Search

Various types of three-dimensional (3D) cameras have been used to analyze real-world objects or environments effectively. However, because most existing 3D cameras capture scenes by statically using one type of camera, there may be a limit to the quality of the captured images. Therefore, in this paper, we build a hybrid camera system that combines passive triangulation (PT)- and active triangulation (AT)-based cameras and suggest a new mechanism of estimating accurate 3D depth by adaptively switching between the two types of cameras depending on the complexity of the environment. The suggested method initially uses initial input images to extract brightness and texture, which are major features representing the current state of the surrounding environment. The method subsequently generates a set of rules that dynamically select the PT- or AT-based camera, whichever can operate more suitably in the current environment, by analyzing the two extracted features. In experimental results, we demonstrate that the proposed adaptive camera-selection approach can be applied to extract 3D depth reliably with reasonable performance in terms of accuracy and time. [ABSTRACT FROM AUTHOR]

Software-Defined Networking (SDN) has emerged as a new architectural paradigm in computer networks, aiming to enhance network capabilities and address the limitations of conventional networks. Despite its many advantages, SDN has encountered numerous attack risks and vulnerabilities. Using an intrusion detection system (IDS) is one of the most important ways to address threats and concerns in the SDN. The great flexibility, adaptability, and programmability of SDN, together with other unique qualities, make the integration of IDS into the SDN network effective. The majority of these methods are less scalable and have poor accuracy. This research suggests an Optimized Fuzzy Based Function Network (OFBFN) to solve this problem. The Modified ResNet152 method is utilized to extract features from the input data. The Binary Waterwheel Plant Algorithm (BWWPA) selects the essential features. To characterize attacks within the InSDN, BOT-IOT, ToN-IoT, and CICIDS 2019 datasets, the system first selects the most efficient features. Then, it employs the FBFN with the Coatis Optimization Algorithm for classification. The suggested system classifies attacks and benign traffic, distinguishes between different types of attacks, and specifies high-performance sub-attacks. Four benchmark datasets were utilized for training and evaluating the proposed system, demonstrating its effectiveness. According to the findings from the experiments, the suggested approach performs better than others at identifying a wide range of threats. [ABSTRACT FROM AUTHOR]

In today's rapidly evolving digital landscape, the surge in smartphone usage is paralleled by an increasing wave of cyberthreats, highlighting the limitations of existing signature-based malware detection methods. To address this problem, our research introduces a Time-Aware Machine Learning (TAML) framework specifically designed for Android malware detection. Our framework extracts the best time-correlated features and then it builds time-aware and time-agnostic machine learning (ML) models. The ML models are trained on the KronoDroid dataset, which contains more than 41,000 benign Android apps and more than 36,000 malicious apps developed between 2008 to 2020. Our experimental evaluation revealed that the Last Modification Date 'LastModDate' feature is a critical variable for time-aware classification. Moreover, our empirical analysis reveals that real-device detection outperforms emulator-based detection. Impressively, the time-correlated features boosts the detection performance and achieving an outstanding 99.98% F1 score in a time-agnostic setting. In addition, on each year, our time-aware experiments outperformed the traditional ML detection models. Our time-aware classifier achieved a 91% F1 score on average and a maximum F1 score of 99% of yearly chunk experiments over 12 years. These experimental results affirm the effectiveness of our proposed method in detecting Android malware. [ABSTRACT FROM AUTHOR]

Developing technologies over digitalization have become more popular and become a threat to society and cybersecurity. Generally, the phishing method is used by hackers to access the data without the influence of users whose data was stolen. Several techniques are used to detect whether the data is phished or non-phished. Some anti-phishing software is used to identify the phishing data. However, few of these techniques did not provide efficient performance. Hence, the proposed model is introduced to overcome the issues obtained and improve the efficiency of detecting whether the data is phished or non-phished. The data is gathered from the Phishstorm dataset, which is pre-processed using the Z score normalization method and data cleaning. Data balancing is done by the Advanced synthetic sampling approach (Adv-SyN) to balance the dataset, and the features are extracted using a Double self-sparse autoencoder (DSelSa). The Opposition Gazelle optimization algorithm (OpGoA) model is used for optimal feature selection, and finally, the data is classified using Multi Head Depth wise Tern integrated long short term memory (MDepthNet). The sooty tern optimization is used to evaluate the loss function of the network model. The performance of the proposed model is analyzed based on some evaluation metrics and compared with other models, which describes the efficiency of the proposed model. The main objective of proposed technique used to detect the phishing attack and phishing or non-phishing. An automated DL methodology introduced for effective detection of phishing attacks for enhancing the cyber security. The accuracy of the proposed model is obtained as 99.45%, and Precision is 99.45%. RMSE and MSE rate of the proposed model is reduced to 0.73 and 0.05 for better performance. [ABSTRACT FROM AUTHOR]

