradient towards the object. The second method, the context-detection method, learns to detect image regions that are likely to contain objects. The accuracy and speed of both methods are evaluated on a face-detection task involving natural and painted faces in a wide variety of contexts. The experimental results show that the gradient method enhances accuracy at the cost of computational speed, whereas the context-detection method optimises speed at the cost of accuracy. The different results of both methods are argued to arise from the different ways in which the methods trade-off accuracy and speed. We conclude that both the gradient method and the context-detection method can be applied to reliable and fast object detection in paintings and that the choice for either method depends on the application and user constraints. [Copyright &y& Elsevier]

Abstract: This paper presented a hybrid optimal estimation algorithm for solving multi-level thresholding problems in image segmentation. The distribution of image intensity is modeled as a random variable, which is approximated by a mixture Gaussian model. The Gaussian’s parameter estimates are iteratively computed by using the proposed PSO+EM algorithm, which consists of two main components: (i) global search by using particle swarm optimization (PSO); (ii) the best particle is updated through expectation maximization (EM) which leads the remaining particles to seek optimal solution in search space. In the PSO+EM algorithm, the parameter estimates fed into EM procedure are obtained from global search performed by PSO, expecting to provide a suitable starting point for EM while fitting the mixture Gaussians model. The preliminary experimental results show that the hybrid PSO+EM algorithm could solve the multi-level thresholding problem quite swiftly, and also provide quality thresholding outputs for complex images. [Copyright &y& Elsevier]

Abstract: Multi-class AdaBoost algorithms AdaBooost.MO, -ECC and -OC have received a great attention in the literature, but their relationships have not been fully examined to date. In this paper, we present a novel interpretation of the three algorithms, by showing that MO and ECC perform stage-wise functional gradient descent on a cost function defined over margin values, and that OC is a shrinkage version of ECC. This allows us to strictly explain the properties of ECC and OC, empirically observed in prior work. Also, the outlined interpretation leads us to introduce shrinkage as regularization in MO and ECC, and thus to derive two new algorithms: SMO and SECC. Experiments on diverse databases are performed. The results demonstrate the effectiveness of the proposed algorithms and validate our theoretical findings. [Copyright &y& Elsevier]

Abstract: In this paper, we present the expert systems for detecting variability of electroencephalogram (EEG) signals. In practical applications of pattern recognition, there are often diverse features extracted from raw data which needs recognizing. Methods of combining multiple classifiers with diverse features are viewed as a general problem in various application areas of pattern recognition. Because of the importance of making the right decision, we are looking for better classification procedures for EEG signals. The mixture of experts (ME) and modified mixture of experts (MME) were tested and benchmarked for their performance on the classification of the studied EEG signals. Decision making was performed in two stages: feature extraction by eigenvector methods and classification using the classifiers trained on the extracted features. The inputs of these expert systems composed of diverse or composite features were chosen according to the network structures. The present study was conducted with the purpose of answering the question of whether the expert system with diverse features (MME) or composite feature (ME) improve the capability of classification of the EEG signals. Our research demonstrated that the MME trained on diverse features achieved accuracy rates which were higher than that of the ME. [Copyright &y& Elsevier]

Abstract: In this paper, we present an algorithm based on equal errors principle, which solves the general version of polygonal approximation problem (GPA). The approximation nodes of GPA can be selected anywhere on the original polygonal curve, while the “classical” (usually used) polygonal approximation problem (PA) demands the vertices to be a subset of the original polygon vertices. Although the proposed algorithm does not guarantee global minimum distortion, in many cases almost optimal solutions are achieved. Moreover, it is very flexible on changes of error criteria and on curve dimension yielding an alternative and in many cases approximations with lower error (by relaxing the simplification constraint) than the optimal PA methods with about the same computation cost. [Copyright &y& Elsevier]

Abstract: Feature selection is a crucial step in pattern recognition. Most feature selection algorithms reported are developed for continuous features. In this paper, we propose a feature selection algorithm for mixed-typed data containing both continuous and nominal features. The algorithm consists of a novel criterion for mixed feature subset evaluation and a novel search algorithm for mixed feature subset generation. The proposed feature selection algorithm is tested using both artificial and real-world problems. [Copyright &y& Elsevier]

Abstract: In this paper, a novel document skew detection algorithm based on wavelet decompositions and projection profile analysis is proposed. First, the skewed document images are decomposed by the wavelet transform. The matrix containing the absolute values of the horizontal sub-band coefficients, which preserves the text’s horizontal structure, is then rotated through a range of angles. A projection profile is computed at each angle, and the angle that maximizes a criterion function is regarded as the skew angle. Experimental results show that this algorithm performs well on document images of various layouts and is also robust to different languages. The effects of various wavelet basis, number of decomposition levels, and parameters of the criterion function are investigated too. [Copyright &y& Elsevier]

