Journal of Circuits, Systems and Computers

Image segmentation is one of the most important parts of clinical diagnostic tools. Medical images mostly contain noise and inhomogeneity. Therefore, accurate segmentation of medical images is a very difficult task. However, the process of accurate segmentation of these images is very important and crucial for a correct diagnosis by clinical tools. We proposed a new clustering method based on Fuzzy C-Mean (FCM) and user specified data. In the postulated method, the color image is converted to grey level image and anisotropic filter is applied to decrease noise; User selects training data for each target class, afterwards, the image is clustered using ordinary FCM. Due to inhomogeneity and unknown noise some clusters contain training data for more than one target class. These clusters are partitioned again. This process continues until there are no such clusters. Then, the clusters contain training data for a target class assigned to that target class; mean of intensity in each class is considered as feature for that class, afterwards, feature distance of each unsigned cluster from different class is found then unsigned clusters are signed to target class with least distance from. Experimental result is demonstrated to show effectiveness of new method.

DNA motif discovery is an important task since it helps to better understand the regulation of the transcription in the protein synthesis process. This paper introduces a novel method for the task of DNA motif finding where the proposed method adopts machine-learning approach by the use of a well-known clustering algorithm, Fuzzy C-Means. The method is explained in detail and tested against DNA sequences extracted from the genome of Saccharomyces cerevisiae and Escherichia coli organisms. Experimental results suggest that the algorithm is efficient in finding statistically interesting features existing in the DNA sequences. The comparison of the algorithm with the well-known motif finding tools, MEME and MDScan, which are built on statistical and word-enumerative models, shows the advantages of the proposed method over the existing tools and the promising direction of the machine-learning approach.

Database classification suffers from two common problems, i.e., the high dimensionality and nonstationary variations within the large historic data. This paper presents a hybrid classification model by integrating a case-based reasoning technique, a Support Vector Machine (SVM), and Genetic Algorithms to construct a decision-making system for data classification in various database applications. The model is mainly based on the concept that the historic database can be transformed into a smaller case-base together with a group of SVM models. As a result, the model can more accurately respond to the current data under classifying from the inductions by these SVM models generated from these smaller case bases. Hit rate is applied as a performance measure and the effectiveness of our proposed model is demonstrated by experimentally compared with other approaches on different database classification applications. The average hit rate of our proposed model is the highest among others.

Optimal planning of large-scale distribution networks is a multiobjective combinatorial optimization problem with many complexities. This paper proposes the application of improved genetic algorithm (GA) for the optimal design of large-scale distribution systems in order to provide optimal sizing and locating of the high voltage (HV) substations and medium voltage (MV) feeders routing, using their corresponding fixed and variable costs associated with operational and optimization constraints. The novel approach presented in the paper, solves hard satisfactory optimization problems with different constraints in large-scale distribution networks. This paper presents a new concept based on MST in graph theory and GA for optimal locating of the HV substations and MV feeders routing in a real-size distribution network. Minimum spanning tree solved with Prim's algorithm is employed to generate a set of feasible population. In the present article, to reduce computational burden and avoid huge search space leading to infeasible solutions, special coding method is generated for GA operators to solve optimal feeders routing. The proposed coding method guarantees the validity of the solution during the progress of the GA toward the global optimal solution. The developed GA-based software is tested in a real-size large-scale distribution system and the well-satisfactory results are presented.

Neural networks (NNs) provide an efficient tool that can be trained to estimate the value of output parameters given certain metrics. In this paper, NNs are used to enhance intelligent agents that negotiate on behalf of their owners aiming to maximize their utility. More specifically, NNs are exploited in order to predict the hybrid negotiation behavior of the agents' opponents, thus achieving more profitable results for the parties these agents represent. The NNs provide the means so that the agents can early detect the cases where agreements are not achievable, thus supporting their decision to withdraw or not from the negotiation threads. The designed NN-assisted negotiation strategies have been evaluated via extensive experiments and are proven to be very useful, as they manage to significantly reduce the average duration of the negotiation threads.