Medical image registration is vital for precise healthcare diagnosis, treatment planning, and disease progression tracking, but traditional methods fail to capture complex spatial transformations and anatomical variations. A Kernel Principal Component Analysis (KPCA) driven Teaching Learning based optimization (TLBO) approach is proposed to overcome these limitations. The proposed approach is categorized into three phases, i.e., pre-processed phase, contour extraction phase with feature extraction using KPCA, and evaluating robust affine transformation parameters leveraging TLBO for accurate alignment. In the pre-processing phase, gaussian filter is applied to remove noise from source and target images, followed by normalization process. Afterwards, contour extraction is carried out to create a feature image that represents the boundaries of an image. Centroid localization is then utilized to compute translation parameters, which determine the spatial alignment between the images. By utilizing KPCA, this method captures non-linear relationships in the data to enhance the representation of image features. TLBO is employed to optimize the rigid transformation parameters to improve the accurate alignment of source and target images. Extensive experiments are carried out on monomodal and multimodal medical images such as CT and MRI taken from the Harvard Brain ATLAS, Kaggle as well as in-house clinical dataset to demonstrate the effectiveness of the proposed approach. The proposed approach significantly outperforms state-of-the-art methods, with improvements of 44.37% in Root Mean Square Error (RMSE), 19.93% in Structural Similarity Index Measure (SSIM), 20.01% in Peak Signal-to-Noise Ratio (PSNR), and 16.21% in Cross Correlation (CC) quality assessment metrics. [ABSTRACT FROM AUTHOR]

Accurate resource utilization estimation is crucial for efficient resource allocation, capacity planning, and cost optimization in cloud systems. In the past, several artificial intelligence, machine learning, and deep learning-based techniques have been evolved to forecast cloud cluster workload. Despite the abundance of available techniques, most existing techniques fail to achieve the desired prediction efficiency and generalization capability. They are computationally inefficient in accurately determining resource utilization at a machine-level granularity. The current research proposes a computationally less-expensive hybrid approach combining cluster analysis and deep neural learning with transfer learning to estimate the machine-level workload. The method implements clustering to identify the similarity patterns among the non-linear usage profiles of machines present in the input dataset. Subsequently, the generalized deep neural learning models are developed considering only a sample dataset belonging to each identified cluster. Lastly, the concept of transfer learning is deployed using pre-trained generalized models to estimate the workload for all remaining machines relating to the clusters. The performance validation of the proposed approach is carried out on the real-world traces dataset of the google cluster. The comparative evaluation of the proposed approach with benchmark approaches verifies the achieved performance benefits and accuracy. [ABSTRACT FROM AUTHOR]

In the backdrop of an ever-expanding urban transportation road network, the dramatic changes in traffic flow make traffic flow forecasting become a challenge. Which encompass intricate spatial correlations and intricate non-linear temporal relationships. Given a complex time series of traffic flow, accurately predicting traffic flow under such circumstances becomes an arduous task, particularly for long-term sequences. To tackle this conundrum, we put together a traffic prediction framework called the multi-scale attention-based graph convolutional recurrent network (MAGRN). Based on GCN and RNN, this framework skilfully extracts dynamic spatial and temporal features from node attributes. By incorporating multi-scale feature extraction and two attention mechanisms within the attention structure, MAGRN adeptly captures features at varying granularities. This enables MAGRN to diligently capture both the temporal interdependencies and spatial correlations inherent in traffic sequences, thereby minimizing errors. Experimental validation on three real-world traffic datasets unequivocally attests to the superior performance of the MAGRN model. [ABSTRACT FROM AUTHOR]

Even though video technologies have continuously advanced, playing a video still requires full length of the videos playtime, high volume of data traffic, and appropriate sound play devices. To overcome these limitations, this paper proposes a method to serve a video in an image carousel format. By providing multiple key frame images in carousel form with full script of the video, the proposed image carousel system transforms any playable video content into a readable form so that users can read the content rather than watch. Reading a video make consumers be freed from sound restrictions, and increase the potential and efficiency in the video linkage and usage in other services. In addition, this approach saves both playtime and data traffic. When a video encoded in H.264 codec with 720p resolution is transformed into the proposed image carousel form, in average 90-95 % of the data traffic were saved. [ABSTRACT FROM AUTHOR]

