Face recognition using an NNSRM classifier in LDA subspace

NNSRM is an implementation of the structural risk minimization (SRM) principle using the nearest neighbor (NN) rule, and linear discriminant analysis (LDA) is a dimension-reducing method, which is usually used in classifications. This paper combines the two methods for face recognition. We first project the face images into a PCA subspace, then project the results into a much lower-dimensional LDA subspace, and then use an NNSRM classifier to recognize them in the LDA subspace. Experimental results demonstrate that the combined method can achieve a better performance than NN by selecting different distances and a comparable performance with SVM but costing less computational time.

Semi-supervised discriminative classification with application to tumorous tissues segmentation of MR brain images

Due to the large data size of 3D MR brain images and the blurry boundary of the pathological tissues, tumor segmentation work is difficult. This paper introduces a discriminative classification algorithm for semi-automated segmentation of brain tumorous tissues. The classifier uses interactive hints to obtain models to classify normal and tumor tissues. A non-parametric Bayesian Gaussian random field in the semi-supervised mode is implemented. Our approach uses both labeled data and a subset of unlabeled data sampling from 2D/3D images for training the model. Fast algorithm is also developed. Experiments show that our approach produces satisfactory segmentation results comparing to the manually labeled results by experts.

Quadtree-based eigendecomposition for pose estimation in the presence of occlusion and background clutter

Eigendecomposition-based techniques are popular for a number of computer vision problems, e.g., object and pose estimation, because they are purely appearance based and they require few on-line computations. Unfortunately, they also typically require an unobstructed view of the object whose pose is being detected. The presence of occlusion and background clutter precludes the use of the normalizations that are typically applied and significantly alters the appearance of the object under detection. This work presents an algorithm that is based on applying eigendecomposition to a quadtree representation of the image dataset used to describe the appearance of an object. This allows decisions concerning the pose of an object to be based on only those portions of the image in which the algorithm has determined that the object is not occluded. The accuracy and computational efficiency of the proposed approach is evaluated on 16 different objects with up to 50% of the object being occluded and on images of ships in a dockyard.

Human action recognition in videos based on the Transferable Belief Model

This paper focuses on human behavior recognition where the main problem is to bridge the semantic gap between the analogue observations of the real world and the symbolic world of human interpretation. For that, a fusion architecture based on the Transferable Belief Model framework is proposed and applied to action recognition of an athlete in video sequences of athletics meeting with moving camera. Relevant features are extracted from videos, based on both the camera motion analysis and the tracking of particular points on the athlete’s silhouette. Some models of interpretation are used to link the numerical features to the symbols to be recognized, which are running, jumping and falling actions. A Temporal Belief Filter is then used to improve the robustness of action recognition. The proposed approach demonstrates good performance when tested on real videos of athletics sports videos (high jumps, pole vaults, triple jumps and long jumps) acquired by a moving camera and different view angles. The proposed system is also compared to Bayesian Networks.

Hierarchical clustering of subpopulations with a dissimilarity based on the likelihood ratio statistic: application to clustering massive data sets

The problem of clustering subpopulations on the basis of samples is considered within a statistical framework: a distribution for the variables is assumed for each subpopulation and the dissimilarity between any two populations is defined as the likelihood ratio statistic which compares the hypothesis that the two subpopulations differ in the parameter of their distributions to the hypothesis that they do not. A general algorithm for the construction of a hierarchical classification is described which has the important property of not having inversions in the dendrogram. The essential elements of the algorithm are specified for the case of well-known distributions (normal, multinomial and Poisson) and an outline of the general parametric case is also discussed. Several applications are discussed, the main one being a novel approach to dealing with massive data in the context of a two-step approach. After clustering the data in a reasonable number of ‘bins’ by a fast algorithm such as k-Means, we apply a version of our algorithm to the resulting bins. Multivariate normality for the means calculated on each bin is assumed: this is justified by the central limit theorem and the assumption that each bin contains a large number of units, an assumption generally justified when dealing with truly massive data such as currently found in modern data analysis. However, no assumption is made about the data generating distribution.

