Search

Principal Component Analysis (PCA) is a commonly used unsupervised exploratory analysis technique. It is also frequently used in dimensionality reduction. This preliminary paper investigates the feasibility of three variants of PCA, i.e. independent PCA (iPCA), sparse PCA (sPCA), and sparse independent PCA (siPCA) on forensic classification of paper based on their IR spectral data. After that, Linear Discriminant Analysis (LDA) models were built using the Principal Components (PCs) produced by the PCA and the three aforementioned variants. The performances of all these four LDA models, i.e. PCA-DA, iPCA-DA, sPCA-DA and siPCA-DA, were evaluated via leave-one-out cross-validation on the data set. The results obtained show that iPCA-DA and siPCA-DA are the most effective models with 100.0% classification accuracy. Then, the effectiveness of siPCA and iPCA models was evaluated based on posterior probability used for predictions of class membership that were derived from leave-one-out cross-validation. As a conclusion, siPCA is identified as the best classification model.

EMG is a medical technique that measures muscle response or electrical activity in response to a nerve’s stimulation of the muscle [10]. In this paper we have classified multiple hand gestures using small number of electrodes. This report aims to review current use of EMG based control system and explain the existing challenges. The stages involved includes data acquisition, feature extraction, classification and controller. Some of the application includes Arm rehabilitation, game controller, wheel chair controllers, etc.

Biometric authentication is a recognition of individual according to their unique physiological and behavioural characteristics. Recently, the application of biometric is the most trending in cyber security technology such as fingerprint, facial recognition, and voice recognition. However, these biometrics have their own drawbacks which allow the unauthorized party to cybercrime and the number of cases is also increased. To encounter this kind of problem, the previous researchers proposed brain signal or electroencephalogram (EEG) as biometric trait. EEG is an electrical activity recorded via non-invasive method using electrode placed on the scalps and measured as voltages. EEG is chosen by the researchers as biometric module because EEG hold its own unique characteristics and more robust against the cybercriminals. This paper presents a review of the EEG-based biometric studies and research. The previous research was reviewed based on their signal acquisition, pre-processing, feature extraction and classification. The general knowledge of EEG and the basic operation of biometric authentication also discussed in this paper. The recent studied and research is chosen with various proposed method respect to the better performance rate. In addition, the deep learning in biometric authentication is found to be the popular among the researchers for classification step because more robust and automatically extracted feature within the network.

The main goal of this paper to evaluate the performance of the proposed long short-term memory (LSTM)- basis relative humidity (RH) prediction model. Computational physics learning theory is a subfield in artificial intelligence. In this paper, we use the real observed weather data of RH from synoptic station location as an important aspect and effect in weather and climate and using machine learning approach basis on the LSTM neural network, which is a type of recurrent neural network (RNN). We use a machine learning approach to compute the relative humidity basis on the LSTM model. The LSTM is trained by using two years of climate data (from January 2008 to December 2009). LSTM is capable of forecasting complex univariate relative humidity time series with strong no-stationarity. The result shows the usefulness of LSTM and out performs traditional forecasting methods in the challenging relative humidity record problem.

The primary objective of this paper is to verify individuals as indicated by their finger surfaces. We propose to remove Finger Texture (FT) highlights of the two finger pictures (center, ring) from a low goal contactless hand picture utilizing LBP strategy. The utilization of Inner-Knuckle-Print (IKP) in biometric recognition is the most broadly proposed validation work. The unique characteristics of the IKP give us the requirement for recognizable proof. During the IKP filtering process, the image created by the scanner might be partially unique. This paper proposes artificial neural networks for effectively coordinating procedures to IKP validation. By utilizing the Back-Propagation method, the algorithm coordinates IKP and relates them to a novel accomplished client. After grouping, the procedure restores to the best counterpart for the given finger impression variables.

In recent years many of the people are suffering from diabetics which may result damaging the human eye sights by damaging the blood vessels of the retinal by form exudates around the optic disc. In this paper, we intend to build retinal exudates from fundus image to predict the severity of diabetics resulting in blood vessels tissue damages. In order to analysis the blood vessels damage and diabetic severity initially we use appropriate image pre-processing techniques to remove any noise from the retinal funds image, to remove the noise in this paper we use wavelet transform and first order Gaussian derivative and matched filter to segment the fundus image by rotating the original image by certain angle. The filtered data is stored in the matched filter bank, then by applying k-NN clustering technique to identify minimum value in each filter bank and marking such minimum value center of k-nearest neighbor value. Further, Support vector machine a supervised learning algorithm is applied to the identified k-nearest neighbor values thereby predicting the severity of blood vessel tissue damage from the fundus image.

Newly developed 3D reconstruction technology has many uses ranging from facial recognition, self-driving vehicles, and producing accurate models of samples. To utilize 3D reconstruction, a depth map of the sample must be generated from a series of images. This can be done using stereoscopic 3D, or in the case of a fixed camera (such as the 10ID On-Axis Microscope at NSLS-II used for this paper) through focus stacking. This paper will explain how focus stacking can be used to generate in-focus images at very shallow depth of fields, and merge these images to create depth maps and 3D models.

