نشریه مهندسی برق و مهندسی کامپیوتر ایران
سال هفتم شماره 4 (پیاپی 20، زمستان 1388)

Two approaches for the recognition of printed Farsi documents based on sub-word shape recognition is proposed. First approach is based on recognition of sub-word shape as a whole and the second is based on the recognition of the body of sub-words. Sub-word body is constructed via removing dots and signs of the sub word. In second approach, information of dots and signs will be added after recognition of the body. Both approaches have two phases: training and test. In training phase, sub-words are clustered based on ISODATA algorithm. Initial centers of the clusters are computed through a hierarchical clustering algorithm. In first approach, sub-word recognition is performed in two stages: finding clusters close to the input sub-word and then finding the best match within the sub-words of these clusters. In the second approach another stage is required to find the final sub-word including dots and signs. Experimental results show that on clean images the first algorithm have better performance; 94% versus 93% in word level. But when dealing with low quality and noisy images, both algorithms are suffering from reduced accuracy. Sometimes this reduction is significant. The reasons of this behavior are inspected and some solutions are presented. Finally we compared both methods and inspected pros and cons of Farsi OCR based on sub-word shape.

In this paper, coherent radar detection of moderately fluctuating targets in Gaussian clutter is performed. In the previous works, the target vector has been modeled as a Gaussian random vector that is independent of clutter vector. In this paper, based on the coherence assumption, by extending the Swerling models for radar targets to the moderate fluctuating cases, we propose a more accurate model for every fluctuating target. The proposed model shows that the Gaussian assumption for moderately fluctuating targets is not correct.Then, based on the proposed model, the optimal detection of moderately fluctuating targets in Gaussian clutter is performed and the GLR (Generalized Likelihood Ratio) test is obtained. Also, the ECD (Estimator-Correlator Detector) and OF (Optimal Filter) (that have been proposed as the best detectors in the previous works) are derived based on the new model.Computer simulation results show that the ECD has a superior performance than the other detectors. Nevertheless, the OF can be considered as an admissible detector because of the simpler algorithm and the performance close to the ECD. Although, the GLR detector has a suitable performance, but its performance is often lower than that ECD and OF while its algorithm is more complicated. Therefore, the GLR detector cannot compete to the ECD and OLD. Computer simulation results also show that these detectors do not have a convenient constant false alarm rate.

Click stream analysis is known as an effective method for customer’s viewing route prediction in a particular web site. Predicting Customer viewing behavior provides considerable advantages in different areas such as e-commerce, e-business and customer relationship management. This paper aims to provide a 0-1 mathematical model based on Markov models for evaluating the most probable viewing route of a customer in a website. This problem can be formulated as an especial case of well-known Prize Collecting Traveling Salesman Problem (PCTSP) which is a NP-hard problem and its sub tour elimination constraints are increased drastically by increasing the model parameters. Also an effective algorithm is introduced in this paper to solve this NP-hard model. For model validation, the proposed model was implemented by using the log files of a university web site server for 20 different users. Comparison of the results with commonly used Giudici algorithm shows that the proposed model yields better and exacter solutions.

In this paper, a model is presented for the optimal power flow of an HVAC/HVDC transmission network. OPF is a fundamental tool in power system operation and planning. The model proposed for optimal power flow includes network control parameters settings and HVDC links parameters tunings. Also, excessive equipment installation for appropriate operation of the network is minimized, while the network security margin is maximized. To solve the proposed model, a hybrid evolutionary algorithm by combining Particle Swarm Optimization (PSO) and Differential Evolution (DE) is proposed. The methodology is tested on IEEE-30 test system and compared with the results from other OPF techniques. Also the impact of HVDC links on OPF studies is illustrated by numerical examples.

In this paper a 5th-Order single-loop Sigma-Delta Modulator with low distortion structure is presented. This structure, which uses integrator and IIR filter concurrently, has relatively less feedforward paths and modulator coefficients. Thus, its sensitivity to coefficient mismatching is reduced. To lower the power consumption of the modulator, the 2-order IIR filter block is implemented by single OTA, and a passive adder is used to realize input quantizer adder. Simulation results show that this structure can achieve 15-bit of resolution and 6 MHz input signal bandwidth, with 1.2 V supply voltage using a 0.13 µm CMOS technology. Power consumption of modulator is 53 mW. Comparing with other structures, the proposed modulator has higher performance because of increasing the DR and input bandwidth of modulator without extra increasing the power consumption.

The iterative angular spectrum (IAS) method has been introduced by Mellin and Nordin for designing finite-aperture diffractive optical elements (FADOEs). We have extended this method to two-dimensional FADOEs and used it to design some optical devices. The first device is a 1-to-7 beamsplitter that couples an optical beam to seven single-mode optical fibers with a diffraction efficiency of 84%. The second device is a beam-shaper that converts a Gaussian beam into a nearly flat beam with a diffraction efficiency of 74.8%. The third design is a 1-to-3 asymmetric beamsplitter. The fourth design includes three microlenses with different focal lengths. The desired intensity distribution patterns of all these designs are located at the near field region. We have investigated the sensitivity of the extended method by comparing the results obtained by this method with those obtained by three-dimensional finite difference time domain (3-D FDTD) method using perfect matched layer (PML). Also, a 1-to-5 beamsplitter is fabricated and the experimental results are presented.

In recent years, automated recognition of gestures in the finger spelling paradigm has become an active research area. Gesture is a combination of hand postures, hand movements, and face gestures; and finger spelling is a way of presenting alphabets of a word that does not exist in the sign language dictionary. In this paper, we present a scheme for hand gesture recognition in finger spelling of Farsi alphabets, where a different shape for hand and fingers denote a different letter in the alphabet. Our scheme has five stages, namely, visual data gathering, preprocessing of the image, detection and extraction of hand’s features, feature reduction and consolidation, and finally, hand gesture recognition. For the last stage (hand gesture recognition), we employ three techniques, namely, the nearest neighbor using the Euclidian distance, the nearest neighbor using the normalized Euclidian distance, and neural networks. For reducing the feature space, we use the discrete cosine transform (DCT), which yields better results as compared to the discrete Fourier transform and Fourier coefficients. We achieved 99.1% correct recognition using neural networks, which is superior to existing schemes.

Differential pulse code modulation (DPCM) system plays an important role in communication systems. In this paper a considerable amount of theoretical analysis has been carried out on DPCM structure, degree of predictor and quantizer unit. Simulation has been done using DPCM scheme. Here, instead of direct coding of signal, difference between two signals, input to DPCM and estimated output, is encoded. Consequently, application of the difference signal to quantizer unit, fewer bits for each sample is achieved at the output of the quantizer unit. In this study, Levinson-Durbin algorithm is utilized to design the optimal predictor. For the optimal quantized levels, classical together with Lloyd method are employed. Finally, second order predictor together with optimal quantized levels, 2-bit for normal case and 3-bit for optimal case, is successfully achieved to process ECG signal at DPCM scheme.