Table of Content

08 August 2015, Volume 0 Issue 8

Previous Issue    Next Issue

Improved Microscopic Cell Image Segmentation Algorithm Based on Template Matching

GE Liang1,2, YU Ka1,2

2015, 0(8):  1-7,12.  doi: 10.3969/j.issn.1006-2475.2015.08.001

Asbtract ( 187 )  

References | Related Articles | Metrics

 Image segmentation algorithm based on template matching has good generality in the microscopic cell image segmentation. In view of the traditional template matching microscopic cell image segmentation algorithm in producing template set produced more redundant templates, which led to a rather long time of image segmentation, an improved template matching microscopic cell image segmentation algorithm was proposed. This algorithm extracted shape feature and calculated similarity of template set on the basis of the traditional template matching algorithm. Then, this algorithm eliminated the more similar templates to reduce template set in the case of least affecting the accuracy of image segmentation. Finally, the reduced template set was used to complete the segmentation. The experiments on U2OS and NIH3T3 image sets show that the improved algorithm is comparable with traditional algorithm in accuracy, and has faster speed and better performance compared with traditional algorithm in image segmentation.

Segmentation Method for Liver Tumor Based on 3D ROI and FCM

JIN Kai-cheng, WANG Yi, ZHENG Shen-hai, OUYANG Zi-peng

2015, 0(8):  8-12.  doi:10.3969/j.issn.1006-2475.2015.08.002

Asbtract ( 190 )  

References | Related Articles | Metrics

 In Computed Tomography (CT) scans of liver it exists many noises. Besides, liver tumor’s gray level is very close to the liver and the tumor has fuzzy boundaries, it is hard to be segmented. During liver tumor segmentation, the traditional level set method is sensitive to initial contours and needs to adjust the parameters manually, and the time complexity is high. According to liver tumor’s fuzziness, this paper proposed a new segmentation method for liver tumor based on three dimension region of interest (3D ROI) and spatial fuzzy c-means clustering (FCMS). First it picks the ROI in three dimensions, then uses FCMS to segment the tumor, then does morphology operation, in the end uses variational B-spline level sets method to smooth the contour. The result of test turns out that, the method proposed in this paper gets better result and higher efficiency, which is also easy to operate.

An Image Representation Computation Model Based on Dynamic Adjustment Mechanism of Non-classical Receptive Fields

FAN Yi-na, LANG Bo, HUANG Jing

2015, 0(8):  13-18.  doi: 10.3969/j.issn.1006-2475.2015.08.003

Asbtract ( 174 )  

References | Related Articles | Metrics

This paper utilizes the physiological mechanism of non-classical receptive field of ganglion cell to design a hierarchical network model for image representation based on neurobiology. It is different from the contour detection, edge detection, and other practices using the classical receptive fields, simulating the non-classical receptive fields physiological mechanism which can be dynamically adjusted according to stimulation for image local segmentation and compression based on image neighborhood region similarity, thus realizes the inner image representation in neural representation level and convenients for extract the semantics of image further. We provide extensive experimental evaluation to demonstrate that, comparing with the traditional methods using single pixel, GC-array can represent image with a low cost. Most importantly, the GC-array model provides a basic infrastructure for image semantic extraction.

 

Online Automatic Recognition and Diagnosis of Electrical Devices via Thermal Panorama

LYU Jun1, WANG Fu-tian1,2, TANG Jin1,2, LUO Bin1,2

2015, 0(8):  19-23.  doi:10.3969/j.issn.1006-2475.2015.08.004

Asbtract ( 212 )  

References | Related Articles | Metrics

Aiming at solving the problem of online automatic recognition and thermal diagnosis of electrical devices, this paper proposes an online automatic device recognition and diagnosis method based on the stitched thermal panorama. Firstly, we collect some thermal infrared images of the scene and stitch them based on SIFT to construct the thermal panorama in an offline fashion. The categories of devices in the scene are further manually annotated. Secondly, given a thermal image captured online, we locate and identify the devices of this thermal image by utilizing the feature matching algorithm, and then effectively diagnose these devices based on their categories. Finally, experiments on the real substation scenarios demonstrate the effectiveness and practicability of the proposed approach.