In the present scenario, the product recommendation system is important to boost sales on many e-commerce websites. The majority of the existing product recommendation systems completely rely on the customer's website browsing history or historical transaction records to precisely predict the user's preferences for online product recommendations that are constrained by the limited information available on the e-commerce websites. Usually, processing the customer's feedback is a difficult procedure, where it is tough in interpreting and analyzing the text information. Therefore, an effective deep learning-based recommendation framework is implemented in this research manuscript. After acquiring text information from the amazon product reviews database, the data pre-processing is carried-out by using stemming, lemmatization, and stop word removal techniques. The undertaken pre-processing techniques remove both inconsistent and duplicate data that eases the process of data interpretation and use. Further, the computational time of the model is decreased by removing inconsistent and duplicate data from the original dataset. Then, the feature values are extracted from the pre-processed text data using the term frequency-inverse document frequency (TF-IDF) technique. The extracted feature values are fed to the enhanced convolutional-neural-networks (CNNs) for sentiment analysis. The enhanced CNN model uses Glove as a pre-trained model with a skip-gram in the word-embedding layer for identifying the relevant product ratings based on the user's experience to further enhance product sentiment analysis. The embedding layer in the enhanced CNN-model effectively converts every word into fixed feature vectors that helps in better word representation along with reduced dimensions. The enhanced CNN model obtained a mean recall of 94.80%, the precision of 93.64%, and accuracy of 96.92% on the amazon product reviews database and it is superior to the traditional CNN. [ABSTRACT FROM AUTHOR]

Emerging scientific simulations on leadership class systems are generating huge amounts of data and processing this data in an efficient and timely manner is critical for generating insights from the simulations. However, the increasing gap between computation and disk I/O speeds makes traditional data analytics pipelines based on post-processing cost prohibitive and often infeasible. In this paper, we investigate an alternate approach that aims to bring the analytics closer to the data using in-situ execution of data analysis operations. Specifically, we present the design, implementation and evaluation of a framework that can support in-situ feature-based objects tracking on distributed scientific datasets. Central to this framework is a scalable decentralized and online clustering, a cluster tracking algorithm, which executes in-situ (on different cores) in parallel with the simulation processes, and retrieves data from the simulations directly via on-chip shared memory. The results from our experimental evaluation demonstrate that the in-situ approach significantly reduces the cost of data movement, that the presented framework can support scalable feature-based objects tracking, and that it can be effectively used for in-situ analytics in large scale simulations. [ABSTRACT FROM AUTHOR]

In the last several decades, application of affine transformation and optimization approaches have received enough attention in the domain of monomodal/multimodal image registration. A novel and robust image registration method is proposed which leverages teaching learning-based optimization (TLBO) to obtain the optimal value of rigid transformation parameters. Further, speeded up robust features (SURF) framework is employed to extract features along with Random sample consensus (RANSAC) algorithm. Afterwards, projective transformation is used to obtain more accurate registered image. To remove noise, gaussian filter is used during pre-processing and then normalization is carried out. TLBO is used to identify the optimal geometric transformation parameters by considering the mutual information (MI) as an objective function. This method detects keypoints (features) using SURF then K-nearest neighbour (KNN) is employed to match the detected features. Furthermore, RANSAC eliminates false matches. The registered image with optimal value of rigid transformation parameters is obtained and then enhanced by using SURF-RANSAC followed by projective transformation. Experiments are conducted on Whole Brain ATLAS and KAGGLE datasets. Robustness of proposed approach is verified by the improvement in evaluation metric, such as structure similarity index measure (SSIM) by 8%, MI by 12.71% on monomodal and root mean square error (RMSE) by 41.44% on multimodal images as compared to recent state-of-the-art methods. [ABSTRACT FROM AUTHOR]

To detect the region and bleeding frame in the wireless capsule endoscopy video, an automatic computer-aided technique is highly demanded to reduce the burden of physicians. The wireless capsule endoscopy (WCE), is an imaging technology which is recently established and doesn't require any wired device. This device detects abnormalities in GI tract, i.e. (colon, esophagus, small intestine, and stomach). A WCE video consists of 57,000 images. It is very hard to examine by clinicians. To determine bleeding photos out of fifty-seven thousand WCE images makes the task very hard and expensive. The main goal is to develop an automatic obscure bleeding detection method by using superpixel segmentation and naive Bayes classifier. Naive Bayes and superpixel segmentation are used for this problem. [ABSTRACT FROM AUTHOR]

For existed problems on fire detection fields, the traditional recognition methods on fire usually based on sensor's signals are easily affected by the external environment elements. Meanwhile, most of the current methods based on feature extraction of fire image are less discriminative to different scene and fire type, and have lower recognition precision if the fire scene and type change. To overcome the drawback on fire recognition, the new fast recognition method for fire image has proposed by introducing color space information into Scale Invariant Feature Transform (SIFT) algorithm. Firstly, the feature descriptors of fire are extracted by SIFT algorithm from the fire images which are obtained from internet databases. Secondly, the local noisy feature points are filtered by introducing the feature information of fire color space. Thirdly, the feature descriptors are transformed into feature vectors, and then Incremental Vector Support Vector Machine classifier is utilized to establish the fast fire recognition model. The experiments are conducted on real-life fire image from internet. The experimental results had shown that for different fire scenes and types, the proposed algorithm has outperformed Kim's method, Dimitropoulos's method and Sumei's method in terms of recognition accuracy and algorithm's running speed. The proposed algorithm has better application prospects than Kim's method, Dimitropoulos's method and Sumei's method. [ABSTRACT FROM AUTHOR]

