Search

The high computational load of conventional adaptive FIR filters applied to long system identification and their weak tracking ability has encouraged researchers to seek for efficient adaptive algorithms for this kind of applications. One of the efficient solutions is the three-level clipped LMS/RLS adaptive algorithm. In this paper, an insight into the performance of these adaptive algorithms that evaluates the amount of steady-state misalignment error of clipped LMS/RLS adaptive algorithms employed for identification of time-invariant and time-varying systems, is presented. Employing it, we compare the misalignment performance with their low-complexity adaptive algorithm counterparts theoretically. In addition, we derive the optimal step size/forgetting factor explicitly and also obtain a relation between the optimal level of the clipping and step size/forgetting factor to achieve the lowest steady-state misalignment and then we explain as to how to improve the performance of these kinds of adaptive algorithms by adjusting the clipping threshold according to the noise level in such a way that a higher performance is achieved. Finally, different adaptive algorithms have been further coded in VHDL in order to evaluate them in terms of speed and hardware resources.

Sea target detection is a vital application for military and navigation purposes. A new supervised clustering method based on the combination of the PSO and FCM techniques is presented for the sea target detection problem. The color components of the target and non-target pixels in the RGB color space are used as features to train the classification algorithm. The new classifier is presented in the form of a new color space which we call the Target-based Color Space (TCS); in fact the RGB color space is converted to this new space through a 3×3 matrix. The Particle Swarm Optimization (PSO) algorithm is then used to search for the optimum weights of the conversion matrix which results in a more discriminating clustering space between the target and non-target pixels. In other words, solving the optimization problem, minimization of the objective function of the FCM clustering technique in linear and quadratic transform domain (with a NP-hard problem in quadratic conversion), is done using the PSO algorithm. The main objective of this work is to demonstrate the efficiency of using just color features, as well as color space conversion in the classification domain. Experimental results show the efficiency of new method in finding sea targets in color images.

Recently, the mixed raster content model was proposed for compound document image compression. Most state-of-the-art document image compression methods, such as DjVu, work on the basis of this model but they have some disadvantages, especially for Farsi and Arabic document images. First, the Farsi/Arabic script has some characteristics which can be used to further improve the compression performance. Second, existing segmentation methods have focused on well-separating the textual objects from the background and/or optimizing the rate-distortion trade-off; nevertheless, they have not considered the text readability and OCR facility. Third, these methods usually suffer from the undesired jaggy artifact and misclassifying the important textual details. In this paper, MRC-based document image compression method is proposed which compromises rate-distortion trade-off better than the existing state-of-the-art document compression methods. The proposed method has higher performance in the aspects of segmentation, bi-level mask layer compression, OCR facility, and the overall compression. It uses a 1D pattern matching technique for compression of mask layer. It also uses a segmentation method which is sensitive enough to the small textual objects. Experimental results show that the proposed method has considerably higher compression performance than that of the state-of-the-art compression method DjVu, as high as 1.75–2.3.

Pattern matching is the most widely used technique for the compression of printed bi-level text images. In some printed scripts, letters normally attach to each other, or some letters have a simple relation to each other, or there may be undesired touching characters. Detecting such situations and exploiting them to reduce the library size, has a rather great effect on the compression ratio. In this paper, a lossy/lossless compression method for printed typeset bi-level text images is proposed for archiving purposes. For this, three techniques are proposed. First, the number of library prototypes is reduced by detecting and exploiting the mentioned situations. Second, a new effective encoding scheme is proposed for patterns and numbers. Third, three levels are proposed for lossy compression. Experimental results show that the proposed method works better, as high as 1.4–3.3 times in lossy case and 1.2–2.7 times in lossless case at 300 dpi, than the best existing compression methods or standards.

In this paper, the architecture of the first Iranian Farsi continuous speech recognizer and syntactic processor is introduced. In this system, by extracting suitable features of speech signal (cepstral, delta-cepstral, energy and zero-crossing rate) and using a hydrid architecture of neural networks (a Self-Organizing Feature Map, SOFM, at the first stage and a Multi-Layer Perceptron, MLP, at the second stage) the Iranian Farsi phonemes are recognized. Then the string of phonemes are corrected, segmented and converted to formal text by using a non-stochastic method. For syntactic processing, the symbolic (by using artificial intelligence techniques) and connectionist (by using artificial neural networks) approaches are used to determine the correctness, position and the kind of syntactic errors in Iranian Farsi sentences, as well.

We have shown that duct modeling using the generalized RBF neural network (DM_RBF), which has the capability of modeling the nonlinear behavior, can suppress a variable-frequency narrow band noise of a duct more efficiently than an FX-LMS algorithm. In our method (DM_RBF), at first the duct is identified using a generalized RBF network, after that stage of time delay of the input signal to the generalized RBF network is applied, then a linear combiner at their outputs makes an online identification of the nonlinear system. The weights of linear combiner are updated by the normalized LMS algorithm. We have showed that the proposed method is more than three times faster in comparison with the FX-LMS algorithm with 30% lower error. Also the DM_RBF method will converge in changing the input frequency, while it makes the FX-LMS cause divergence.