The security of most of the cryptosystems depend on the secret key generator. However, implementation for hardwares, of this key generator is inefficient because secret key generators depend on mathematical problem to generate the high randomness quality. Cellular automata (CA) pseudorandom number generator (PRNG) is more efficiently implemented rather than mathematical problem based PRNGs because a structure of CA PRNG is highly regular and simpler than the other PRNGs. In this paper, a virtual three-dimension (3D) CA PRNG based on the Moore neighborhood structure is proposed. The proposed PRNG uses new methods which are the rule numbering function that provides a high-quality randomness and cell position function that diminishes correlations between global states. In order to evaluate the quality of randomness, the ENT and DIEHARD test suites are used. The results of these tests show that the quality of randomness is superior to previous PRNGs.

The emCGA is a new extension of the compact genetic algorithm (CGA) that includes elitism and a mutation operator. These improvements do not increase significantly the computational cost or the memory consumption and, on the other hand, increase the overall performance in comparison with other similar works. The emCGA is applied to the problem of object recognition in digital images. The objective is to find a reference image (template) in a landscape image, subject to distortions and degradation in quality. Two models for dealing with the images are proposed, both based on the intensity of light. Several experiments were done with reference and landscape images, under different situations. The emCGA was compared with an exhaustive search algorithm and another CGA proposed in the literature. The emCGA was found to be more efficient for this problem, when compared with the other algorithms. We also compared the two proposed models for the object. One of them is more suitable for images with rich details, and the other for images with low illumination level. Both models seem to perform equally in the presence of distortions. Overall, results suggested the efficiency of emCGA for template matching in images and encourages future developments.

This paper presents a proposal based on an evolutionary algorithm for imputing missing observations in time series. A genetic algorithm based on the minimization of an error function derived from their autocorrelation function, mean, and variance is presented.

All methodological aspects of the genetic structure are presented. An extended description of the design of the fitness function is provided. Four application examples are provided and solved by using the proposed method.

Narrowband jammer excision is formulated as an optimization problem in this paper. Optimal filter weights are calculated/searched for by the computational intelligence techniques. We compare the error rate performance, complexity, and implementation issues of various computational intelligence techniques like Particle Swarm Optimization (PSO), Genetic Algorithm, and Least Mean Square (LMS). These techniques update the excision filter weights iteratively till the convergence criteria has been achieved. Bit Error Rate performance shows these techniques effectively suppress the Narrow Band Interference. It has been observed that PSO-based algorithm with tuned parameters outperforms other schemes of PSO and the other algorithms. It approaches the optimum performance in fewer iterations with ease in implementation.

Many attributes contribute to product failures that result in warranty claims. In particular, there are situations where several attributes are used together as criteria for judging the warranty eligibility of a failed product. For example, automobiles warranty coverage has both age and mileage limits. The warranty policy characterized by a region in a two-dimensional plane with one axis representing product age and the other axis representing product usage is known as the "two-attribute" warranty policy. A number of procedures have been developed for analyzing the two-dimensional warranty policy. These procedures use many crisp data obtained from strictly controlled reliability tests. However, in real situations, these requirements might not be fulfilled. In extreme cases, the warranty claims data come from users whose reports are expressed in a vague way. This may be due to subjective and imprecise perception of failures by a user, imprecise warranty data record, or imprecise rate of usage record. This paper suggests fuzziness as an alternative to randomness for describing the two-dimensional warranty uncertainty. A new sets-as-points geometric view of fuzzy warranty sets is developed in this study. This view can reduce many errors of estimation and prediction of the cost associated with a variety of warranty policies including the "two-attribute" warranties and some reliability improvement warranties.

