This paper estimates and segments the moving objects based on center of mass model to decrease the search window and to provide a new algorithm, which achieves an accurate and rapid tracking. Furthermore, a novel method is proposed to update the template size adaptively by using estimation and segmentation of moving objects. The estimated results of moving target is transformed to wavelet domain and target tracking is performed in that domain. To improve the algorithm center of mass model is performed in wavelet domain. By using kalman predictor and thresholding method, a new approach is presented for object tracking failure and recovery.

Object tracking is one of the major subjects in machine vision and plays a main role in detection of major events in indoor soccer matches. In this paper, a novel approach for tracking the ball and players is proposed. In this method, the ground lines are segmented and eliminated using a fast and effective method. Then, the remaining non-field pixels are considered and labeled as players and the ball. A fast level set contour is used to track players and the ball. The proposed method can track players and the ball in presence of occlusion. Experiments show that the proposed method is robust to occlusion and different field colors.

This paper, proposes the use of a reinforcement learning approach for a target tracking sensor network application. Harsh and unpredictable situations of sensor nodes in such an application requires a selftuning mechanism for the nodes to adapt their behavior over time. The method is examined under high dynamic network conditions and compared with a similar method called SORA over different performance measures. The results show a significant improvement over the compared method in the environments with high level of dynamism.

This paper proposes an integrated framework for speaker indexing which includes both speaker segmentation and speaker clustering. Speaker indexing systems has wide domains of application with different requirements which make a general speaker indexing framework hard to accomplish. The main source of performance degradation in speaker indexing is the probable existence of short speech utterances which makes the speaker turns hard to distinguish and also exposes the segment modeling to data insufficiency. This paper introduces a speaker indexing framework with high average performance which uses Support Vector Machines (SVM) as the core approach. The main contribution of this framework is the SVM based clustering approach which makes the indexing more robust against the short speech segments. This framework is evaluated on a domestic conversational speech dataset and the results were satisfactory.

One of the most important phases of speaker indexing is speaker clustering which aims to find the number of speakers in a speech document and merge the speech segments corresponding to a single speaker. The most critical source of problem in speaker clustering is the speech segments duration which may be so short that proper segment modeling becomes hard to achieve. An alternative suggestion in these situations is to adapt global models with new data instead of building the speaker models from the ground. In this paper we investigate two adaptation techniques in eigen-voice space for improving clustering performance especially for shorter speech utterances. These techniques were embedded in a clustering framework and evaluated on a set of domestic conversational speech. We have also compared the proposed methods with some other known techniques. The experiments show a considerable improvement in speaker clustering performance.

Though numerous approaches have been proposed for face recognition. In this paper we propose a novel face recognition approach based on adaptively weighted patch Local Statistic in Multi dimensional (LMDS) when only one exemplar image per person is available. In this approach, a face image is decomposed into a set of equal-sized patches in a nonoverlapping way. In order to obtain Local Multi Dimensional Statistic Features in each patch, we calculated mean and standard deviation of all pixels along some directions. An adaptively weighting scheme is used to assign proper weights to each LMDS features to adjust the contribution of each local area of a face in terms of the quantity of identity information that a patch contains. An extensive experimental investigation is conducted using AR face databases covering face recognition under controlled/ideal conditions and different facial expressions. The system performance is compared with the performance of four benchmark approaches. The encouraging experimental results demonstrate that our approach can be used for face recognition and patch-based local statistic features provides a novel way for face.

PEnT1 is an automatic English to Persian text translator. It translates simple English sentences into Persian, exploiting a combination of rule based and semantic approaches. It covers all the twelve tenses in English in both passive and active verbs for indicative, negative, interrogative sentences. In this paper, introducing PEnT1, we propose a new WSD method by presenting a hybrid measure to score different senses of a word. We also discuss prototyping some linguistic resources to test our methods.

Vehicle tracking is an essential requirement of any vision based Intelligent Transportation System for extracting different traffic parameters, efficiently. Handling inter-object occlusion is the most challenging part of tracking as a process of finding and following interested objects in a sequence of video frames. In this paper we present a system, based on code-book background model for motion segmentation and Kalman filter for tracking with a new approach for occlusion. This approach separates occluded vehicles based on longest common substring of chain codes. We use this tracking system to estimate some traffic parameters. Experimental results show the efficiency of the method

In this paper, a new fuzzy fading memory (FFM) is developed in order to aid a modified input estimation (MIE) technique and enhance its performance in tracking high maneuvering targets. The MIE has been introduced recently and performs well in tracking low and medium maneuvering targets. However, due to some modeling errors, the accuracy of this tracker may be seriously degraded in presence of high maneuvers. To cope with this difficulty, an intelligent approach based on FFM is presented in this paper. Simulation results prove the efficiency of the proposed method in tracking high maneuvering targets.

In this paper, we proposed a new feature subset selection approach. In proposed approach first, the entire dataset are classified and the best number of clusters over it are found according to silhouette value. Then according to this value, each feature is alone classified with the same cluster number and accordingly the proposed entropy fuzzy measure is found for them. We examine our method on some traditional datasets. The results show a good performance of proposed method.