Reverse engineering the way humans rank textures

We argue that in order to understand which features are used by humans to group textures, one must start by computing thousands of features of diverse nature, and select from those features those that allow the reproduction of perceptual groups or perceptual ranking created by humans. We use the Trace transform to produce such features here. We compare these features with those produced from the co-occurrence matrix and its variations. We show that when one is not interested in reproducing human behaviour, the elements of the co-occurrence matrix used as features perform best in terms of texture classification accuracy. However, these features cannot be “trained” or “selected” to imitate human ranking, while the features produced from the Trace transform can. We attribute this to the diverse nature of the features computed from the Trace transform.

A multi-stage classification system for detecting intrusions in computer networks

A serial multi-stage classification system for facing the problem of intrusion detection in computer networks is proposed. The whole decision process is organized into successive stages, each one using a set of features tailored for recognizing a specific attack category. All the stages employ suitable criteria for estimating the reliability of the performed classification, so that, in case of uncertainty, information related to a possible attack are only logged for further processing, without raising an alert for the system manager. This permits to reduce the number of false alarms. On the other hand, in order to keep low the number of missed detections, the proposed system declares a connection as normal traffic only if all the stages do not detect an attack. The proposed multi-stage intrusion detection system has been tested on three different services (http, telnet and ftp) of a standard database used for benchmarking intrusion detection systems and also on real network traffic data. The experimental analysis highlights the effectiveness of the approach: the proposed system behaves significantly better than other multiple classifier systems performing classification in a single stage.

Multi-level classification of emphysema in HRCT lung images

Emphysema is a common chronic respiratory disorder characterised by the destruction of lung tissue. It is a progressive disease where the early stages are characterised by a diffuse appearance of small air spaces, and later stages exhibit large air spaces called bullae. A bullous region is a sharply demarcated region of emphysema. In this paper, it is shown that an automated texture-based system based on co-training is capable of achieving multiple levels of emphysema extraction in high-resolution computed tomography (HRCT) images. Co-training is a semi-supervised technique used to improve classifiers that are trained with very few labelled examples using a large pool of unseen examples over two disjoint feature sets called views. It is also shown that examples labelled by experts can be incorporated within the system in an incremental manner. The results are also compared against “density mask”, currently a standard approach used for emphysema detection in medical image analysis and other computerized techniques used for classification of emphysema in the literature. The new system can classify diffuse regions of emphysema starting from a bullous setting. The classifiers built at different iterations also appear to show an interesting correlation with different levels of emphysema, which deserves more exploration.

---

Motion estimation based color transfer and its application to color video compression

In this paper a novel scheme for color video compression using color transfer technique is proposed. Towards this, a new color transfer mechanism for video using motion estimation is presented. Encoder and decoder architectures for the proposed compression scheme are also presented. In this scheme, compression is achieved by firstly discarding chrominance information for all but selected reference frames and then using motion prediction and discrete cosine transform (DCT) based quantization. At decompression stage, the luminance-only frames are colored using chrominance information from the reference frames applying the proposed color transfer technique. To integrate color transfer mechanism with hybrid compression scheme a new color transfer protocol is defined. Both compression scheme and color transfer work in YCbCr color space.

Optimizing two-pass connected-component labeling algorithms

We present two optimization strategies to improve connected-component labeling algorithms. Taking together, they form an efficient two-pass labeling algorithm that is fast and theoretically optimal. The first optimization strategy reduces the number of neighboring pixels accessed through the use of a decision tree, and the second one streamlines the union-find algorithms used to track equivalent labels. We show that the first strategy reduces the average number of neighbors accessed by a factor of about 2. We prove our streamlined union-find algorithms have the same theoretical optimality as the more sophisticated ones in literature. This result generalizes an earlier one on using union-find in labeling algorithms by Fiorio and Gustedt (Theor Comput Sci 154(2):165–181, 1996). In tests, the new union-find algorithms improve a labeling algorithm by a factor of 4 or more. Through analyses and experiments, we demonstrate that our new two-pass labeling algorithm scales linearly with the number of pixels in the image, which is optimal in computational complexity theory. Furthermore, the new labeling algorithm outperforms the published labeling algorithms irrespective of test platforms. In comparing with the fastest known labeling algorithm for two-dimensional (2D) binary images called contour tracing algorithm, our new labeling algorithm is up to ten times faster than the contour tracing program distributed by the original authors.