In the image handling and image processing areas, most operations can be executed in a pixel-by-pixel or cluster-by-cluster manner. These parallel and simultaneous executions have many benefits, and many researchers showed remarkable improvements. In this paper, we started from a specific and practical image handling and feature extraction sequences. We focused on the detailed design and robust implementation on the modern massively parallel architecture of CUDA. We present the enhanced features of our implementation and their design details. Our final result shows 13 times faster execution speed, in comparison with its previous CPU-based implementation. These methods can be applied to the variety of image manipulation processes. [ABSTRACT FROM AUTHOR]

Telecom customer churn data is not publicly available because involving users' personal privacy. In 2009, the French telecommunications company Orange for knowledge discovery and data mining (KDD) competition provides a telecom customer churn data set KDD Cup 09. In order to solve the high dimensional problem of KDD Cup 09, a new feature reduction method is used to explore the influence of different features on the prediction of classification model. In this paper, a new K- local maximum margin feature extraction algorithm (KLMM) is proposed. Through researching on the diversification subspace partition rules, the corresponding potential field structure is constructed. According to the data source in the dimension of scalability, the intrinsic link between data attributes and classification results is revealed. The extracted features can reduce the dimension of the churn prediction in telecom data. The KLMM method adapts auto selection sigma factor to reflect the anisotropy of features. The potential function is used to assess the weights of attributes and find the potential important weight. Experiments and analysis show that the extracted features by KLMM are more likely to find a classification hyperplane which can separate data points of the different classes. [ABSTRACT FROM AUTHOR]

In the modern healthcare system, the function of the Internet of Things (IoT) and the data mining methods with cloud computing plays an essential role in controlling a large number of big data for predicting and diagnosing various categories of diseases. However, when the patients suffer from more than one disease, the physician may not identify it properly. Therefore, in this research, the predictive method using the cloud with IoT-based database is proposed for forecasting the diseases that utilized the biosensors to estimate the constraints of patients. In addition, a novel Generalized Fuzzy Intelligence-based Ant Lion Optimization (GFIbALO) classifier along with a regression rule is proposed for predicting the diseases accurately. Initially, the dataset is filtered and feature extracted using the regression rule that data is processed on the proposed GFIbALO approach for classifying diseases. Moreover, suppose the patient has been affected by any diseases, in that case, the warning signal will be alerted to the patients via text or any other way, and the patients can get advice from doctors or any other medical support. The implementation of the proposed GFIbALO classifier is done with the use of the MATLAB tool. Subsequently, the results from the presented model are compared with state of the art techniques, and it shows that the presented method is more beneficial in diagnosis and disease forecast. [ABSTRACT FROM AUTHOR]

Due to the rapid development of diversified technology, people may use multiple electronic devices, such as personal computers, tablets, and smartphones, to connect to the Internet in their daily lives. Switching between devices enables a user to use e-commerce on various platforms. The complexity of consumer behavior is directly proportional to the number of involved devices. Additionally, since the personal privacy regulations nowadays are getting more strict, the user data on the Internet starts to be anonymous. Thus, determining how the devices are related is an indispensable step in achieving precision marketing or developing customized applications. In this research, the dataset provided by the CIKM Cup 2016 Challenge is used. The representation of a device is created by extracting features from browsing logs. The computation cost is reduced by filtering candidates of a target device instead of comparing them in pairs. Latent semantic indexing representations and techniques of supervised learning are used to accomplish filtering. Performing word embedding can turn literature semantic into vectors through an unsupervised neural ensemble. The addition of feature engineering on the input vectors of supervised classification can enhance the classifier's discrimination. The classification is used to determine the probability of any two instances belonging to the same user. The significant benefit of the implementation is to form the sequences mentioned above by a cross-device linking mechanism to provide a baseline for aligning with the computation limitation and boosting the performance. [ABSTRACT FROM AUTHOR]

Biometric Recognition and authentication is used in every application for the secured identification of the persons. Several Researches has been carried out in the past decade that concerns only to enhance the accuracy of biometric algorithms for various kinds of features by ignoring the aspects of robustness and reliability. With this objective, new algorithm called effective linear scale authentication has been proposed which works on the principle of hybrid fusion of two feature inputs such as Hand geometry and iris of the users. The two different techniques have been adopted for the Feature Extraction. One is effective linear binary patterns and other is scale invariant fourier transform and it is stored in the databases for the further verification. This algorithm has been tested with the CASIA Image Datasets and the results had clearly proven that the novel idea of fusing the local features to address the scalar and angular inefficiencies of the existing methods. [ABSTRACT FROM AUTHOR]