Abstract: An intrinsic property of real aperture imaging has been that the observations tend to be defocused. This artifact has been used in an innovative manner by researchers for depth estimation, since the amount of defocus varies with varying depth in the scene. There have been various methods to model the defocus blur. We model the defocus process using the model of diffusion of heat. The diffusion process has been traditionally used in low level vision problems like smoothing, segmentation and edge detection. In this paper a novel application of the diffusion principle is made for generating the defocus space of the scene. The defocus space is the set of all possible observations for a given scene that can be captured using a physical lens system. Using the notion of defocus space we estimate the depth in the scene and also generate the corresponding fully focused equivalent pin-hole image. The algorithm described here also brings out the equivalence of the two modalities, viz. depth from focus and depth from defocus for structure recovery. [Copyright &y& Elsevier]

Abstract: This paper addresses a fundamental limitation of the existing distortion measures embedded in vertex-based operational-rate-distortion shape coding techniques by introducing a new accurate distortion measurement algorithm based upon the actual distance, rather than either the shortest absolute distance or a distortion band. [Copyright &y& Elsevier]

Abstract: In this paper, inverted coordinates image space (ICIS) is proposed as a solution for the problem of the unbounded accumulator space in the automatic detection of the finite/infinite vanishing points in image space. Since the ICIS is based on the direct transformation from the image space, it does not lose any geometrical information from the original image and it does not require camera calibration as opposed to the Gaussian sphere based methods. Moreover, the proposed method can accurately detect both the finite and infinite vanishing points under a small fixed memory amount as opposed to the conventional image space based methods. Experiments are conducted on various real images in architectural environments to show the advantages of the proposed approach over conventional methods. [Copyright &y& Elsevier]

Abstract: Most adaptive smoothing approaches damage image fidelity. In this paper we propose to improve the structure-adaptive anisotropic filtering approach using an elliptical kernel, a non-linear filtering function, and a more robust-to-noise technique for oriented pattern direction estimation. The proposed filter outperforms some others known filters and is more robust to noisy images. [Copyright &y& Elsevier]

Abstract: This paper deals with recognition of known 3-D objects in different orientations on cluttered backgrounds. As a recognition technique we apply support vector machines (SVMs). To cope with the cluttered background a tree structure of masks is introduced for each object. SVMs are then computed by masking the training sets with the appropriate masks. One- and two-class SVMs are combined in the recognition process. One-class SVMs, used at the first stage, allow us to avoid the “non-object” class generation usually required to classify unknown objects or other parts of a scene. Two-class SVMs are further applied to resolve the recognition process when necessary. The proposed method is compared with two other approaches and as demonstrated by experimental results it is robust to cluttered backgrounds. The advantage of the method is its ability to classify the pattern as unknown which has a valuable effect on false positive rate. [Copyright &y& Elsevier]

Abstract: This paper presents a novel shape representation algorithm based on mathematical morphology. It consists of two steps. Firstly, an input shape is decomposed into a union of meaningful convex subparts by a recursive scheme. Each subpart is obtained by repeatedly applying condition expansion to a seed, which is selected by utilizing the skeleton information. Secondly, the shape of each subpart is approximated by a morphological dilation of basic structuring elements. The location and direction of the subpart are represented respectively by two parameters. Thus the given shape is represented by a union set of a number of three-dimensional vectors. Experiments show that the new algorithm is immune to noise and occlusion, and invariant under rotation, translation and scaling. Compared to other algorithms, it achieves more natural looking shape components and more concise representation at lower computation costs and coding costs. [Copyright &y& Elsevier]

Abstract: Identity verification is one of the critical issues in the sector of security and has been emerging as an active research area. In recent years, technologies using biological features to address problems of identity verification have attracted numerous research interests. For examples, fingerprint recognition, voice recognition and pattern of blood vessels in the retina have spanned many commercial applications. However, special and expensive equipments such as fingerprint readers and iris scanners are often required and people have to be in unpleasant poses occasionally. This paper presents a study on computer vision technique and its application in face recognition to achieve identity verification. With multiple facial images taken from different view angles, relative affine structures are computed and are used as measurements. To that end, the explicit relationship between relative affine structure and the cross ratio which is a view-invariant under perspective projection is also addressed. The proposed method neither requires camera calibration nor reconstructs 3D models. According to simulation results, the developed approach can achieve satisfactory results given the feature points of facial images. [Copyright &y& Elsevier]

Abstract: This paper proposes a new scheme based on independent component analysis (ICA) for object recognition with affine transformation and for affine motion estimation between video frames. For different affine shapes of a recognized object, an invariant descriptor can be extracted by ICA, and it can solve some object recognition problems. This method also can be used to estimate the affine motion between two frames, which is important in high compression rate coding such as MPEG4 or MPEG7 standard. Simulation results show that the proposed method has a better performance than other traditional methods in pattern recognition and affine motion estimation. [Copyright &y& Elsevier]