A prediction scheme for Ethernet traffic data using a Multiscale-Bilinear Recurrent Neural Network with Adaptive Learning (M-BRNN-AL) is proposed and presented in this paper. The proposed predictor integrates an M-BRNN and an AL algorithm. In M-BRNN, the wavelet transform is employed to decompose the original traffic signals into several simple traffic signals. A BRNN is then used to predict each decomposed traffic signal. An AL algorithm is also applied in order to improve the learning process at each resolution level in M-BRNN-AL. Experiments and results on a set of Ethernet network traffic predictions show that the proposed scheme converges faster and archives better prediction performance than the other conventional models such as the Multi-layer Perception Type Neural Network, BRNN, and the original M-BRNN in terms of the normalized mean square error.

Graphics processing unit (GPU) has surfaced as a high-quality platform for computer vision-related systems. In this paper, we propose a straightforward system consisting of a registration and a fusion method over GPU, which generates good results at high speed, compared to non-GPU-based systems. Our GPU-accelerated system utilizes existing methods through converting the methods into the GPU-based platform. The registration method uses point correspondences to find a registering transformation estimated with the incremental parameters in a coarse-to-fine way, while the fusion algorithm uses multi-scale methods to fuse the results from the registration stage. We evaluate performance with the same methods that are executed over both CPU-only and GPU-mounted environment. The experiment results present convincing evidences of the efficiency of our system, which is tested on a few pairs of aerial images taken by electro-optical and infrared sensors to provide visual information of a scene for environmental observatories.

In this paper, two different methodologies respectively based on an unsupervised self-organizing (SOM) neural network and on a graph matching are shown and discussed to validate the performance of a new 3D facial feature identification and localization algorithm. Experiments are performed on a dataset of 23 3D faces acquired by a 3D laser camera at eBIS lab with pose and expression variations. In particular results referred to five nose landmarks are encouraging and reveal the validity of this approach that although low computational complexity and the small number of landmarks guarantees an average face recognition performance greater than 80%.

Image segmentation is a preliminary stage in diagnosis tools and the accurate segmentation of medical images is crucial for a correct diagnosis by these tools. Sometimes, due to inhomogeneity, low contrast, noise and inequality of content with semantic, automatic methods fail to segment image correctly. Therefore, for these images, it is necessary to use user help to correct method's error. We proposed to upgrade FAST FCM method to use training data to have more accurate results. In this paper, instead of using pixels as training data which is usual, we used different gray levels as training data and that is why we have used FAST FCM, because the input of FAST FCM is gray levels exist in image (histogram of the image). We named the new clustering method improved fast fuzzy C-mean (FCM). We use two facts to improve fast FCM. First, training data for each class are the member of the class. Second, the relevance distance of each input data from the training data of a class show the distance of the input data from the class. To cluster an image, first, the color image is converted to gray level image; then, from histogram of image, user selects training data for each target class, afterwards, the image is clustered using postulated clustering method. Experimental result is demonstrated to show effectiveness of the new method.

Human brains exhibit a possibility to control directly the intelligent computing applications in form of brain computer/machine interfacing (BCI/BMI) technologies. Neurophysiological signals and especially electroencephalogram (EEG) are the forms of brain electrical activity which can be easily captured and utilized for BCI/BMI applications. Those signals are unfortunately usually very highly contaminated by external noise caused by the presence of different devices in the environment creating electromagnetic interference. In this paper, we first decompose each of the recorded channels, in multichannel EEG recording environment, into intrinsic mode functions (IMF) which are a result of empirical mode decomposition (EMD) extended to multichannel analysis. We present novel and interesting results on human mental and cognitive states estimation based on analysis of the above-mentioned stimuli-related IMF components. The IMF components are further clustered for their spectral similarity in order to identify only those carrying responses to present stimuli to the subjects. The resulting targets only reconstruction allows us to identify when and to which stimuli intelligent application user is tuning at a time.

