Nasser M. Nasrabadi

Channel Assignment in Cellular Radio Using Genetic Algorithms

The channel assignment problem has become increasingly important in mobile telephone communication. Since the usable range of the frequency spectrum is limited, the optimal assignment problem of channels has become increasingly important. Recently Genetic Algorithms (GAs) have been proposed as new computational tools for solving optimization problems. GAs are more attractive than other optimization techniques, such as neural networks or simulated annealing, since GAs are generally good at finding an acceptably good global optimal solution to a problem very quickly. In this paper, a new channel assignment algorithm using GAs is proposed. The channel assignment problem is formulated as an energy minimization problem that is implemented by GAs. Appropriate GAs operators such as reproduction, crossover and mutation are developed and tested. In this algorithm, the cell frequency is not fixed before the assignment procedures as in the previously reported channel assignment algorithm using neural networks. The average generation numbers and the convergence rates of GAs are shown as a simulation result. When the number of cells in one cluster are increased, the generation numbers are increased and the convergence rates are decreased. On the other hand, with the increased minimal frequency interval, the generation numbers are decreased and the convergence rates are increased. The comparison of the various crossover and mutation techniques in a simulation shows that the combination of two points crossover and selective mutation technique provides better results. All three constraints are also considered for the channel assignments: the co-channel constraint, the adjacent channel constraint and the co-site channel constraint. The goal of this paper is the assignment of the channel frequencies which satisfied these constraints with the lower bound number of channels.

Hierarchical block truncation coding

We present a new image coding technique called hierarchical block truncation coding (HBTC). HBTC is a combination of the block truncation coding (BTC) technique and the quadtree segmentation method. Quadtree segmentation is used to decompose an image into homogeneous regions so that the BTC method can exploit the nonstationary characteristics of the image data. The resulting bit rate is lower than that of conventional BTC, depending on the characteristics and complexity of the digital image. We investigated the performance of the encoder on both still and moving images. A small reduction in the bit rate is achievable for still images, but false contours become apparent as the rate declines. The proposed method works well on difference images from a sequence since the moving areas are encoded with greater resolution than the stationary background. A significant bit rate reduction is seen for sequence transmission. The bit rate is reduced from the 1.625 bits per pixel (bpp) required for a conventional BTC implementation down to 1.19 bpp for the least detailed still image. A typical CCITT image sequence was encoded at an average rate of 1.21 bpp. The bit rate was further reduced to an average of 0.39 bpp with a small degradation in the quality of the reconstructed images by transmitting only the most varying portions of the sequence.

An adaptive segmentation algorithm using iterative local feature extraction for hyperspectral imagery

We present an adaptive segmentation algorithm based on the iterative use of a modified minimum-distance classifier. Local adaptivity is achieved by gradually updating each class centroid over a local region whose size is reduced progressively during a segmentation process. The proposed method provides improved segmentation performance over template matching segmentation techniques because it adapts to the local context. The proposed algorithm can be applied to virtually any hyperspectral image regardless of size, dimensionality, and spectral sensitivity. Experimental results on a set of visible to near-infrared hyperspectral images using both the proposed algorithm and a standard template matching technique are presented.

A committee of networks classifier with multi-resolution feature extraction for automatic target recognition

A neural network-based classifier has been applied to the problem of automatic target recognition (ATR) using forward-looking infrared (FLIR) imagery. The target classifier consists of several neural networks that form a committee for classification. Each neural network in the committee receives inputs from features extracted from only a local region of a target, known as a receptive field, and is trained independently from other committee members. The classification results of the individual neural networks are combined to determine the final classification. Our experiments show that this committee of networks classifier is superior to a fully connected neural network classifier in terms of complexity (number of weights to be learned) and performance (classification rate). The proposed classifier shows a high noise immunity to clutter or target obscuration due to the independence of the individual neural networks in the committee. Performance of the proposed classifier is further improved by the use of multi-resolution features and by the introduction of a higher level neural network on the top of committee, a method known as stacked generalization.