Abstract: This paper presents an adaptive classification method that utilizes ellipsoidal regions for multidimensional pattern classification problems with continuous input variables. The classification method fits a finite number of the ellipsoidal regions to data pattern by using adaptive operations iteratively. The method adaptively expands, rotates, shrinks, and/or moves the ellipsoidal regions while each ellipsoidal region is separately handled with a fitness value assigned. The adaptation procedure is combined with a variable selection process in the outer loop, where significant input variables for the ellipsoids are determined by using a stepwise selection method. The performance of the method is evaluated on well-known classification problems from the UCI machine learning repository. The evaluation result shows that the proposed method can exert equivalent or superior performance, with smaller number of rules, to other classification methods such as fuzzy rules, decision trees, or neural networks. [Copyright &y& Elsevier]

Abstract: Most intrusion detection system (IDS) with a single-level structure can only detect either misuse or anomaly attacks. Some IDSs with multi-level structure or multi-classifier are proposed to detect both attacks, but they are limited in adaptively learning. In this paper, two hierarchical IDS frameworks using Radial Basis Functions (RBF) are proposed. A serial hierarchical IDS (SHIDS) is proposed to identify misuse attack accurately and anomaly attacks adaptively. A parallel hierarchical IDS (PHIDS) is proposed to enhance the SHIDS’s functionalities and performance. The experiments show that the two proposed IDSs can detect network intrusions in real-time, train new classifiers for novel intrusions automatically, and modify their structures adaptively after new classifiers are trained. [Copyright &y& Elsevier]

Abstract: Multidimensional scaling (MDS) is a statistical tool for constructing a low-dimensional configuration to represent the relationships between the objects. Although MDS has been widely used in various fields, it is difficult to evaluate similarity/ dissimilarity between the complex systems by human judgment. Even though we can divide a complex system into subsystems, which can be more easily evaluated, the relative weights of the subsystems are also a crucial problem. Because of these subsystems usually exist interdependence and feedback, the weights of the subsystems are hard to obtain. This paper proposes a method which combines the methods of the interpretive structural modeling (ISM) and the analytic network process (ANP) procedures to deal with the problem of the subsystems’ interdependence and feedback. In addition, we also provide a numerical example to illustrate the proposed method. [Copyright &y& Elsevier]

Abstract: If the training pattern set is large, it takes a large memory and a long time to train support vector machine (SVM). Recently, we proposed neighborhood property based pattern selection algorithm (NPPS) which selects only the patterns that are likely to be near the decision boundary ahead of SVM training [Proc. of the 7th Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD), Lecture Notes in Artificial Intelligence (LNAI 2637), Seoul, Korea, pp. 376–387]. NPPS tries to identify those patterns that are likely to become support vectors in feature space. Preliminary reports show its effectiveness: SVM training time was reduced by two orders of magnitude with almost no loss in accuracy for various datasets. It has to be noted, however, that decision boundary of SVM and support vectors are all defined in feature space while NPPS described above operates in input space. If neighborhood relation in input space is not preserved in feature space, NPPS may not always be effective. In this paper, we show that the neighborhood relation is invariant under input to feature space mapping. The result assures that the patterns selected by NPPS in input space are likely to be located near decision boundary in feature space. [Copyright &y& Elsevier]

Abstract: In this paper we compare the accuracy of a range of advanced density models for gas identification from sensor array signals. Density estimation is applied in the construction of classifiers through the use of Bayes rule. Experiments on real sensors’ data proved the effectiveness of the approach with an excellent classification performance. We compare the classification accuracy of four density models, Gaussian mixture models, Generative topographic mapping, Probabilistic PCA mixture and K nearest neighbors. On our gas sensors data, the best performance was achieved by Gaussian mixture models. [Copyright &y& Elsevier]

Abstract: Recently a large number of studies has been published in the area of Fractal Analysis. In this paper we review briefly the IFS (Iterated Function System) theory, and we show how this theoretical tool leads to new applications in Pattern Recognition and more precisely in the analysis of handwritten texts. Among other results, this research work shows that fractal analysis is no longer only restricted to a descriptive role, but has entered in a functional phase. More precisely we developed a concrete application, based on automatic feature selection by fractal approach, and dedicated to alphabet discrimination. [Copyright &y& Elsevier]

This paper presents a new method for similarity measures between intuitionistic fuzzy sets (IFSs). We will present a method to calculate the distance between IFSs on the basis of the Hausdorff distance. We will then use this distance to generate a new similarity measure to calculate the degree of similarity between IFSs. Finally we will prove some properties of the proposed similarity measure and use several examples to compare the proposed similarity measure with existing methods. Numerical results show that the proposed similarity measure is much simpler than existing methods and is well suited to use with linguistic variables. [Copyright &y& Elsevier]

In this paper, a new method to measure the degree of similarity between generalized fuzzy numbers is presented. Eighteen sets of generalized fuzzy numbers are used to compare the proposed method with the existing similarity measures. The results show that the new similarity measure can overcome the drawbacks of the existing methods. Finally, the proposed similarity measure is applied to pattern recognition. [Copyright &y& Elsevier]