Real-time multiple people tracking using competitive condensation

The CONDENSATION (Conditional Density Propagation) algorithm has a robust tracking performance and suitability for real-time implementation. However, the CONDENSATION tracker has some difficulties with real-time implementation for multiple people tracking since it requires very complicated shape modelling and a large number of samples for precise tracking performance. Further, it shows a poor tracking performance in the case of close or partially occluded people. To overcome these difficulties, we present three improvements: First, we construct effective templates of people's shapes using the SOM (Self-Organizing Map). Second, we take the discrete HMM (Hidden Markov Modelling) for an accurate dynamical model of the people's shape transition. Third, we use the competition rule to separate close or partially occluded people effectively. Simulation results shows that the proposed CONDENSATION algorithm can achieve robust and real-time tracking in the image sequences of a crowd of people.     

Automatic 3d free form shape matching using the graduated assignment algorithm

Three-dimensional free form shape matching is a fundamental problem in both the machine vision and pattern recognition literatures. However, the automatic approach to 3D free form shape matching still remains open. In this paper, we propose using k closest points in the second view for the automatic 3D free form shape matching. For the sake of computational efficiency, the optimised k-D tree is employed for the search of the k closest points. Since occlusion and appearance and disappearance of points almost always occur, slack variables have to be employed, explicitly modelling outliers in the process of matching. Then the relative quality of each possible point match is estimated using the graduated assignment algorithm, leading the camera motion parameters to be estimated by the quaternion method in the weighted least-squares sense. The experimental results based on both synthetic data and real images without any pre-processing show the effectiveness and efficiency of the proposed algorithm for the automatic matching of overlapping 3D free form shapes with either sparse or dense points.     

Face recognition using a fuzzy fisherface classifier

In this study, we are concerned with face recognition using fuzzy fisherface approach and its fuzzy set based augmentation. The well-known fisherface method is relatively insensitive to substantial variations in light direction, face pose, and facial expression. This is accomplished by using both principal component analysis and Fisher's linear discriminant analysis. What makes most of the methods of face recognition (including the fisherface approach) similar is an assumption about the same level of typicality (relevance) of each face to the corresponding class (category). We propose to incorporate a gradual level of assignment to class being regarded as a membership grade with anticipation that such discrimination helps improve classification results. More specifically, when operating on feature vectors resulting from the PCA transformation we complete a Fuzzy K-nearest neighbor class assignment that produces the corresponding degrees of class membership. The comprehensive experiments completed on ORL, Yale, and CNU (Chungbuk National University) face databases show improved classification rates and reduced sensitivity to variations between face images caused by changes in illumination and viewing directions. The performance is compared vis-à-vis other commonly used methods, such as eigenface and fisherface.     

bigVAT: Visual assessment of cluster tendency for large data sets

Assessment of clustering tendency is an important first step in cluster analysis. One tool for assessing cluster tendency is the Visual Assessment of Tendency (VAT) algorithm. VAT produces an image matrix that can be used for visual assessment of cluster tendency in either relational or object data. However, VAT becomes intractable for large data sets. The revised VAT (reVAT) algorithm reduces the number of computations done by VAT, and replaces the image matrix with a set of profile graphs that are used for the visual assessment step. Thus, reVAT overcomes the large data set problem which encumbers VAT, but presents a new problem: interpretation of the set of reVAT profile graphs becomes very difficult when the number of clusters is large, or there is significant overlap between groups of objects in the data. In this paper, we propose a new algorithm called bigVAT which (i) solves the large data problem suffered by VAT, and (ii) solves the interpretation problem suffered by reVAT. bigVAT combines the quasi-ordering technique used by reVAT with an image display of the set of profile graphs displaying the clustering tendency information with a VAT-like image. Several numerical examples are given to illustrate and support the new technique.     

Automatic model selection for the optimization of SVM kernels

This approach aims to optimize the kernel parameters and to efficiently reduce the number of support vectors, so that the generalization error can be reduced drastically. The proposed methodology suggests the use of a new model selection criterion based on the estimation of the probability of error of the SVM classifier. For comparison, we considered two more model selection criteria: GACV (‘Generalized Approximate Cross-Validation’) and VC (‘Vapnik-Chernovenkis’) dimension. These criteria are algebraic estimates of upper bounds of the expected error. For the former, we also propose a new minimization scheme. The experiments conducted on a bi-class problem show that we can adequately choose the SVM hyper-parameters using the empirical error criterion. Moreover, it turns out that the criterion produces a less complex model with fewer support vectors. For multi-class data, the optimization strategy is adapted to the one-against-one data partitioning. The approach is then evaluated on images of handwritten digits from the USPS database.     