Distributional-based texture classification using non-parametric statistics

Texture classification is an important problem in image analysis. In the present study, an efficient strategy for classifying texture images is introduced and examined 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 and its superiority over other well-established 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 aspect Bernoulli model: multiple causes of presences and absences

We present a probabilistic multiple cause model for the analysis of binary (0–1) data. A distinctive feature of the aspect Bernoulli (AB) model is its ability to automatically detect and distinguish between “true absences” and “false absences” (both of which are coded as 0 in the data), and similarly, between “true presences” and “false presences” (both of which are coded as 1). This is accomplished by specific additive noise components which explicitly account for such non-content bearing causes. The AB model is thus suitable for noise removal and data explanatory purposes, including omission/addition detection. An important application of AB that we demonstrate is data-driven reasoning about palaeontological recordings. Additionally, results on recovering corrupted handwritten digit images and expanding short text documents are also given, and comparisons to other methods are demonstrated and discussed.

Periodicity and stability issues of a <Emphasis Type="Italic">chaotic</Emphasis> pattern recognition neural network

Traditional pattern recognition (PR) systems work with the model that the object to be recognized is characterized by a set of features, which are treated as the inputs. In this paper, we propose a new model for PR, namely one that involves chaotic neural networks (CNNs). To achieve this, we enhance the basic model proposed by Adachi (Neural Netw 10:83–98, 1997), referred to as Adachi’s Neural Network (AdNN), which though dynamic, is not chaotic. We demonstrate that by decreasing the multiplicity of the eigenvalues of the AdNN’s control system, we can effectively drive the system into chaos. We prove this result here by eigenvalue computations and the evaluation of the Lyapunov exponent. With this premise, we then show that such a Modified AdNN (M-AdNN) has the desirable property that it recognizes various input patterns. The way that this PR is achieved is by the system essentially sympathetically “resonating” with a finite periodicity whenever these samples (or their reasonable resemblances) are presented. In this paper, we analyze the M-AdNN for its periodicity, stability and the length of the transient phase of the retrieval process. The M-AdNN has been tested for Adachi’s dataset and for a real-life PR problem involving numerals. We believe that this research also opens a host of new research avenues. Research partially supported by the Natural Sciences and Engineering Research Council of Canada.

---

Quantitatively measuring object-oriented couplings

One key to several quality factors of software is the way components are connected. Software coupling can be used to estimate a number of quality factors, including maintainability, complexity, and reliability. Object-oriented languages are designed to reduce the number of dependencies among classes, which encourages separation of concerns and should reduce the amount of coupling. At the same time, the object-oriented language features change the way the connections are made, how they must be analyzed, and how they are measured. This paper discusses software couplings based on object-oriented relationships between classes, specifically focusing on types of couplings that are not available until after the implementation is completed, and presents a static analysis tool that measures couplings among classes in Java packages. Data from evaluating the tool on several open-source projects are provided. The coupling measurement is based on source code, which has the advantage of being quantitative and more precise than previous measures, but the disadvantage of not being available before implementation, and thus not useful for some predictive efforts.

