Volume 3 Number 3 December 2013
1 |
Protein Sequence Motif Detection Using Novel Rough Granular Computing Mode Abstract: Protein sequence motifs information is essential for the analysis of biologically significant regions. Discovering sequence motifs is a key task to realize the connection of sequences with their structures. Protein sequence motifs have the potential to determine the function and activities of the proteins. Many algorithms or techniques are used to determine motifs which require a predefined fixed window size. Our input dataset is extremely large as a result, an efficient technique is demanded. So we apply three different granular computing models to find protein motif information which transcend protein family boundaries. The constructed segments from 3000 protein sequences are divided into granules using Rough K-Means and then K-Means has been applied on each granule. The highly structured clusters are further considered to find motif patterns. This approach is compared with Adaptive Fuzzy Granular model. The proposed Rough Granular computing model generates more number of highly structured motif patterns. |
2 |
Effective Clustering Algorithm for Gas Sensor Array Drift Dataset Abstract: Much work has been done in the last fifteen years to develop adapted techniques and robust algorithms. The problem of data correction in presence of simultaneous sources of drift, other than sensor drift, should be also investigated since it is often the case in practical situations. To this, one idea could be combining semi-supervised methods able to learn the actual source of drift, which might clearly change with the measured samples, with adaptive drift correction strategies that can account for the continuous drift direction change in the feature space. Cluster validity checking is one of the most important issues in cluster analysis related to the inherent features of the dataset under concern. It aims at the evaluation of clustering results and the selection of the scheme that best fits the underlying data. This paper studies clustering methods K-Means, Fuzzy C-Means (FCM) and Rough K-Means algorithm have implemented for the Gas Sensor Array Drift Dataset without considering class labels. Then they obtained results are compared with the original class labels through the confusion matrix. It is found that the Rough K-Means is performing well comparatively to get the valid data from the drift dataset. |
3 |
Tree-Based Mining with Sentiment Analysis for Discovering Patterns of Human Interaction in Meetings: Tamil Document Abstract: Human interface is vital individuality of group social dynamics in conference. The order of human interaction is generally represented as a tree. Tree structure is used to capture how the person interacts in meetings and to find out the interactions. The human interaction are offering as an thought, giving comments, ask opinion, acknowledge, etc., Frequent interaction tree pattern mining algorithm and Frequent interaction sub tree pattern mining algorithm are utilized to analysis the structure and to extract interaction flow patterns, where co-occurring only the tags are considered. To conquer this problem, Sentiment Analysis (SA) is proposed work to the entire flow of interaction in meetings. A sentiment analysis approach extracts sentiments associated with opinions of positive or negative for specific subjects from the Tamil document instead of classifying the whole Tamil document into positive or negative. Sentiment analysis approach identifies the semantic relationship between the sentiment expressions and subject properly and also improve the performance of discovering pattern of Human interactions in meetings. |
4 |
Mammogram Image Feature Extraction Using Pulse-Coupled Neural Network Abstract: A typical mammogram image processing system generally consists of mammogram image acquisition, pre-processing, segmentation, feature extraction, feature selection and classification. The texture description methods such as GLCM, GLDM, SRDM, and GLRLM are widely used to extract features in mammogram images for analysis and identification of micro calcification. The Pulse-Coupled Neural Networks (PCNN) is found a very good feature extraction model widely used in the area of image processing. The PCNN features are extracted from the mammogram images and analyze classification performance along with GLCM, GLDM, SRDM, and GLRLM features, extracted from the same mammogram images. These processes are executed and analyses the features. The performance of the proposed PCNN Feature extraction Method is examined and the experimental results are illustrated. |
5 |
A Novel Method for Person Identification Based on Iris Abstract: The biometric person authentication technique based on the pattern of the human iris is well suited to be applied to any access control system requiring a high level of security. In this paper, a iris-based biometric identification system that increases the accuracy and the performance of a typical human iris recognition system is proposed. This system detects, isolates, and extracts the iris region from the eye images. The phase responses, obtained from convolving the polar images with 1D log Gabor filter, are quantized to generate the binary iris templates which are compared using the similarity measures like Cosine similarity, Jaccard Coefficient and Pearson Correlation Coefficient. The proposed method takes images from CASIA Iris database V3. |
