Atilla Özmen

Residual Separation of Magnetic Fields Using a Cellular Neural Network Approach

Abstract — In this paper, a Cellular Neural Network (CNN) has been applied to a magnetic regional/residual anomaly separation problem. CNN is an analog parallel computing paradigm defined in space and characterized by the locality of connections between processing neurons. The behavior of the CNN is defined by the template matrices A, B and the template vector I. We have optimized weight coefficients of these templates using Recurrent Perceptron Learning Algorithm (RPLA). The advantages of CNN as a real-time stochastic method are that it introduces little distortion to the shape of the original image and that it is not effected significantly by factors such as the overlap of power spectra of residual fields. The proposed method is tested using synthetic examples and the average depth of the buried objects has been estimated by power spectrum analysis. Next the CNN approach is applied to magnetic data over the Golalan chromite mine in Elazig which lies East of Turkey. This area is among the largest and richest chromite masses of the world. We compared the performance of CNN to classical derivative approaches.

Separation of Bouguer anomaly map using cellular neural network

Abstract In this paper, a modern image-processing technique, the Cellular Neural Network (CNN) has been firstly applied to Bouguer anomaly map of synthetic examples and then to data from the Sivas–Divrigi Akdag region. CNN is an analog parallel computing paradigm defined in space and characterized by the locality of connections between processing neurons. The behaviour of the CNN is defined by two template matrices and a template vector. We have optimised the weight coefficients of these templates using the Recurrent Perceptron Learning Algorithm (RPLA). After testing CNN performance on synthetic examples, the CNN approach has been applied to the Bouguer anomaly of Sivas–Divrigi Akdag region and the results match drilling logs done by Mineral Research and Exploration (MTA).

Correlation of ternary liquid--liquid equilibrium data using neural network-based activity coefficient model

Liquid--liquid equilibrium (LLE) data are important in chemical industry for the design of separation equipments, and it is troublesome to determine experimentally. In this paper, a new method for correlation of ternary LLE data is presented. The method is implemented by using a combined structure that uses genetic algorithm (GA)--trained neural network (NN). NN coefficients that satisfy the criterion of equilibrium were obtained by using GA. At the training phase, experimental concentration data and corresponding activity coefficients were used as input and output, respectively. At the test phase, trained NN was used to correlate the whole experimental data by giving only one initial value. Calculated results were compared with the experimental data, and very low root-mean-square deviation error values are obtained between experimental and calculated data. By using this model tie-line and solubility curve data of LLE can be obtained with only a few experimental data.

Detection of Trojans in integrated circuits

This paper presents several signal processing approaches in Trojan detection problem in very large scale integrated circuits. Specifically, wavelet transforms, spectrograms and neural networks are used to analyze power side-channel signals. Trojans in integrated circuits can try to hide themselves and become almost invisible due to process and measurement noises. We demonstrate that our initial results with these techniques are promising in successful detection. Discrete wavelet transforms and spectrograms can provide clear visual assistance in detecting Trojans by catching the time-scale differences and time-frequency activities introduced by the Trojans. Furthermore, neural networks with sufficient training are also used and simulation results show that correct decisions are possible with a very high success rate.

Amplitude and Frequency Modulation behaviours of Cellular Neural Networks

As is well known, Amplitude Modulation (AM) and Frequency Modulation (FM) are the most popular modulation techniques in communications [1]. Classically, they are designed as individual systems at transmitters; however in this study, it is shown that, both can be realized under a single architecture, specifically on a Cellular Neural Network with an opposite sign template (CNN-OST). First, the CNN-OST is introduced and then, AM and FM behaviours of this structure are explained. Simulations are carried out where satisfactory results are found and finaly, the advantages and drawbacks of the proposed system are discussed.

Channel equalization with cellular neural networks

In this paper, a dynamic neural network structure called Cellular Neural Network (CNN) is employed for the equalization in digital communication. It is shown that, this nonlinear system is capable of suppressing the effect of intersymbol interference (ISI) and the noise at the channel. The architecture is a small-scaled, simple CNN containing 9 neurons, thus having only 19 weight coefficients. Proposed system is compared with linear transversal filters as well as with a Multilayer Perceptron (MLP) based equalizer.

Amplitude and Frequency Modulations with Cellular Neural Networks

Amplitude and frequency modulations are still the most popular modulation techniques in data transmission at telecommunication systems such as radio and television broadcasting, gsm etc. However, the architectures of these individual systems are totally different. In this paper, it is shown that a cellular neural network with an opposite—sign template, can behave either as an amplitude or a frequency modulator. Firstly, a brief information about these networks is given and then, the amplitude and frequency surfaces of the generated quasi-sine oscillations are sketched with respect to various values of their cloning templates. Secondly it is proved that any of these types of modulations can be performed by only varying the template components without ever changing their structure. Finally a circuit is designed, simulations are presented and performance of the proposed system is evaluated. The main contribution of this work is to show that both amplitude and frequency modulations can be realized under the same architecture with a simple technique, specifically by treating the input signals as template components.

Genetic algorithm based broadband equalizer design with ripple level control

In this paper, broadband equalizer design with ripple control via genetic algorithm has been studied. The equalizer is defined as a lossless two-port terminated by load impedance, and the coefficients of its describing scattering polynomials have been optimized via genetic algorithm. During the optimization process, ripple level of the transducer power gain has been controlled. An example has been given to illustrate the utilization of the proposed approach.

DETECTION AND CLASSIFICATION OF VIEWER AGE RANGE SMART SIGNS AT TV BROADCAST

In this paper, the identification and classification of “Viewer Age Range Smart Signs”, designed by the Radio and Television Supreme Council of Turkey, to give age range information for the TV viewers, are realized. Therefore, the automatic detection at the broadcast will be possible, enabling the manufacturing of TV receivers which are sensible to these signs. The most important step at this process is the pattern recognition. Since the symbols that must be identified are circular, various circle detection techniques can be employed. In our study, first,two different circle segmentation methods for still images are analyzed, their advantages and drawbacks are discussed. A popular neural network structure called Multilayer Perceptron is employed for the classification. Afterwards, the same procedures are carried out for streaming video.All of the steps depicted above are realized on a standard PC.

Real time detection and classification of age range smart signs at TV broadcast

In this paper, real time detection and classification of the “Smart Signs”, which appear at the beginning and after the commercial breaks of the programs at national TV channels, that indicates the appropriate viewer ages for the broadcast, is realized by using a fast image processing technique. For this purpose, circular Hough Transform (CHT) and a circle detection algorithm depending on the determination of center of gravity of closed objects is carried out for the segmentation of these signs and afterwards a Multilayer Perceptron (MLP) is utilized for the classification process at the video signals.