The Effects of Navigation Sensors and Spatial Road Network Data Quality on the Performance of Map Matching Algorithms

Map matching algorithms are utilised to support the navigation module of advanced transport telematics systems. The objective of this paper is to develop a framework to quantify the effects of spatial road network data and navigation sensor data on the performance of map matching algorithms. Three map matching algorithms are tested with different spatial road network data (map scale 1:1,250; 1:2,500 and 1:50,000) and navigation sensor data (global positioning system (GPS) and GPS augmented with deduced reckoning) in order to quantify their performance. The algorithms are applied to different road networks of varying complexity. The performance of the algorithms is then assessed for a suburban road network using high precision positioning data obtained from GPS carrier phase observables. The results show that there are considerable effects of spatial road network data on the performance of map matching algorithms. For an urban road network, the results suggest that both the quality of spatial road network data and the type of navigation system affect the link identification performance of map matching algorithms.

On detecting faults for Boolean expressions

Fault based testing aims at detecting hypothesized faults based on specifications or program source. There are some fault based techniques for testing Boolean expressions which are commonly used to model conditions in specifications as well as logical decisions in program source. The MUMCUT strategy has been proposed to generate test cases from Boolean expressions. Moreover, it detects eight common types of hypothesized faults provided that the original expression is in irredundant disjunctive normal form, IDNF. Software practitioners are more likely to write the conditions and logical decisions in general form rather than IDNF. Hence, it is interesting to investigate the fault detecting capability of the MUMCUT strategy with respect to general form Boolean expressions. In this article, we perform empirical studies to investigate the fault detection capability of the MUMCUT strategy with respect to general form Boolean expressions as well as mutated expressions. A mutated expression can be obtained from the original given Boolean expression by making a syntactic change based on a particular type of fault.

Development of assessment criteria for clustering algorithms

In this paper, new measures—called clustering performance measures (CPMs)—for assessing the reliability of a clustering algorithm are proposed. These CPMs are defined using a validation measure, which determines how well the algorithm works with a given set of parameter values, and a repeatability measure, which is used for studying the stability of the clustering solutions and has the ability to estimate the correct number of clusters in a dataset. These proposed CPMs can be used to evaluate clustering algorithms that have a structure bias to certain types of data distribution as well as those that have no structure biases. Additionally, we propose a novel cluster validity index, V _I index, which is able to handle non-spherical clusters. Five clustering algorithms on different types of real-world data and synthetic data are evaluated. The first dataset type refers to a communications signal dataset representing one modulation scheme under a variety of noise conditions, the second represents two breast cancer datasets, while the third type represents different synthetic datasets with arbitrarily shaped clusters. Additionally, comparisons with other methods for estimating the number of clusters indicate the applicability and reliability of the proposed cluster validity V _I index and repeatability measure for correct estimation of the number of clusters.

Building measure-based prediction models for UML class diagram maintainability

The usefulness of measures for the analysis and design of object oriented (OO) software is increasingly being recognized in the field of software engineering research. In particular, recognition of the need for early indicators of external quality attributes is increasing. We investigate through experimentation whether a collection of UML class diagram measures could be good predictors of two main subcharacteristics of the maintainability of class diagrams: understandability and modifiability. Results obtained from a controlled experiment and a replica support the idea that useful prediction models for class diagrams understandability and modifiability can be built on the basis of early measures, in particular, measures that capture structural complexity through associations and generalizations. Moreover, these measures seem to be correlated with the subjective perception of the subjects about the complexity of the diagrams. This fact shows, to some extent, that the objective measures capture the same aspects as the subjective ones. However, despite our encouraging findings, further empirical studies, especially using data taken from real projects performed in industrial settings, are needed. Such further study will yield a comprehensive body of knowledge and experience about building prediction models for understandability and modifiability.

Advances in GML for Geospatial Applications

This paper presents a study of Geography Markup Language (GML), the issues that arise from using GML for spatial applications, including storage, parsing, querying and visualization, as well as the use of GML for mobile devices and web services. GML is a modeling language developed by the Open Geospatial Consortium (OGC) as a medium of uniform geographic data storage and exchange among diverse applications. Many new XML-based languages are being developed as open standards in various areas of application. It would be beneficial to integrate such languages with GML during the developmental stages, taking full advantage of a non-proprietary universal standard. As GML is a relatively new language still in development, data processing techniques need to be refined further in order for GML to become a more efficient medium for geospatial applications.