Application of PCA and Eros Affinity Propagation Clustering in Financial Data Sets

LIAO Hong-yi， WANG Xin

2015, 0(8):  24-28.  doi:10.3969/j.issn.1006-2475.2015.08.005

Asbtract ( 202 )  

References | Related Articles | Metrics

The multi-dimensional characteristics and high noise characteristics of the financial data set make it hard to analyze the time series. This paper puts forward an algorithm based on the principal component analysis and the Eros affinity propagation clustering. First it uses the principal component analysis method to extract the main eigenvalues of the multivariate financial time series data; then uses the Eros affinity propagation clustering to analyze the extracted eigenvalues. This kind of clustering algorithm can make the individual data as an attribute of the original data, through iterate competition to achieve optimal, do not need to find the number of clusters. The research results show that, this integrated method greatly reduces the dimension of the time series, and has a highly correct classification rate. It proves that this algorithm is very effective.


Design and Simulation of Hydraulic System in Paper Machine Based on Automation Studio

WU Xin-sheng1,2, YANG Dong-shan3

2015, 0(8):  29-31,66.  doi: 10.3969/j.issn.1006-2475.2015.08.006

Asbtract ( 150 )  

References | Related Articles | Metrics

A hydraulic system provides the press rollers with stable and reliable high-pressure to finish wet paper’s dewatering by squeezing in paper machine. A hydraulic system in paper machine was designed by Automation Studio. The working processes including falling and pre-pressing, pressure increasing, lifting and reset of top rollers were simulated. The simulation results show that the system could finish squeezing work. The results of the hydraulic system tested in a paper machine are basically the same as the simulation results. This design and simulation method can accurately simulate the systems working state. The hydraulic system is of simplified modeling process and conveniently adjusted parameters. It is a good tool for designing and working optimization of a hydraulic system.

HBase Model Transformation Based on Meta-model

QIN Si-feng, GU Ping, ZHANG Chao

2015, 0(8):  32-37.  doi:10.3969/j.issn.1006-2475.2015.08.007

Asbtract ( 216 )  

References | Related Articles | Metrics

Model Driven Architecture (MDA) can reduce the influence of requirement change on software engineering, and improve the efficiency of software development and portability and maintainability of the system at the same time. The meta-models of Platform Independent Model (PIM) and Platform Specific Model (PSM) were built based on MDA. The idea of model transformation from UML class diagram to HBase model was proposed on the meta-model level. At last, the transformation was implemented through the ATL model transformation framework, which verified the feasibility of automatically generating database target model from model transformation and the feasibility of applying MDA in software engineering to some extent.


Storage Life Forecasting for Missiles Based on Improved Gray Model and RBF Optimization Model

]XU Ting-xue 1, ZHU Hui-chuan2, DONG Qi2

2015, 0(8):  38-42.  doi:10.3969/j.issn.1006-2475.2015.08.008

Asbtract ( 192 )  

References | Related Articles | Metrics

The combined forcasting model based on the gray model and RBF neural network optimization model was proposed, for solving the problem of little failure data in storage period and difficult to forecast, and the combined model was established to forecast the storage life of missiles. The result shows that the combined model is better than the single forecasting model, and is regarded of high practice value.


Design of Image Processing System Platform Based on Zynq Chip

LIU Hong, FU Yi-de

2015, 0(8):  43-47.  doi: 10.3969/j.issn.1006-2475.2015.08.009

Asbtract ( 193 )  

References | Related Articles | Metrics

 From the angle of the chip, the characteristics of the current image processing system were deeply analyzed. Aiming at the shortcomings of the current embedded image processing system based on ARM that in the anterior the acquisition speed is slow and in the middle the image processing algorithms are not easy to be accelerated by FPGA hardware, in order to improve the speed of the anterior acquisition, also in order to facilitate various image processing algorithms to be accelerated by FPGA hardware in the middle and the design of the transmission and display in the posterior, Zynq chip whose internal integrated ARM and FPGA architecture was selected, a image processing system platform prototype was designed by software and hardware co-design method, and the performance of the system platform was analyzed. Test results show that, the anterior acquisition speed of the system platform compared with the use of pure ARM chip solution is improved by 234 times, the system platform also provides extensible FPGA hardware accelerated channel for the middle image processing algorithms, and also provides good support for the posterior results show.

