نشریه فناوری اطلاعات و ارتباطات ایران
سال دوم شماره 5 (پاییز و زمستان 1389)

GMM is one of the most successful models in the field of automatic language identification. In this paper we have proposed a new model named adapted weight GMM (AW-GMM). This model is similar to GMM but the weights are determined using GMM-VSM LID system based on the power of each component in discriminating one language from the others. Also considering the computational complexity of GMM-VSM، we have proposed a technique for constructing bigram sequences of components which could be used for higher sequence orders and decreases the complexity. Experiments on four languages of OGI corpus including English، Farsi، French and German have shown the effectiveness of proposed techniques.

Increment of complexity and costs of information technology systems have made many problems about infrastructure and manpower for organizations which have been decreased by the use of outsourcing. All organizations try to increase the success of outsourcing projects by using different ways. One of the important reasons for failure of these projects especially in IT area- because of its major role in acquisition of competitive advantage- is selecting inappropriate contractor. In order to existence of different and contradictive criteria، this selection is complex. The purpose of this research is to determine important criteria and specify the weights of each criterion and finally design a fuzzy expert system for selecting the best contractor in IT outsourcing. The method of knowledge acquisition from experts-which are managers and experts of IT- is a questionnaire. Also in order to evaluate the validity of system، it was used in an IT company. The results show the favorable performance of contractor selection expert system.

Broadcasting is one of the most fundamental operations in mobile ad hoc networks. Broadcasting serves as a building block in many routing protocols. The simplest approach for broadcasting is flooding. However، it generates many redundant messages that cause to loss of scarce resources such bandwidth and battery powers and it may lead to contention، collision and extreme decrease of network performance. Many schemes have been developed to improve the performance of flooding algorithm that are mainly classified into two basic approaches: deterministic and probabilistic approaches، of which the second is more considered. However، the existing schemes either increase the latency or decrease the reach ability of the algorithm. In this paper، we propose a new scheme to improve the performance of flooding algorithm. The basis of the new method is the probabilistic rebroadcasting based on local observations. In this method، the probability function of each node is adjusted dynamically based on local observations. Simulation results show that the new scheme considerably decrease average latency compared with similar existing schemes while maintaining the reach ability and saving messages.

The quality of the image formed by an optical system is reduced by aberrations. This paper points out and attempts to correct blindness of lens aberration. Therefore modified Zernike moments were introduced to present lens aberration model within which their coefficients were estimated through poly-spectral analysis. The model parameters are divided into asymmetric and symmetric which are estimated through bi-coherence and tri-coherence respectively. The obtained precision compares favorably to the aberration given by state of the art ploy-spectral analysis and reaches a RMSE of 0.1 pixels.

Image compression is one of the important research fields in image processing. Up to now، different methods are presented for image compression. Neural network is one of these methods that has represented its good performance in many applications. The usual method in training of neural networks is error back propagation method that its drawbacks are late convergence and stopping in points of local optimum. Lately، researchers apply heuristic algorithms in training of neural networks. This paper introduces a new training method based on the Gravitational Search Algorithm. Gravitational Search Algorithm is the latest and newest version of swarm intelligence optimization approaches. In this algorithm، the candidate answers in search space are masses that interact with each other by gravitational force and change their positions. Gently، the masses with better fitness obtain more mass and effect on other masses more. In this research، an MLP neural network by GSA method is trained for images compression. In order to efficiency evaluation of the presented compressor، we have compared its performance toward PSO and error back propagation methods in compression of four standard images. The final results show salient capability of the proposed method in training of MLP neural networks.

With development of digital multimedia technology and rapid growth of the Internet، illegal copy and exchange of digital multimedia sources is also spread. In such environment، copyright protection plays an essential role. In this paper a new semi- blind image watermarking algorithm for proof of ownership is proposed. At first، the original image is transformed to transform domain and the low frequency sub-band is selected to make reference image. Then، 1-level wavelet decomposition is applied to reference and grey-scale watermark images. Finally، the embedding is done by modifying the singular values of reference image’s sub-bands with the equivalent singular values of watermark’s sub-bands. In the proposed method، the reference image is needed during the extraction process، so it is called semi blind method. Robustness of the proposed method against various attacks that are applied to the watermarked image is investigated. The experimental results show that the proposed method is more robust than previous works against different attacks and the watermarked image looks visually identical to the original image.