Concept graph is a graph that represents the relationships between language concepts. In this structure the relationship between any two words is demonstrated by a weighted edge such that the value of this weight is interpreted as the degree of the relevance of two words. Having this graph, we can obtain most relevant words to a special term. In this paper, we propose a method for improving the classification of documents from unknown sources by means of concept graph. In our method, initially some features are selected from a training set by a well-known feature selection algorithm. Then, by extracting most relevant words for each class from the concept graph, a more effective feature set is produced. Our experimental results identify an improvement of 1% and 8% in precision and recall measures, respectively.

In this paper we present a new approach for evolving an optimized neural network architecture for a three layer feedforward neural network with a mutation based genetic algorithm. In this study we will optimize the weights and the network architecture simultaneously through a new presentation for the three layer feedforward neural network. The goal of the method is to find the optimal number of neurons and their appropriate weights. This optimization problem so far has been solved by looking at the general architecture of the network but we optimize the individual neurons of the hidden layer. This change results in a search space with much higher resolution and an increased speed of convergence. Evaluation of algorithm by 3 data sets reveals that this method shows a very good performance in comparison to current methods.

A model of episodic memory is derived to propose algorithms of text categorization with semantic space models. Performances of two algorithms are contrasted using textual material of the text-mining context ‘DEFT09’. Results confirm that the episodic memory metaphor provides a convenient framework to propose efficient algorithm for text categorization. One algorithm has already been tested with LSA. The present paper extends these algorithms to another model of Word Vector named Random Indexing.

Gender identification based on speech signal has become gradually a matter of concern in recent years. In this context 6 feature types including MFCC, LPC, RC, LAR, pitch values and formants are compared for automatic gender identification and three best feature types are selected using four feature selection techniques. These techniques are GMM, Decision Tree, Fisher’s Discriminant Ratio, and Volume of Overlap Region. A dimension reduction is done on the best three feature types and the best coefficients are then selected from each feature vector. Selected coefficients are evaluated for gender classification using three types of classifiers including GMM, SVM and MLP neural network. 96.09% gender identification performance was obtained as the best performance using the selected coefficients and MLP classifier.

In this paper a two step method based on Gauss- Newton and factor graphs algorithm is proposed for localization to enhance accuracy of localization. The Gauss-Newton algorithm is accurate method for positioning. The most important challenge of this method is senility to initial point; this problem is solved in positioning based on factor graphs. So, in this paper, first positioning equations using angle of arrival is considered based on factor graphs algorithm. Second, final location estimation is performed using Gauss Newton algorithm with error near to Cramer-Rao bound. Simulation results shows that positioning error using two step method has maximum 6% gap to Cramer-Rao Bound.

The presence of undetected direct path conditions in the TOA (Time of Arrival)-based indoor geolocation introduces major errors into distance measurements. In the proposed location estimation, an improved procedure is introduced in Ultra-wideband (UWB) systems in Line of Sight (LOS) and Non-Line of Sight (NLOS) multipath environments by incorporating the skewness as new statistic information of multipath channel. Weighted least square error (WLS) algorithm is used to estimate source position in this paper. Simulation results show effectiveness of the proposed algorithm, by reducing source localization error. Uncertainty region between NLOS and LOS identification is reduced in the proposed method compared to previous methods..

Human can determine optimal behaviors which depend on the ability to make planned and adaptive decisions. In this paper, we have proposed a predictive structure based on neuropsychological evidences to model human decision making process by concentrating on the role of frontal brain regions which are responsible for predictive control of human behavior. We have considered a model-based reinforcement learning framework to implement the relations between these brain areas. Finally, we have designed an experimental test to compare the function of model with human behavior in a maze task. Our results reveal that there is more than reward and punishment in human behavior, and considering higher cognitive functions such as prediction will help to have more reliable models which could better describe human behavior.

In this paper, we introduce a refinement approach to develop probabilistic programs formally. To achieve this goal, we first present a way to specify probabilistic programs in a Z-based notation. We then use an existing method of translating Z into a refinement calculus to transform our Z-style specifications of probabilistic programs into specification statements of the refinement calculus. We finally add new laws to the refinement calculus helping us to refine the resulting specification statements into probabilistic choice constructs of a probabilistic imperative language. In this way, we will provide a completely formal way for developing probabilistic programs.

Due to the changing and dynamic environment of web services, Quality of Service (QoS) becomes a key factor to differentiate service providers. Since current web service standards and technologies suffer from the lack of QoS Management, having architectures capable of supporting QoS verification, selection, negotiation and monitoring is inevitable. Software agents have been recognized as a promising technology for managing web services. Using FIPA compliant Multi Agents we were able to propose a Multi Agents based web service QoS Management Architecture. We also introduced a QoS based web service clustering method which helps us to select the best service that suits user quality preferences.

The developers of a system are accepted as one of the most important stakeholders of an Architecture Description. The Development Viewpoint is suggested to satisfy the needs of the developers throughout the development process via codeline organization descriptions, programming models, etc. However, the available models for such purposes, if any, barely cross informal natural language descriptions and checklists. This paper introduces the idea of enhancing the description of the Development viewpoint using lightweight Domain- Specific Languages and presents the application of such languages in two industrial case studies. This language enables the architect to provide the necessary guidelines that constrains the implementers during the development process and it is also used as a means to discover the deviation of the code from the architecture as the development goes on.