A multilayer incremental neural network architecture for classification

A new multilayer incremental neural network (MINN) architecture and its performance in classification of biomedical images is discussed. The MINN consists of an input layer, two hidden layers and an output layer. The first stage between the input and first hidden layer consists of perceptrons. The number of perceptrons and their weights are determined by defining a fitness function which is maximized by the genetic algorithm (GA). The second stage involves feature vectors which are the codewords obtained automaticaly after learning the first stage. The last stage consists of OR gates which combine the nodes of the second hidden layer representing the same class. The comparative performance results of the MINN and the backpropagation (BP) network indicates that the MINN results in faster learning, much simpler network and equal or better classification performance.     

Multi-sink load balanced forwarding with a multi-criteria fuzzy sink selection for video sensor networks

Congestion is a challenging problem in wireless sensor networks, which exacerbates with the high volume of data traffic imposed by video applications such as video surveillance and target tracking. Deployment of multiple sinks is a candidate solution for congestion and is also promising in terms of reliability and energy-efficiency. In order to gain the maximum benefit from multiple sinks, it is essential to distribute the load among them evenly. In this paper, we propose a cross layer geographic forwarding scheme MLBRF (Multi-Sink Load Balanced Reliable Forwarding) which aims to provide reliable and energy efficient video delivery in a multi-sinked sensor network for target tracking. In order to provide load balancing among the sinks, MLBRF proposes a sink selection mechanism based on fuzzy logic for the frame forwarding which evaluates the traffic density in the direction of each sink by combining two dynamic criteria which are the number of contenders and the buffer occupancy levels in the neighborhood with the static distance criterion. The performance of the fuzzy sink selection mechanism is compared using simulation with various sink selection mechanisms. The results show that MLBRF gains the maximum benefit from deploying multiple sinks in terms of reliability, latency and energy efficiency by using the proposed fuzzy sink selection mechanism.     

Fast algorithms for the discrete Fourier preprocessing transforms

It is pointed out that the two-stage Fourier algorithms are useful in a variety of applications in signal/image processing and recognition. The first stage, known as the DFPT (discrete Fourier preprocessing transform), has the potential of very fast and low-cost implementation, say, in a VLSI chip, as well as high-quality performance in applications without the second stage. The DFT (discrete Fourier transform), the RDFT (real DFT), or any other DFPT can be obtained if the second stage is completed. Since the second stage consists of independent blocks of circular correlations, they can be computed in fast sequential or parallel architectures. There are two groups of fast algorithms for the computation of DFPTs. The first group involves the representation of a DFPT in terms of skew-circular correlations (SCCs), which are then computed by fast SCC algorithms or parallel architectures such as semisystolic arrays. The second group involves the fast computation of a DFPT, including the DFT and the RDFT, by the implementationally simplest DFPT, such as class 2, case V DFPT     

Image coding with the discrete cosine-III transform

The discrete cosine-III transform (DC3T) is the same as the discrete symmetric cosine transform (DSCT) with a specific preprocessing of input data. It has less computational complexity than the discrete cosine transform (DCT) in terms of multiplications. The DC3T is also related to the DCT by a weighting matrix. The performance of the DC3T is compared to the DCT when compression is performed by adaptive coding. Experimental results show that a significant improvement in visual performance and mean square reconstruction error can be achieved over what is possible with the DCT. The better performance is attributed to the weighting of the DCT coefficients, which is indirectly achieved at reduced computational cost     

Quality assessment of surface and groundwater in Solaklı Basin (Trabzon, Turkey)

The paper reports a study of the surface and ground water in Solakl? Basin, Trabzon, from an assessment of 33 sampling points. Many of the parameters analysed are only marginally acceptable as potable water, even though the samples were collected during the wet season. The highest concentrations of contaminants were found downstream of the centres of population. The results indicate regular monitoring is required and some measures should be taken to protect the limited resources.