Cloud based Personal health record (PHR) is an emerging cloud based platform for exchanging a person's health information in a secure manner. There occur many security issues when records of the data owners are outsourced through the third party cloud providers. The health records which are sensitive must be stored and retrieved through a secure source without any loss in the data. Nevertheless sharing and searching of the data is the key aspect, but when it is outsourced undeniably it is a cumbersome task. It may lead to unveil the sensitive information and so the records may become vulnerable to the hackers. In this report, we have suggested a novel access control structure called as user usage based encryption constructed on the searchable attribute based encryption to guarantee the data protection. Usage is mapped as credential with a time frame to every private attribute. The data user can decipher a fortified attribute only if there is a match between the credentials associated with the attribute. Using the feature extraction algorithm the searchable encryption scheme enables a consistent routing of encrypted attributes. Multi-Credential routing is applied to strengthen the confidentiality of the fragile records. We allow the data user to perpetuate the credentials according to their usage criteria also the user receives the keys as labels along with the credentials. The data owner will be able to associate each enciphered attribute with a set of credentials. Before beginning the encryption scheme we apply the singular value decomposition algorithm to the unutilized or less used attributes to reduce the attribute set. Additionally the data user confidentiality pitfalls are tackled using the semantic clustering of data user. To preserve the data user confidentiality reliable overlay privacy preserving protocol is designed. We manifest a complete security analysis so that our recommended system dominates the up to date approaches in terms of communication and ciphering cost. [ABSTRACT FROM AUTHOR]

Distinctive human cancer could be a standout amongst the foremost dangerous disorder that is essentially caused by hereditary insecurity of various sub-atomic changes. Among many sorts of human growth, neoplasm is that the most generally recognized one. To acknowledge skin growth at a starting amount we'll examine and break down them through totally different ways named as division and highlight extraction. Here, we have a tendency to center dangerous malignant melanoma skin growth, (because of the high centralization of Melanoma-Heir we provide our skin, within the stratum layer of the skin) recognition. In this, we have a tendency to used our ABCD manage dermoscopy innovation for harmful malignant melanoma neoplasm recognition. During this framework distinctive advance for malignant melanoma skin injury portrayal i.e., initial the Image Acquisition Technique, pre-preparing, division, characterize highlight for skin Feature choice decides sore portrayal, characterization ways. within the Feature extraction by advanced image getting ready strategy incorporates, symmetry recognition, border detection, shading, and dimension discovery and moreover we have a tendency to used LBP for separate the surface based mostly highlights and In order to obtain pattern wise features. Here we have a tendency to plan the rear propagation neural network to rearrange the type or harmful stage. Because neural network can give multi classification results and can back propagate to hidden layers. [ABSTRACT FROM AUTHOR]

In traditional appearance based face recognition approaches, the facial images in standard face databases are used as it is for testing standard statistical face recognition algorithms. Due to the camera angle, position and lightening conditions, the facial image captured contains occlusion. Human pose, facial expressions and alignment of face with respect to camera axis also result in occluded image. Sometimes the background of the face also covered during the facial image capturing. We proposed an approach, called wavelet based facial image decomposition (WBFD) and in this approach the facial images are decomposed using two levels wavelet transform. In the faces obtained after two level wavelet decomposition, the occlusion and background images are suppressed. The face images from Yale database are used for our experiments. In order to test the performance of proposed WBFD approach, the face images obtained from WBFD approach, resized Yale face and Yale face from the original Yale face database are tested with five standard statistical appearance based face recognition algorithms such as eigen face, Fischer discriminant analysis, kernel principal component analysis, independent component analysis and 2D principal component analysis. The experimental results obtained show that, the proposed WBFD approach gives high recognition accuracy and require minimum recognition time in seconds for all the five face recognition methods. [ABSTRACT FROM AUTHOR]

Biometric recognition identifies an individual with the help of biological traits of humans like fingerprint, iris, voice etc. Estimation of similarity between the biometric images stored in cloud databases needs a closer attention due to its essentiality. To improve the privacy concerns, the proposed system encrypts the biometric data via feature extracted data hashing with the private key. The minutiae feature extraction procedure adequately removes edge detected features from the sensed data. The keys generated are stored in the cloud database. This work focuses on the comparison of the various distance similarity measure functions used for matching the stored template key with the query key. It has been found that by comparing with many existing similarity distance measures Minkowski distance method performs better in terms of accuracy, recall and F-score. Since the work involves fingerprint as the biometric trait, the performance is measured using false matching rate, false non-matching rate, false positive identification rate, false negative identification rate, accuracy, recall, F-score. The implementation has been done with MATLAB. [ABSTRACT FROM AUTHOR]

