مجله رایانش نرم و فناوری اطلاعات
سال چهارم شماره 4 (زمستان 1394)

Segmentation of medical images is one of the initial pre-processing for designing automated diagnosing systems. Magnetic Resonance Imaging (MRI) of the brain is associated with intensity uncertainty due to destructive artificial factors in the imaging process such as noise and Intensity Non-Uniformity (INU). As a result, segmentation of these images is a challenging issue. Due to uncertainty in brain MRI, researchers have employed fuzzy methods for segmenting brain MRI. BCFCM is one of the fuzzy segmentation methods in which information of neighboring pixels are also used for segmentation. This method has different parameters which inappropriate selection of that, greatly reduces the performance of the method. In this paper an optimized BCFCM with two structure is proposed for brain MRI segmentation. In the optimization process, GA and PSO are used. Appropriate performance of the proposed method is demonstrated by simulation results of standard Brain-Web datasets using Tanimoto and Dice similarity measures.

Data clustering is a basic tool for understanding the structure of data collections. The process puts the data into groups of similar objects is called clustering. Clustering is one of the main issues of unsupervised clustering to find the structure in a set of unlabeled data. Clustering algorithms can be divided into two categories according to the type of data: Clustering algorithms for numerical data and clustering algorithms for categorical data. The clustering algorithms for categorical data are more important than clustering algorithms for numerical data because of the nature and application of these data. In this paper, at first the nature of this type of data is described and then the clustering algorithms and similarity measures presented in this area are reviewed. Finally, a hybrid method is proposed based on the combination of the hierarchical clustering algorithm and the partitioning clustering algorithm. The experiments show that the proposed method in this paper improves the results of clustering.

Data streams outlier mining is an important and active research issue in anomaly detection. Outliers are large deviate from others data points. They are often not the errors, and may carry important information. Recently, many studies on outlier detection in the database are done. Many algorithms have been proposed to detect outliers, but most of them are effective on static data. As data streams evolve during the time, traditional methods cannot perform well on them. These algorithms often can lead us to a wrong decision. The false positive rate of the algorithms will be high. In this paper, an algorithm is proposed to divide the streams to pieces evenly and compute local outlier factor for every data. The proposed algorithm uses a list as candidate list for the outliers. The proposed algorithm detects outliers and unusual patterns by postponing at outlier detection. The experimental results on synthetic and real datasets show that the proposed algorithm was successful in reducing false positive rate and increasing its accuracy.

Close-range Photogrammetry is widely used in industrial applications for measuring size, shape and deformation of objects. In industrial photogrammetry, circular targets are often utilized to increase the automation as well as the accuracy of measuring process. Edge based and area based methods can be applied for locating the centroid of these targets. In this paper, an improved ellipse fitting technique is proposed to increase the accuracy and precision of determining center location of the targets. Various tests including size, orientation and elongation under varying conditions are performed on both simulated and real targets to evaluate the reliability of the proposed method. Obtained results proved the higher accuracy of the proposed method in comparison with the other traditional methods.

Routing is a critical issue in telecommunication networks. In routing algorithms, the changes in data packets by intermediate nodes are not allowed. Network coding is a new method for data transmission in telecommunication networks which has advantages such as the improvement in network performance and throughput. In this method, intermediate nodes send a combining code of incoming packets to the destination node. Opportunistic routing has the potential to substantially increase wireless network throughput. In this approach, no node is selected as the intermediate node, and each node which receives the data can transfer them to the destination. In this paper, a method is proposed based on opportunistic routing in network coding to increase the throughput of wireless networks. In the proposed method, the random linear coding of forwarding packets from a source is used; and decoding of the packets is not required in all intermediate nodes. In addition, a fixed path is not considered for data transmission and opportunistic routing is used. The performance of proposed methods is simulated using simulator NS-2 and compared with COPE and EXOR. The simulation results show that the proposed method improves the network throughput.

Terrestrial Laser Scanner (TLS) acquired 3D information¡ Intensity image and color image around settlement point simultaneously. The output of this device can be categorized intwo groups: 2D images and 3D point cloud. Physical and geometric properties of discontinuities and effects¡ as well as explain the position of the laser scanner and camera effects cause changes in the amount of reflected energy¡ lighting and explain the depth of the point cloud and the image is recorded. Such as machine vision¡ properties of 2D images and 3D point cloud are complementary in the field of surveying¡ can be combined and used these to understanding and objects detection. In this paper¡ we combined the image processing techniques in 2D and 3D data¡ for occlusion extraction. The point cloud and images recorded in K.N.T University was used for this purpose. Canny algorithm for edge detection in combination with Range Border Detection was used. This method has a high ability to find hidden areas¡ as one of the main problems is point cloud data. In the data sample that used¡ obtained 65 region¡ with a total area of 4719 square meters in the 8000 square meters scan area.