6 |
Analysis of Path selection policy of the Routing Protocols using Energy Efficient Metrics for Mobile Ad-Hoc Networks Abstract: As routing protocols in MANET are very essential, reducing power consumption is an important in ad hoc wireless networks. Our aim is to improve energy performance of DSR (Dynamic Source Routing) protocol in mobile ad hoc networks. This routing protocol looks for shortest paths which jointly improve packet latency and network life time. Our proposal for a new routing module based on energy metrics. We have tried to minimize the total power needed to transmit packets, maximize the life time of every single node. In this paper, we have performed the comparison analysis of an energy-efficient DSR and AODV protocols by testing energy aware metrics such as Minimum Total Transmission Power Routing, Minimum Battery Cost Routing and Minimum Drain Rate. |
7 |
Performance Analysis of Entropy based methods and clustering methods for Brain tumor segmentation Abstract: Brain tumor is the most deadly disease that affects human life span. To segment the brain tumor part, many segmentation techniques have been emerged in image processing like region based Segmentation, Boundary based segmentation. In this paper, several entropies based methods and several cluster techniques are compared and analyzed for brain tumor segmentation. Several entropies such as rough entropy, Shannon entropy, Renyi entropy, Min entropy, Log Energy, entropy and several clustering methods such as K-Means segmentation, Fuzzy C-Means segmentation and improved Fuzzy C-Means clustering based on measure of medium truth degree are applied for segmentation of brain tumor MRI image and the result is compared and analyzed. Entropy and clustering methods are applied to segment the different parts of the image based on threshold. The proposed segmentation gives higher accuracy when compared with other methods like Region based segmentation, pixel based segmentation. Image accuracy is calculated using Peak Signal noise ratio (PSNR), Mean square error (MSE) for each entropy method and for each clustering method and the results show that Rough entropy gives better results for segmentation. |
8 |
Mammogram Classification using Fuzzy Neural Network Abstract: Breast cancer is one of the major causes for the increased mortality among women especially in developed countries. It is second most common cancer in women. The World Health Organization’s International estimated that more than 1, 50,000 women worldwide die of breast cancer in year. In India, breast cancer accounts for 23% of all the female cancer death followed by cervical cancer which accounts to 17.5% in India. Early detection of cancer leads to significant improvements in conservation treatment. However, recent studies have shown that the sensitivity of these systems is significantly decreased as the density of the breast increased while the specificity of the systems remained relatively constant. Mammography is a medical imaging technique that combines, low-dose radiation and high-contrast, high resolution film for examination of the breast and screening for breast cancer. Another disadvantage is false-positive result. This research proposes a fuzzy neural network for classifying mammograms. Results of screening the mammograms are organized by classification and finally grouped into three categories i.e., Normal, malignant and Benign. Experimental results show that this method performs well with the classification accuracy reaching nearly 82% in comparison with the already existing algorithms. The fuzzy neural network provided high accuracy in the early diagnosis of Mammography, which can provide quantitative indicators for early clinical diagnosis and serve as a convenient diagnostic tool for physicians. |
9 |
Evaluation of LSB Based Image Steganography technique for various file formats Abstract: Steganography is derived from the Greek word steganos which literally means “Covered” and graphy means “Writing”, i.e. covered writing. Stegnography is the art and science of hiding messages in such a way that no one apart from sender and receiver identify the message. The paper describes the steganalysis technique for the detection of secret message in the image. The strong and weak point of this technique is mentioned briefly. Steganography function is used to hide a secret message in any media such as text, image, audio and video. There are many algorithms used for hiding the information. One of the simplest and best known techniques is Least Significant Bit (LSB). This paper focuses on image Steganography and hiding the message in the Least Significant Bit (LSB) method. We also discuss the LSB method used for various file formats. |
10 |
Fuzzy Clustering using Credibilistic Critical Values Abstract: In this paper, the utility of credibilistic critical values in crisp conversion of fuzzy data sets is considered. Conversion of this type becomes essential mainly when clustering of fuzzy data sets is carried out. In this paper performance of two popular clustering algorithms namely Fuzzy c–means and Fuzzy c–medoids algorithms are evaluated under credibilistic critical value crisp conversion is carried out. Two synthetic data sets of varying nature are used in the comparative study. Some popular fuzzy clustering validity measures were employed in this study. |