Nowadays, speech recognition systems are used in various environments, namely, healthcare, robotics, vehicle control and unmanned aerial vehicle system. In recent years, many speech recognition systems have been developed to solve various issues in real world applications. We have proposed a novel speech recognition system using enhanced mel frequency cepstral coefficient with windowing and framing method. Windowing and framing method is used to remove the Gaussian white noise present in the input speech signal. The de-noising block effectively uses the nonnegative matrix factorization algorithm for factorizing the Mel-magnitude spectra of noisy input audio signal. Moreover, the mel-frequency cepstral coefficients (MFCC) is used for finding the more important features exist in the speech signal. Finally, Laplace smoothing technique is used as the language model for recognizing the audio signals. MATLAB software is used for demonstrating the proposed Mel frequency cepstral coefficient with Windowing and Framing based speech recognition system. We have compared the proposed speech recognition system with wavelet based feature extraction and artificial neural network based feature extraction methods for speech recognition. The experimental results proved the good performance of the proposed Mel frequency cepstral coefficient with windowing and framing based speech recognition system. [ABSTRACT FROM AUTHOR]

In the proposed work introduced a pose-invariant face recognition process for the videos so as todiminish the calculation intricacy of the traditional technique. The intended approach includes four phases, namely (i) preprocessing, (ii) object detection, (iii) feature extraction along with (iv) dimensionality reduction. The initial phase is segmenting the database video clips into the frames where the preprocessing is done using an adaptive median filtering such that to eradicate the noise. The subsequent phase is detecting the preprocessed image through the viola–jones technique. With this technique, the mouth, left eye, face, nose, right eye were identified. After that, the training image attributes like the texture, edge, and wavelet are extorted with the assistance of dissimilar feature extortion approaches. The extorted attributes will then be inputted to the PCA approach. The similar process is replicated for query images. At last, perform the similarity measurement amid the trained database images and query images such that to attain the distinguished image. In addition, the genuine general performance assesses of the projected technique is assessed with that of the traditional ICA and also LDA techniques. [ABSTRACT FROM AUTHOR]

Texture medical image classification is a major task in many areas of computer vision and pattern recognition, including biomedical imaging. In reality, optical surfaces can be abused to recognize particular tissues or cells in a true example, to feature complex chemical reactions between molecules, and additionally to identify subcellular designs that can be confirmation of specific pathologies. It makes automated texture analysis fundamental in many applications of biomedicine, such as the accurate detection and grading of multiple types of cancer, the differential diagnosis of autoimmune diseases, or the study of physiological processes. Due to their specific characteristics and challenges, the design of texture analysis systems for biological images has attracted ever-growing attention in the last few years. In this work, approach applies the multi-scale gabor rotation-invariant local binary pattern (MGRLBP) used to analysis the texture features of a bio medicinal image and joined the weight aspect that is presented by the direct measure to obtain the last texture feature of a biomedical picture. By using multi-scale binary pattern (MSBP) classifier with the direct action and multi-scale gabor rotation-invariant local binary pattern algorithm. Both quantitative and qualitative methods are applied to assess the classification results. The simulation work does with the MATLAB2013a environment by using proposed MGRLBP and MSBP technique. The simulation results demonstrate the usefulness of the suggested technique and its ability to classify the texture medical images. Hence the proposed model produces better features for texture in medical images. Over 93% efficiency achieved by using MGRLBP and MSBP method. [ABSTRACT FROM AUTHOR]

Network communication load state recognition has important practical application value, in order to improve the network communication load state recognition accuracy is proposed based on relevance vector machine load of the network traffic state identification model. When the network throughput overload, need to adjust the network communication strategy to ensure the overall performance of sensor network, so first of all need to network transmission state identification. Firstly, uses wavelet analysis in order to eliminate the noise signal, and extract the features, and then the artificial bee colony optimization algorithm to determine the optimal parameters of relevance vector machine, establishing the network communication load state identification model, finally through with other network communication load condition recognition models to be compared with the experimental. The results show that the model can describe the relationship between the state and the characteristics of network communication, and improve the recognition accuracy of the network traffic load state, and the recognition results are higher than the comparison model. [ABSTRACT FROM AUTHOR]

The rapid expansion of e-commerce improves customer experience and online businesses largely. Customers buy products based on reviews provided by the users of the product through cloud. Opinion of customers plays a huge role in decision making for business. Opinion mining helps in comparison of products by identifying the main features commented upon by customers and also ranking the products. Most of the opinion mining systems identify only the explicitly mentioned features from the reviews. Identification of implicit features and considering the context of the reviewer about a product plays a huge role in better decision making. The proposed system uses a novel approach for identifying explicit features, implicit features and the context of the reviewer through soft clustering techniques for forming the clusters. Since opinion comparison and ranking of products are based on multi-alternative criteria, a multi criteria decision making approach is used for ranking the products. An experiment is designed to test the proposed method using the reviews collected from Amazon cloud set for mobile phone and camera reviews. The result shows that the proposed method identifies the key features of the products and provides the better ranking of products as it considers the explicit features, implicit features and the context of the reviewer. [ABSTRACT FROM AUTHOR]