A Review on Multi-Label Learning

LIANG Wei-chao， SONG Bin

2015, 0(8):  48-56.  doi:10.3969/j.issn.1006-2475.2015.08.010

Asbtract ( 364 )  

References | Related Articles | Metrics

Multi-label learning is different from the traditional supervised learning. It is a framework which is proposed to represent objects which might have multiple semantic meanings simultaneously in the external world. Under this framework, an instance might be associated with a set of labels. Over the past decades, a lot of research results of multi-label learning have been achieved and gotten extensive application. This paper provides a systematic and detailed review in this area. Firstly, fundamentals on multi-label learning including formal definition and evaluation metrics are given. Secondly, some important and classic multi-label learning algorithms are presented. Finally, some valuable research directions in this area are discussed.


Solar Energy Monitoring System for Oil Pumping Units Based on Single-chip Microcomputer

CHEN Xue-song, ZHANG Hui-yu, ZHAO Zi-hao

2015, 0(8):  57-60.  doi:10.3969/j.issn.1006-2475.2015.08.011

Asbtract ( 194 )  

References | Related Articles | Metrics

To overcome the shortcomings of current oil production industry, in which monitoring of beam-pumping units at night consumes a lot of human resources and material resources, this paper uses solar energy as the  energy source, single-chip microcomputer STC12C5A60S2 as the intelligent core module to design an oil pumping units monitoring system. It is used for real-time monitoring of the oil pumping unit at night working state. The system not only realizes the remote monitoring of the oil pumping unit, but also saves energy and protects environment. It is small, portable, low cost, and is conducive to large-scale use of oilfield surface mining.


A WSN Location Algorithm Against Wormhole Attacks

GE Jie-li, LIU Yuan

2015, 0(8):  61-66.  doi:10.3969/j.issn.1006-2475.2015.08.012

Asbtract ( 160 )  

References | Related Articles | Metrics

DV-Hop is a classic algorithm for WSN localization, which is applied in various conditions because of its simplicity. However, it is vulnerable to wormhole attack easily as DV-Hop uses the average distance of each hop. In order to overcome the shortage, we make the improvement of DV-Hop. Firstly, according to the characteristics of the wormhole attack, we filter out suspicious wormhole nodes, determine the location and replace it with other path hops between nodes, secondly, introduce the conception of secondary location, in the third period, take a weighted average based on all anchor nodes average distance per hop, and finally, use the weighted average as average distance of unknown nodes. The experimental result shows that the algorithm has a good performance of the localization, and the error is small. Moreover, it can resist wormhole attack to some extent.


Application of Local Outlier Mining Based on Cluster Merging Algorithm in Intrusion Detection

MEI Xiao-hui, LONG Yuan, ZHANG Jian-bo

2015, 0(8):  67-70.  doi: 10.3969/j.issn.1006-2475.2015.08.013

Asbtract ( 135 )  

References | Related Articles | Metrics

According to the problem of high dimension of network security data, the traditional outliers based on cluster cannot effectively detect network intrusion behavior data in detail. This paper put forward an improved DBSCAN algorithm of outlier mining called LDBSCAN-CM. First, this paper introduced a concept of local outlier mining for traditional DBSCAN algorithm, calculated local outlier factors of candidate objects, and generated a number of clusters. Next, this paper merged clusters in order to improve the mining efficiency. Eventually, the KDD Cup99 dataset was applied to conduct simulation experiment on the application of the improved algorithm in intrusion detection. The results indicate that the improved algorithm LDBSCAN-CM can guarantee higher detection rate and lower false alarm rate.


Network Security Situation Assessment Based on Combination Model

ZHANG Tian-dan

2015, 0(8):  71-74.  doi:10.3969/j.issn.1006-2475.2015.08.014

Asbtract ( 172 )  

References | Related Articles | Metrics

In order to obtain more ideal estimate effect of network security situation, this paper proposed an estimation model based on network security situational combination method. Firstly, samples of network security situation were collected and processed to get the learning samples. Secondly, the training sample sets were input to the BP neural network to learn, and the cuckoo search algorithm was adopted to select the most reasonable parameters of the BP neural network. Finally the simulation experiments were used to analyze the performance of model. The results show that, the proposed model greatly reduces the fitting error and the prediction error of network security situation, it is a scientific, reasonable estimation model of network security situation, and the estimation results have a certain practical application value.


Research and Application of Rule Engine on Access Control

JI Kai-wei, LE Hong-bing

2015, 0(8):  75-79.  doi: 10.3969/j.issn.1006-2475.2015.08.015

Asbtract ( 120 )  

References | Related Articles | Metrics

In order to apply the ideal of separating strategies from mechanism to access control, we proposed a kind of rule engine technology in access control model. The access control strategies were represented by rules. In the engine, attributes were loaded on real-time and rules were executed dynamically by Java reflective technology, which gets better flexibility. Then this paper proposed a rule matching algorithm based on the sharing node in multi-rule by researching on the Rete algorithm, which can improve the matching efficiency.