Switching class labels to generate classification ensembles

Ensembles that combine the decisions of classifiers generated by using perturbed versions of the training set where the classes of the training examples are randomly switched can produce a significant error reduction, provided that large numbers of units and high class switching rates are used. The classifiers generated by this procedure have statistically uncorrelated errors in the training set. Hence, the ensembles they form exhibit a similar dependence of the training error on ensemble size, independently of the classification problem. In particular, for binary classification problems, the classification performance of the ensemble on the training data can be analysed in terms of a Bernoulli process. Experiments on several UCI datasets demonstrate the improvements in classification accuracy that can be obtained using these class-switching ensembles.     

Identification of semi-physical and black-box non-linear models: the case of MR-dampers for vehicles control

The topic of this paper is the identification of an accurate model for magneto-rheological (MR) dampers. A semi-active MR-damper is a dynamic system, where the inputs are the elongation velocity and the command current; the current is the control input which modulates at high-bandwidth the damping characteristic through the variation of a magnetic field. The output is the force delivered by the damper. Among the broad set of applications where MR-dampers can be used, the results proposed in this work refer to MR-dampers for the control of vehicle dynamics.MR-damper are highly non-linear systems, and their accurate modeling is a non-trivial task. MR-dampers can be modeled using two different model classes: semi-physical models and black-box models. Both approaches are considered in this work.The purpose of this brief paper is to make a concise but complete presentation and discussion of a non-trivial system identification problem. The problem considered herein is particularly interesting from the system identification point of view: from one side, the MR-damper is a very attractive actuator, which is likely to become the key device for many dynamics and vibration control systems in the near future; on the other side, it is an example of an application problem where the accurate modeling of the actuation device is one of the most crucial part of the whole control design problem.     

Structured coprime factor model reduction based on LMIs

In this paper we discuss dynamic model reduction methods which preserve a certain structure in the underlying system. Specifically, we consider the situation where the reduction must be consistent with a partition of the system states. This is motivated, for instance, in situations where state variables are associated with the topology of a networked system, and the reduction should preserve this. We build on the observation that imposing block structure to generalized controllability and observability gramians automatically yields such state-partitioned model reduction. The difficulty lies in ensuring feasibility of the resulting Lyapunov inequalities, which is in general very restrictive. To overcome this, we consider coprime factor model reduction. We derive an LMI characterization of expansive and contractive coprime factorizations that preserve structure, and use this to build a more flexible method for structured model reduction. An example is given to illustrate the method.     

A level-set based variational method for design and optimization of heterogeneous objects

A heterogeneous object is referred to as a solid object made of different constituent materials. The object is of a finite collection of regions of a set of prescribed material classes of continuously varying material properties. These properties have a discontinuous change across the interface of the material regions. In this paper, we propose a level-set based variational approach for the design of this class of heterogeneous objects. Central to the approach is a variational framework for a well-posed formulation of the design problem. In particular, we adapt the Mumford-Shah model which specifies that any point of the object belongs to either of two types: inside a material region of a well-defined gradient or on the boundary edges and surfaces of discontinuities. Furthermore, the set of discontinuities is represented implicitly, using a multi-phase level set model. This level-set based variational approach yields a computational system of coupled geometric evolution and diffusion partial differential equations. Promising features of the proposed method include strong regularity in the problem formulation and inherent capabilities of geometric and material modeling, yielding a common framework for optimization of the heterogeneous objects that incorporates dimension, shape, topology, and material properties. The proposed method is illustrated with several 2D examples of optimal design of multi-material structures and materials.     

On the development of a haptic system for rapid product development

This paper presents a system development that extends haptic modeling to a number of key aspects in product development. Since haptic modeling has been developed based on physical laws, it is anticipated that a natural link between the virtual world and practical applications can be established based on haptic interaction. In the proposed system, a haptic device is used as the central mechanism for reverse engineering, shape modeling, real time mechanical property analysis, machining tool path planning and coordinate measuring machine (CMM) tolerance inspection path planning. With all these features in a single haptic system, it is possible to construct a three dimensional part by either haptic shape modeling or reverse engineering, then performing real-time mechanical property analysis in which the stiffness of a part can be felt and intuitively evaluated by the user, or generating collision free cutter tool path and CMM tolerance inspection path. Due to the force feed back in all of the above activities, the product development process is more intuitive, efficient and user-friendly. A prototype system has been developed to demonstrate the proposed capabilities.