International Journal of Intelligent Engineering and Systems

8

Authors: Dennis Arturo Ludeña Romaña, Shinichiro Kubota, Kenichi Sugitani, Yasuo Musashi

We carried out an entropy study on the DNS query traffic from the outside of a university campus network to the top domain DNS server when querying about reverse resolution on the PC room terminals through April 1st, 2007 to April 30th, 2008. The following interesting results are given: (1) The total DNS query traffic changes in a mild manner until January 16th, 2008, however it drastically changes after January 17th, 2008. (2) In January 17th, 2008, the DNS query traffic is mainly dominated by several specific IP addresses as their query keywords. (3) We carried out forensic analysis on the PC room terminals in which IP addresses are found in the several specific keywords and it is concluded that the PCs become spam bots when inserting USB based key disk storage.

8

Authors: Qiongshui Wu, Libo Zeng, Ningning Guo

Counting of different classes of white blood cells in bone marrow smears can provide pathologists with valuable information regarding various hematological disorders. For automatic imaging analysis techniques, precise segmentation of White Blood Cells is quite challenging due to the complex contents in bone marrow smears. Far more different from traditional color imaging analysis method, we introduced multispectral imaging techniques. After an image with high quality was acquired, the spectrum of each pixel was directly fed into a trained Support Vector Machine (SVM) for classification, then morphological binary operations were performed to correct the small error-classified regions. Masses of experiments show that our multispectral imaging method is device-independent to a large extent, and the segmentation results are highly satisfactory and inspiring. It proves that the introduction of multispectral imaging analysis techniques into White Blood Cells detection is a success. Multispectral imaging analysis is a promising technique in biomedicine.

10

Authors: Suojun Lu, Jian'an Fang, Qingying Miao

The topology of a network often plays a crucial role in determining its dynamical features. There is increasing interest in trying to understand the relationship between the structural properties of networks and their behaviors. However, this problem has not been fully considered in many neural networks, which mimic certain biological processes. Here, we construct a recurrent network model to simulate the decision-making process of brain. In this model, we examined the effects of network topology on the performance of decision-making by constructing three different topological networks: the regular network, the random network, and the small-world network. We found that the regular network and the small-world network show significant better performance of decision-making than that in the random network when the internal noise of the networks is low. However, following the increase of the internal noise, the random network, instead of two other networks, shows better ability to resist the noise. Finally, to mimic neurodegeneration or neural injury, we introduced two types of neuronal damages: clustered damage and distributed damage. We found that three networks exhibit different network behaviors in the case of neuronal damages. The regular network and the small-world network display severe decrease of the performance in the distributed damage pattern, but not in the clustered damage pattern. The random network shows similarly gradual decrease of the performance in both damage patterns. Furthermore, the small-world network shows the best performance in the high levels of distributed damage. Together, our results indicate that network topology significantly influence the network behaviors in our model of decision-making.

7

Authors: Yan Liu ,Kai Liu , Wei-liang Tao, Ye Zhu

Compared to other methods, insulator’s contact angles detection can obtain more precise and more accurate hydrophobicity in parts of insulator, getting in-depth information like hydrophobicity of shed material, migration of the hydrophobicity after polluted insulator, attenuation and recovery characteristics of the insulation performance under certain conditions, all of which are its particular advantages. A new detection method of insulator’s performance is proposed in this paper with measuring contact angles. With this method, the boundary curve of water drop on insulator’s surface was extracted with image processing, then the polynomial fitting was deployed to fit the boundary curve, and contact angles of water drop were calculated by differentiate the polynomial deduced from the polynomial fitting. So the performance of the insulator can be detected and the process doesn’t need manual intervention which makes the detect result accurate and objective. Experiments show that the root mean square error of the measurement is below 2%.

8

Authors: Yan Liu, Wei Liangtao, Kai Liu, Ying Xu

Recently, several data storage schemes have been developed to store massive sensor data in the wireless sensor networks (WSNs). A crucial task of WSNs is to disseminate useful information to users efficiently and handle the data storage architecture incorporating their extreme resource constraints. A Distributed Index based Multi-resolution Storage Architecture (DIMSA) for data storage and retrieval at multiple levels of detail in WSNs is proposed in this paper. The proposed scheme could automatically disseminate useful information to users in correspondence to the user defined level of detail to save energy. In the condition that some attributes are included in many events, the values for these attributes have to be replicated and stored at different places in the network for each individual event in traditional data-centric storage schemes. This will consume some extra energy. To address this problem, DIMSA deploys an attribute allocation scheme for finding a good distribution of attributes to the sensor networks. In addition, a distributed index scheme is also proposed to avoid query flooding and extra data transmission. Analysis and simulations are conducted to evaluate the performance of the proposed scheme. The results show that DIMSA outperforms the traditional data-centric storage scheme like GHT.

6

Authors: Yuchen Yan, Qinghu Chen, Wei Deng, Feng Yuan

In this paper, a spectral feature extraction method based on fast Fourier transformation for writer identification is presented. According to the constructibility of the texture image of handwriting, we put forward an estimation method for mathematical expectation value of texture image’s spectral features. This method eliminates the randomness of spectral features and gets stable spectral features. The experiment results show that this approach enhances the identification accuracy to a large extent using data sets with large handwriting samples.