Hadoop Speculation Execution Algorithm in Heterogeneous Environments

QI Peng-nian, ZHU Jin, HAO Jun-hui, XU Feng-ping

2015, 0(8):  80-83,88.  doi:10.3969/j.issn.1006-2475.2015.08.016

Asbtract ( 185 )  

References | Related Articles | Metrics

This article researches and analyzes the poor performance of the Hadoop speculation execution algorithm in heterogeneous environments, and puts forward a new improvement algorithm after researching source code deeply. The new algorithm can adjust the execution of backup task automatically to make it balanced according to system load condition, and get more precise stragglers queues using the way of putting the residual time value greater than 0.2 in task queue to judge the stragglers, based on the historical average completion time proposed by Zaharia. The new algorithm to a certain extent improves the performance of speculation execution in the heterogeneous environments.


Web Page Title Real-time Extraction Method Based on Hyperlink and DOM Tree

ZHANG Bing1， TANG Jin1,2， LUO Bin1,2

2015, 0(8):  84-88.  doi:10.3969/j.issn.1006-2475.2015.08.017

Asbtract ( 211 )  

References | Related Articles | Metrics

Correct extraction of Web title is significant to Web text information mining. This paper proposed a method which can get a real-time Web page title extraction. This method first used a real-time analysis model though the catalog page, and then used the hyperlink-based travelsal approach, and used the correspondence between the title and the release time to get the URL of the page and the corresponding anchor text. If the anchor text we have was not the title of the text page, we should get the Web page HTML source code and build a DOM tree for the corresponding theme-based Web page. Based on the visual characteristics of the Web page title, we traversed the DOM tree in depth-first order. The experimental results demonstrate that this method is of high accuracy and can be simply implemented and so on.


Question Recommendation Mechanism in Community Question Answering Systems

JIANG Zong-li， LI Li-xin

2015, 0(8):  89-92.  doi:10.3969/j.issn.1006-2475.2015.08.018

Asbtract ( 203 )  

References | Related Articles | Metrics

In the community question answering system, there are so many questions that lead to the user very difficult to find his interested in and be good at questions to answer. In order to solve this problem, this paper based on user’s interest and then put his recent activity into the recommendation algorithm. The experiment results show that this modified method can improve the efficiency of algorithm to some degree.


Text Classification Model Based on Cooperation of Dual Features of Concept and Root

GU Ping, WU Ting-jun, WEN Jing-yun

2015, 0(8):  93-97.  doi:10.3969/j.issn.1006-2475.2015.08.019

Asbtract ( 180 )  

References | Related Articles | Metrics

Traditional semi-supervised text classification methods were built based on the features of root, however, the common disadvantage of neglecting the importance of semantic features resulted in low precision of classification. In order to take account of the influence of semantic on classification, a text classification model comprehensively making use of dual features of concept and root was brought forward. Under the framework of cooperative training, this algorithm considered WordNet as ontology library and built double classifiers based on both concept and root for cooperative training. Through experiments, we analyzed the accuracy rate and recall rate of new classification model, and the results showed the promotions of both accuracy rate and recall rate in new model comparing with old model. It indicates that the new algorithm based on cooperation of dual features of concept and root is more effective than the old algorithm.


A Collaborative Filtering Algorithm Based on Radial Basis Function Interpolation and SVM

ZHAN Zeng-rong, ZENG Qing-song

2015, 0(8):  98-103.  doi:10.3969/j.issn.1006-2475.2015.08.020

Asbtract ( 282 )  

References | Related Articles | Metrics

 Due to the fact that the data sparseness leads to the low accuracy of a collaborative filtering recommendation system, this paper proposed a collaborative filtering algorithm based on Radial Basis Function (RBF) interpolation and SVM. The algorithm first uses the RBF interpolation to fill the missing data in training data set, and then a SVM classifier is introduced to predict the label for testing data by using the interpolated data set as training set. The test result shows that the method overcomes the impact of data quality on the recommended algorithm, and it outperforms other methods in accuracy and stability.