A modular neural network vector predictor for predictive VQ

In this paper, we present a modular neural network vector predictor in order to improve the predictive component of a predictive vector quantization (PVQ) scheme. The proposed vector prediction technique consists of five dedicated predictors (experts), where each expert predictor is optimized for a particular class of input vectors. An input vector is classified into one of five classes based on its directional variances. One expert is optimized for stationary blocks, and each of the other four experts are optimized to predict horizontal, vertical, 45/spl deg/, and 135/spl deg/ diagonally oriented edge-blocks, respectively. An integrating unit is then used to select or combine the outputs of the experts in order to form the final output of the modular network. Therefore, no side information is required to transmit to the receiver about the predictor selection. Experimental results show that the proposed scheme gives an improvement of 1 to 1.5 dB better than a single multilayer perceptron (MLP) predictor. However, if the information about the predictor selection is sent to the receiver, the improvement could be up to 3 dB over the single MLP predictor. The perceptual quality of the predicted images is also significantly improved.

Next-state functions for finite-state vector quantization

In this paper, a finite-state vector quantizer called Dynamic Finite-State Vector Quantization (DFSVQ) is investigated with regard to its subcodebook construction. In DFSVQ each input vector encoded by a small codebook, called subcodebook, is created from a much larger codebook called supercodebook. The subcodebook is constructed by selecting (reordering procedure) a set of appropriate codevectors from the supercodebook. The performance of the DFSVQ depends on this reordering procedure, therefore, several reordering procedures are introduced and their performances are evaluated in this paper. The reordering procedures that are investigated are the conditional histogram, address prediction, vector prediction, nearest neighbor design, and the frequency usage of codevectors. The performance of the reordering procedures are evaluated by comparing their hit ratios (the number of blocks encoded by the subcodebook) and their computational complexity. Experimental results are presented for both still images and video. It is found that for still images the conditional histogram performs the best and for video the nearest neighbor design performs the best.

Asymptotical analysis of a modular neural network

Modular neural networks have been used in several applications because of their superiority over a single neural network in terms of faster learning, proper data representation, and feasibility of hardware implementation. This paper presents an asymptotical performance analysis showing that the performance of a modular neural network is always better than or as good as that of a single neural network when both neural networks are optimized. The minimum mean square error (MSE) that can be achieved by a modular neural network is also obtained.

Combination of two learning algorithms for automatic target recognition

Composite classifiers consisting of a number of component classifiers have been designed and evaluated on the problem of automatic target recognition (ATR) using a large set of real forward-looking infrared (FLIR) imagery. Two existing classifiers are used as the building blocks for our composite classifiers. The performance of the proposed composite classifiers are compared based on their classification ability and computational complexity. It is demonstrated that the composite classifier based on a cascade architecture greatly reduces the computational complexity with a statistically insignificant decrease in performance in comparison to standard classifier fusion algorithms.

Neural network architectures for vector prediction

A vector predictor is an integral part of the predictive vector quantization (PVQ) scheme. The performance of a predictor deteriorates as the vector dimension (block size) is increased. This makes it necessary to investigate new design techniques in order to design a vector predictor which gives better performance when compared to a conventional vector predictor. This paper investigates several neural network configurations which can be employed in order to design a vector predictor. The first neural network investigated in order to design the vector predictor is the multi-layer perceptron. The problem with multi-layer perceptron is the long convergence time which is undesirable when the on-line training of the neural network is required. Another neural network called functional link neural network has been shown to have fast convergence. The use of this network as a vector predictor is also investigated. The third neural network investigated is a recurrent type neural net. It is similar to the multi-layer perceptron except that a part of the predicted output is fed back to the hidden layer/layers in an attempt to further improve the current prediction. Finally, the use of a radial-basis function (RBF) network is also investigated for designing the vector predictor. The performances of the above mentioned neural network vector predictors are evaluated and compared with that of a conventional linear vector predictor.

Wavelet-based learning vector quantization for automatic target recognition

An automatic target recognition classifier is constructed that uses a set of dedicated vector quantizers (VQs). The background pixels in each input image are properly clipped out by a set of aspect windows. The extracted target area for each aspect window is then enlarged to a fixed size, after which a wavelet decomposition splits the enlarged extraction into several subbands. A dedicated VQ codebook is generated for each subband of a particular target class at a specific range of aspects. Thus, each codebook consists of a set of feature templates that are iteratively adapted to represent a particular subband of a given target class at a specific range of aspects. These templates are then further trained by a modified learning vector quantization (LVQ) algorithm that enhances their discriminatory characteristics. A recognition rate of 69.0 percent is achieved on a highly cluttered test set.