8

Authors: Kee-Baek Kim, Jong-Su Kim, Sangkeun Lee, Jong-Soo Choi

Image registration has been widely used in many image processing-related works. However, the exiting researches have some problems: first, it is difficult to find the accurate information such as translation, rotation, and scaling factors between images; and it requires high computational cost. To resolve these problems, this work proposes a Fourier-based image registration using pyramid edge images and a simple line fitting. Main advantages of the proposed approach are that it can compute the accurate information at sub-pixel precision and can be carried out fast for image registration. Experimental results show that the proposed scheme is more efficient than other baseline algorithms particularly for the large images such as aerial images. Therefore, the proposed algorithm can be used as a useful tool for image registration that requires high efficiency in many fields including GIS, MRI, CT, image mosaicing, and weather forecasting.

5

Authors: Dennis Arturo Ludeña Romaña, Kenichi Sugitani, and Yasuo Musashi

The DNS query traffic from the inside and the outside of the campus network in a university was statistically investigated through January 1st, to July 31st, 2007. The following interesting results are obtained, as follows: (1) The unique source IP address-based entropy value is usually less than the unique DNS query keyword based one in the DNS query traffic from the campus network, however, the unique source IP address-based entropy value is greater than the unique DNS query keyword based one in the DNS query traffic from the outside of the campus network. (2) Two types of entropy changes were found in the unique source IP addresses- and the unique DNS resolution query keywords. In the both entropies, one is a parallel change, and another one is a symmetrical one. Although the latter change type can be conventionally observed in 2006, the former change type can recently observed in 2007. Therefore, it can be concluded that the recent spam bots send a lot of spam E-mails to the next victim PCs via the local vulnerable E-mail servers.

8

Authors: Kei Eguchi, Takahiro Inoue, Ichirou Oota, Hongbing Zhu, Fumio Ueno

For mobile applications, a switched-capacitor (SC) DC-DC converter which offers a 2x/ 1.5x/ 1x mode of operation is presented, and its properties are analyzed theoretically in this paper. Aimed at back-lighting applications, battery applications, etc., the proposed converter provides stepped-up/stepped-down voltages. Compared with the conventional approaches employing doubler circuits, the proposed converter can realize high efficiency and flexibility of output voltages. The validity of the proposed converter is confirmed by SPICE simulations and theoretical analyses. Concerning the power efficiency and the optimal duty factor, the confirmation result shows that the theoretical results correspond to the simulation results well. Furthermore, under the conditions that the input voltage Vin=3.7V, the output load RL, and the output voltage Vout=3.3 - 5.3V, the proposed converter can improve the power efficiency more than twice from that of the conventional circuit.

8

Authors: Rong Chen, Yongfeng Cao, Hong Sun, Wen Yang

Support vector machine is a new technique of machine learning developed from the middle of 1990s. There are two important things in the process of face recognition using SVM. One is the feature vectors extracted from face images, the other is the classifier we choose. Our approach optimizes traditional method from these two points. The adaptively weighted Fisherface purposes on a better feature extraction; the size condensing of training data is used to speed up the modeling of the SVMs; and binary tree structure SVMs is selected here to extend SVM to multi-class cases without much additional complexity. We do our experiments on ORL face database and get good recognition results. The accuracy of recognition results are compared to show the influence of the dimension of the Fisherface. At last, the benefit of adaptively weighted Fisherface is shown by a compare with the traditional Fisherface.

6

Authors: Longxin Lin, Haijun Xu, Haibin Xie, Lincheng Shen

Bionic underwater robot is a new-type underwater robot which propelled and controlled by several bionic fins. Heaving motion is an indispensably function for most bionic underwater robots. Bionic bladder system is an apparatus which can realize both heaving and pitching control. It was inspired by biologic fish’s bladder. It was composed of a cylinder, a piston, a pipe, a rolling guide, two ball screws and two servomotors. The dynamic models of the bionic bladder system for heaving and pitching motion were established and based on which a dual-velocity control system was designed. Two control channels realized decoupling by setting different control cycles, and a PID controller are designed for each control channel. Simulations with the real design parameters and different control cycles were performed. Results indicated that the bionic bladder system can get a good control performance with proper control parameters.

7

Authors: Huahua Chen, Xueyi Ye

3D map building is an important task of autonomous land vehicle(ALV) for obstacle detection, path planning. This paper analyzes the defects of the traditional 3D map building method, and then proposes a novel method based on projection of virtual height line(PVHL). The method projects virtual height line(VHL) into the stereo vision system by transforming the world coordinate of each point on VHL into its image coordinate, and generates two projection lines(PLs) in two images. On the two PLs, only the correspondence that is the projection of the intersection point of VHL and the surface of the scene has the maximum similarity metric. Then finding the height of the scene point is equivalent to proving whether the real correspondence of VHL has the maximum similarity metric or not. Experiments show that the proposed method is effective. And compared with that of the traditional, the results show that the proposed avoids unnecessary computation time and has better performance than the traditional, and can build the interested region only.

9

Authors: R A. Raja Mahmood , A H. Muhamad Amin , and A. I. Khan

A lightweight, fast and efficient pattern classifier based on Distributed Hierarchical Graph Neuron (DHGN) is proposed and implemented. A classifier with such features is essential for the emerging networks such as wireless sensor networks and mobile ad hoc networks, where resources such as bandwidth and energy are limited. The proposed classifier adopts an in-network processing algorithm and has a one cycle learning capability. DHGN network is a new form of neural networks which consist of a hierarchical graph-based representation of input patterns. This paper compares the proposed solution with the well known Self-Organizing Map (SOM) classifier, in relation to accuracy and computational complexity. The results show that our solution offers lower computational complexity than SOM while guaranteeing satisfactory accuracy.