A Dynamic Link Travel-time Model by Signal Controlling

WEN Kai-ge

2015, 0(8):  104-106,111.  doi:10.3969/j.issn.1006-2475.2015.08.021

Asbtract ( 133 )  

References | Related Articles | Metrics

The vehicles travel time was divided into multiple components, and the trajectories of vehicles driving in steady mode were classified and their travel times were considered respectively. A new dynamic link travel time model was proposed. This new model is simple and can be applied to forecast dynamic link travel time in real traffic network. An example of the analysis shows that, the link travel time is not only simply an increasing function of the traffic flow, but also affected by both the entry time and the signal control scheme.


A Hierarchical Layout Algorithm Based on Graph Matching

ZHAO Yu-cong, ZHONG Zhi-nong, WU Ye, JING Ning

2015, 0(8):  107-111.  doi:10.3969/j.issn.1006-2475.2015.08.022

Asbtract ( 234 )  

References | Related Articles | Metrics

In order to draw the undirected graph with big and even degrees, we propose a hierarchical layout algorithm based on graph matching (GMH). Firstly, a series of coarser and coarser graphs are generated by graph matching. Secondly, the optimal FR layout for the coarsest graph can be found cheaply. In the end, the layout on the coarser graphs is recursively prolonged to the finer graphs by the mass-center method. The new algorithm can not only advance the efficiency of graph visualization and improve the effect of layout, but also show a series of hierarchical graphs.


An Image Quality Assessment Algorithm Based on Phase Feature for Wavelet Domain

LI Ai-hua, WANG Li-bin

2015, 0(8):  112-115.  doi:10.3969/j.issn.1006-2475.2015.08.023

Asbtract ( 174 )  

References | Related Articles | Metrics

For the pixel-based method for image quality assessment ignores the structural features of natural images, which fails to measure some particular distortion types. Therefore, an image quality assessment algorithm by using phase feature is proposed. This method is proposed based on the fact that human eye understands an image mainly according to its low-level features. First, the phase congruency is considered as the primary feature in computing the assessment function. Then, the modulus of wavelet transform is considered as the second feature in computing the assessment function. Finally, the whole image quality assessment value can be obtained based on these two features. Experimental results on standard image database demonstrate that the effectiveness of the proposed method. Meanwhile, it has a good consistency with the subjective assessment of human beings.


An Improved Distributed Lazy Associative Classification Algorithm

YANG Hao-min, MA Chao, WU Hai-yan

2015, 0(8):  116-120.  doi:10.3969/j.issn.1006-2475.2015.08.024

Asbtract ( 142 )  

References | Related Articles | Metrics

Distributed lazy associative classification algorithm (DLAC) refers to a lazy associative classification algorithm using distributed association rules mining. The existing DLAC algorithm has two main problems: one is the inefficiency of classifying multiple test samples; the other is that projection operation is not distributed. Hence, this paper proposed an improved distributed lazy associative classification algorithm—PDLAC algorithm. Firstly, it clustered the test samples using KMeans method， secondly, judged whether it satisfied the aggregating condition or not for each clustered test samples, if it satisfied, aggregated the clustered test samples, if not, let each of the clustered test samples to be one clustered test sample. Then, it executed distributed projection and mined association rules using C-DMA algorithm. Finally, it constructed classifier to classify one or more test sample at the same time. Experiments were conducted with setting the degree of parallelism to 15. The time consumption of PDLAC algorithm was far less than DLAC algorithm, and its performance was much better as the number of testing samples increased. The test results show that PDLAC algorithm is a good solution to both two problems mentioned above.


A Simplex Coding Multi-class Boosting Optimizing Algorithm

ZHAO Chun-lan

2015, 0(8):  121-126.  doi:10.3969/j.issn.1006-2475.2015.08.025

Asbtract ( 194 )  

References | Related Articles | Metrics

The multi-classification problem was divided into multiple independent binary problems in existing image classification mechanism, numbers of class directly influenced demand sizes of binary classifier. The number of classes in image classification problem was very large, which led to long training time, high computing demand and high test cost. In order to effectively solve these problems, this paper designed a multi-class boosting optimizing algorithm based on simplex coding(SCOBoost). Firstly, based on simplex coding, combining with the least squares support vector machine (LS-SVM) objective function, this paper proposed multi-classification improvement goal based on simplex coding; secondly, selected the weak classifiers which are not associated with the number of classes as the kernel function, and used iterative methods of boosting to solve. Experiment results on different data sets showed, SCOBoost not only had higher classification performance, but also had lower algorithm complexity, fast test speed and training time which is not affected by the number of classes and so on.