One comparative analysis method of English comparative literature based on big data Bayesian method has been proposed to improve the calculation accuracy of comparative analysis process of English comparative literature. Firstly, one multilayer Bayesian classification recognition algorithm mechanism has been proposed for the problem of low accuracy and poor calculation efficiency of Bayesian classification algorithm under the big data environment; Gabor multilayer feature extraction algorithm has been proposed for multilayer Bayesian algorithm, Gabor multilayer feature extraction Bayesian algorithm has been designed and realized. Secondly, the algorithm process of comparative analysis system of English comparative literature based on this improved algorithm has been designed and the success rate of comparative analysis of English comparative literature has been improved effectively based on the features of comparative analysis of English comparative literature. Finally, the effectiveness of algorithm has been verified through experimental simulation. [ABSTRACT FROM AUTHOR]

Emotion has a most important aspect in terms of interactions among the humans and this would become ideal for human emotions to get mechanically identified by the machines and primarily for enhancing the communication among the human–machine. In the recent work Enhanced Bat algorithm with Simulated Annealing (EBSA) are introduced for solving emotion recognition problem. Here the removal of noises from the speech samples and reduction in the number of speech features becomes very difficult task which reduces the accuracy of the classifier. To solve this problem this research work involves detection of emotions from speech which stimulates machines understanding human behavioral tasks namely reasoning, decision making and interaction. EBSA is used in the previous system to identify the happy, sad and neutral emotions from speech input. The performance of the previous system has been decreased due to recognition accuracy and feature selection. Improved Artificial Bee Colony (IABC) with Neuro Optimized Emotional Classifier (NOEC) solves this issue in the proposed system. The Iterative Conditional Entropy Kalman filtering (ICEKF) is initially processed to effectively filter the noisy features from the inputted speech data. Mel Frequency Cepstrum Coefficient (MFCC), pitch, energy, intensity and formants are extracted as speech features. Every extracted feature is maintained in the database and annotated along with their emotional class label. IABC algorithm chooses the feature optimally, which in turn employs the best fitness function values. From the optimally selected dataset, the NOEC is processed. Emotions can be identified from the Tamil news speech dataset with the help of the supervised machine learning technique, which demands the training set (collection of emotional speech recordings). Every recording or sample in the dataset is named with the emotional class and they are indicated as n-dimensional vector of spectrum coefficients which in turn is extracted from the Tamil news speech dataset. This dataset is collected from real time via using the search engine sites like Google, YouTube, twitter etc. By implementing IABC with NOEC classification process, the work segregates the emotional classes such as happy, sad, anger, fear and neutral emotions perfectly. From the experimental verification, it is confirmed that the proposed method IABC with NOEC gives better performances with respect to accuracy, precision, recall and f-measure values. [ABSTRACT FROM AUTHOR]

The traditional human tracking systems are often prone to problems which can be deduced either from the subject, camera or background movements which includes major changes in posture, appearance, clothing and lighting of the background. The work introduced here proposes a system for extracting and tracking objects from a video sequence by initializing the process of feature extraction of the selected area of the human. The need for cascade extraction by using Haar-like features, is to basically decrease the use of crude or raw pixel values and then make classification easier. The major issue here is the problem in extracting the required finite set of input for identifying the characteristics necessary for domain encoding rule In the proposed work the major aim is to address the tracking procedure by facilitating the exclusive histogram of oriented gradient (HOG) methodology, using which only the human's unit element is embraced in the usual tracking system, making it problematic for robust monitoring. The features of Haar-like and HOG methods are combined to propose a tracking system which uses the Haar characteristics for the object's structure and the HOG features for the edge. A set of mixed features is developed with these two features. Boosting Online's selection feature is used to select important features and update these online features to understand the optimal choice using cascading SVM classifier. [ABSTRACT FROM AUTHOR]

Biometric recognition is an effective method for discovering a person's identity. Multimodal biometric recognition employs multiple sources of information about a human for authentication. Recently, many research works are designed for multimodal biometric recognition using classification techniques. However, the performance of conventional techniques was not efficient for achieving higher recognition rate. In order to overcome such limitations, an ensembled support vector machine based kernel mapping (ESVM-KM) technique is proposed for multimodal biometric recognition. The ESVM-KM technique is designed for improving the accuracy of multimodal biometric recognition with human face, finger print and iris images. The ESVM-KM technique initially performs the preprocessing in order to remove noise and to improve the image quality for human recognition. After that, ESVM-KM technique carried outs Gabor wavelet transformation based feature extraction process in which features of human face, finger print and iris images are efficiently extorted for classification. Finally, the ESVM-KM technique used ensembled SVM classifier for enhancing the recognition rate of multimodal biometric system. The ESVM-KM technique conducts simulation work on the metrics such as computational time, recognition rate, and true positive rate. The simulation results demonstrate that the ESVM-KM technique is able to improve the recognition rate and also reduces computational time of multimodal biometric recognition system when compared to state-of-the-art works. The results got through ESVM-KM are stored in cloud environment for easy and future access. [ABSTRACT FROM AUTHOR]