In this paper, we propose a multi-objective optimization framework for SVM hyperparameters tuning. The key idea is to manage a population of classifiers optimizing both False Positive and True Positive rates rather than a single classifier optimizing a scalar criterion. Hence, each classifier in the population optimizes a particular trade-off between the objectives. Within the context of two-class classification problems, our work introduces "the receiver operating characteristics (ROC) front concept" depicting a population of SVM classifiers as an alternative to the receiver operating characteristics (ROC) curve representation. The proposed framework leads to a noncost sensitive SVM training relying on the pool of classifiers. The comparison with a traditional scalar optimization technique based on an AUC criterion shows promising results on UCI datasets.

Previous studies on tumor classification based on feature extraction from gene expression profiles (GEP) were proven to be effective, but some of such methods lack biomedical meaning to some extent. To deal with this problem, we proposed a novel feature extraction method whose experimental results are of biomedical interpretability and helpful for gaining insight into the structure analysis of gene expression dataset. This method first applied rank sum test to roughly select a set of informative genes and then adopted factor analysis to extract latent factors for tumor classification. Experiments on three pairs of cross-platform tumor datasets indicated that the proposed method can obviously improve the performance of cross-platform classification and only several latent factors, which can represent a large number of informative genes, would obtain very high predictive accuracy on test set. The results also suggested that the classification model trained on one dataset can successfully predict another tumor dataset with the same tumor subtype obtained on different experimental platforms.

Machine learning is widely applied to gene expression profiles based molecular tumor classification, but sample imbalance problem is often overlooked. This paper proposed a subclass-weighted neighborhood classifier to address the imbalanced sample set problem and a novel neighborhood rough set model to select informative genes for classification performance improvement. Experiments on three publicly available tumor datasets demonstrated that the proposed method is obviously effective on imbalanced dataset with obscure boundary between two subtypes and informative gene selection and it can achieve higher cross-validation accuracy with much fewer tumor-related genes.

This article introduces a new kind of self-adaptation in discovery mechanism of learning classifier system XCS. Unlike the previous approaches, which incorporate self-adaptive parameters in the representation of an individual, proposed model evolves competitive population of the reduced XCSs, which are able to adapt both classifiers and genetic parameters. The experimental comparisons of self-adaptive mutation rate XCS and standard XCS interacting with 11-bit, 20-bit, and 37-bit multiplexer environment were provided. It has been shown that adapting the mutation rate can give an equivalent or better performance to known good fixed parameter settings, especially for computationally complex tasks. Moreover, the self-adaptive XCS is able to solve the problem of inappropriate for a standard XCS parameters.

Ant Miner algorithms are used for the classification task of data mining. This paper proposes a new version of Ant Miner algorithm called CAnt Miner that considers compatibility between the attributes, when constructing a rule. In addition to a new heuristic function based on compatibility among attributes, we also introduce a strategy to dynamically stop adding terms in the rule antecedent part instead of using a user specified threshold that specifies when to stop adding terms. We compare the performance of proposed approach with other Ant Miner versions and C4.5 in nine public domain data set. The experiment results show that the proposed approach achieves higher accuracy rate then other approaches.

Performance related problems are becoming more and more strategic in the software development, especially recently with the advent of Web Services and related business-oriented composition techniques (software as a service, Web 2.0, orchestration, choreography, etc.). In particular, an early integration of performance specifications in the SDP has been recognized during the last few years as an effective approach to improve the overall quality of a software. The goal of our work is the definition of a software development process that integrates performance evaluation and prediction. The software performance engineering development process (SPEDP) we specify is focused on performance, which plays a key role driving the software development process, thus implementing a performance/QoS-driven (software) development process. More specifically, in this paper our aim is to formally define the SPEDP design process, posing particular interest on the basis, on the first step of SPEDP, the software/system architecture design, modeling and/or representation. We define both the diagrams to use and show how to model the structure of the software architecture, its behavior and performance requirements. This is the first mandatory step for the automation of the SPEDP into a specific tool, which we have partially implemented as a performance plug-in for ArgoUML, ArgoPerformance.