Educational data mining (EDM) has emerged as a research area in recent years for researchers all over the world from different and related research areas. The EDM obtained knowledge can be used to offer suggestions to the academic planners in higher education institutes to enhance their decision-making process. Literature has suggested various prediction models for predicting the student's performance. This work proposes the cluster based distributed architecture for predicting the student's performance. The proposed cluster-based distributed architecture performs the prediction with clustering through Bayesian fuzzy clustering, feature extraction through Kernel-based principal component analysis, and prediction through the proposed Lion–Wolf based deep belief network (LW-DBN). The proposed architecture uses the LW algorithm to find the optimal weights for the DBN. The experimentation of the proposed work is done by collecting a real-time database and measuring the prediction performance through the mean square error (MSE) and root MSE (RMSE). The proposed LW-DBN model has achieved lower error performance than other models with MSE and RMSE values of 0.222606 and 0.050435, for the database. [ABSTRACT FROM AUTHOR]

In the field of pattern recognition and data learning process the dimensionality reduction is a necessary method for providing a personalized data in perfect numbers. The increase in research proclaims the importance of data reduction in current trends. The old traditional reduction techniques use ranking systems. But this differs from others in reduction of data without lagging its performance in terms of increased efficiency and less percentage of error occurrences. The area of dimensionality reduction not only ends in pattern recognition and also in many high dimensional data processing elements such as text categorization, indexing of documents and mainly in gene expression data. The feature extraction and feature selection are the two steps of process in reduction. The process proposed is statistical pattern recognition process and a paradigm for this approach to the problem is summarized herein. The four data process includes is (1) evaluation (2) acquisition (3) feature selection, and (4) statistical model of feature selection. We show how to reduce the dimensionality by utilizing group and convexity model. This paper concluded by an integrated approach to the implementation dimensionality reduction process with an effort to maximize the efficiency and accuracy. [ABSTRACT FROM AUTHOR]

In processing voice with environment noise, the noise must be eliminated to improve the vocabulary recognition rate. In this process, noise elimination and feature extraction for model-estimate technologies are utilized. Concerning these noise-elimination and model-estimate technologies, the most important part is to estimate mixed noise in the source signal and eliminate it. In a vocabulary recognition system, if unexpected noise appears in the signal, or if quantization noise is basically added to digital signals, the source signal is changed or damaged, which decreases the recognition rate. If a source signal is transformed or changed by being mixed with diverse kinds of noise, the hidden Markov model (HMM) is used for effective noise elimination. The HMM forms a model by extracting features to flexibly respond to diverse vocabulary changes found in voice and text, etc. The method is applicable to data changing over time, and can establish a more effective model as the number of parameters constituting the model grows larger. It can provide a robust model estimate by using a parameter set for structured models. HMM-based vocabulary recognition shows discriminating distribution of recognition probability regarding recognition vocabulary models, and has lower computational complexity for recognition. But it produces a relatively lower recognition rate. To solve that problem, a vocabulary recognition-model optimization method is proposed based on a similar phoneme-recognition process and efficient feature extraction. In vocabulary recognition, a similar phoneme-recognition process is applied to HMM to recognize models adjacent to the model group. Efficient feature extraction is used to optimize the recognition model to enhance the recognition rate. For vocabulary composition, a Gaussian-mixture feature-extraction model is optimized and used as a vocabulary recognition model. Then, it is processed with similar-phoneme recognition regarding the vocabulary recognition model. [ABSTRACT FROM AUTHOR]

The Insider threat is minimally addressed by current information security practices, yet the insider poses the most serious threat to organization through various malicious activities. Forensic investigation is a technique used to prove the presence of malicious insider with digital evidence. The proposed surveillance mechanism for countering insider threats operates in two phases. In phase one, the network has to be monitored for incoming and outgoing packets. The information is transferred using packets, and these packets are monitored and captured and the important features are extracted. By performing investigation on the captured packets, information related to suspicious activities can be obtained. In phase two, we mine various log files which are considered to posses vital traces of information when insider attack has been performed. The analysis of the log files is performed in order to extract the key pattern from files. The extracted patterns from log files are further processed. The suspicious data patterns are grouped into clusters to trace the anomaly. They are classified as legal and anomaly pattern with the help of KNN classifier .If anomaly is traced, the user's past activities are referred and a cross check is made with the features of captured packets the computational intelligence based on Dempster-Shafer theory is applied to prove with digital evidence, the presence of malicious insider in the critical networks with utmost accuracy. [ABSTRACT FROM AUTHOR]