Hypertensive Retinopathy (HTR) is a disease caused by high blood pressure flowing into the retinal blood vessels, resulting in thickening of blood vessel walls and reducing blood flow in the retina. Complications arising from this disease are diverse and dangerous, ranging from retinal vein occlusion, eye nerve damage, even blindness. This paper proposed a system for hypertensive retinopathy detection by using Principal Component Analysis (PCA) dan Backpropagation Neural Network (BNN). Retinal image was taken from STARE database which separated into learning and testing data with a ratio 7 to 3. PCA as fundus-image-dimension-reduction method has successfully reduced fundus image raw data by 99.9% reduction, thus cutting computation load for neural network training. This paper presented Backpropagation Neural Network (BNN) as main classification algorithm which done by setting its parameters, learning and testing the data. So the model could classify retinal image into one of two classes, namely the normal retina and retina with high blood pressure, based on the BNN output result. The proposed model result showed that testing accuracy up to 86.36%.

Computational Thinking is introduced as a problem-solving ability that is important for future generations to master. The mastery of computational thinking skills from an early age prepares children to anticipate competition and pursue success in the future. Robotics devices are widely advocated as interactive learning media to facilitate Computational Thinking development. Educational robotics have grown from ideas that represent critical stages of Computational Thinking. The application of robotics in teaching computational thinking skills is increasingly used and is evident in the literature. The paper reviews the development of an innovative robotics device in facilitating understanding of computational thinking in young children in terms of several aspects of computational thinking indicators. A review was also conducted to see the advantages and disadvantages of commercially available robotics devices. This paper was prepared by applying critical analysis methods to the literature published from 1952 to 2017 in international journals and proceedings. As a recommendation for future research, this paper proposes for an educational robotic development that facilitates the of computational thinking skills, especially for early childhood education.

Pattern analysis has been shown to have higher reliability for melanoma and nevus classification. The ABCD method, which is common to be used as a melanoma diagnosis method, has been shown to have inappropriate weighting for each parameter. In addition, pattern analysis has been shown to have a higher success for diagnosing melanoma. In this paper, we choose the Grey Level Co-occurrence Matrix (GLCM) as a texture parameter to represent the pattern of melanoma. We also choose Deep Neural Network (DNN) to retrieve information from the data set. DNN has the capability of analyzing data with a high level of abstraction. Therefore, we choose DNN as a method to classify melanoma and nevus. We use the International Skin Imaging Collaboration (ISIC) archive database as our training and validation data. We use 773 nevi with 870 melanoma images as training data and separate 200 Nevus and 200 Melanoma images as validation data. We achieve 81.75% diagnostic accuracy, 75.5% sensitivity, and 88% specificityPattern analysis has been shown to have higher reliability for melanoma and nevus classification. The ABCD method, which is common to be used as a melanoma diagnosis method, has been shown to have inappropriate weighting for each parameter. In addition, pattern analysis has been shown to have a higher success for diagnosing melanoma. In this paper, we choose the Grey Level Co-occurrence Matrix (GLCM) as a texture parameter to represent the pattern of melanoma. We also choose Deep Neural Network (DNN) to retrieve information from the data set. DNN has the capability of analyzing data with a high level of abstraction. Therefore, we choose DNN as a method to classify melanoma and nevus. We use the International Skin Imaging Collaboration (ISIC) archive database as our training and validation data. We use 773 nevi with 870 melanoma images as training data and separate 200 Nevus and 200 Melanoma images as validation data. We achieve 81.75% diagnostic accuracy, 75.5% sensitivity, and 88% specificity

A method to create mammography phantoms with radiopaque ink and an inkjet printer was developed. A mammography image was printed on several sheets of bond paper, modulated by a calibration function, creating a phantom image that can emulate the original object in a clinical image. The radiopaque ink on paper was stable, at least, over a month. The representation of the mammography image was visually equivalent, and the limitations of the method come from noise, alignment difficulties, and dynamic range compression.

This paper presents a computational method to extract optimum rectangular Regions of Interest (RoI) in images with an associated saliency map. Although saliency maps provide an individual relevance measure for each pixel, to find the sub-image (i.e., rectangular region) that contains the set of the most relevant pixels requires an optimisation procedure to define the boundaries of the best RoI. This is achieved by the method devised in the paper, by following an approach based on balancing the amount of relevant information that is included and excluded from the RoI. The results show that such method is capable of finding the most relevant rectangular RoI and thus to extract the optimum sub-images according to the relevance measure given by a generic saliency map. Since the method is not tied to any particular type of images, it finds application in quite different fields, such as salient object extraction and processing in industry and surveillance, image compression using attention modelling, biomedical imaging, etc.

With the rapid development of Internet social media, excavating the emotional tendencies of the short text information from the Internet, the acquisition of useful information has attracted the attention of researchers. At present, the commonly used can be attributed to the rule-based classification and statistical machine learning classification methods. Although micro-blog sentiment analysis has made good progress, there still exist some shortcomings such as not highly accurate enough and strong dependence from sentiment classification effect. Aiming at the characteristics of Chinese short texts, such as less information, sparse features, and diverse expressions, this paper considers expanding the original text by mining related semantic information from the reviews, forwarding and other related information. First, this paper uses Word2vec to compute word similarity to extend the feature words. And then uses an ensemble classifier composed of SVM, KNN and HMM to analyze the emotion of the short text of micro-blog. The experimental results show that the proposed method can make good use of the comment forwarding information to extend the original features. Compared with the traditional method, the accuracy, recall and F1 value obtained by this method have been improved.

By studying the defense scheme of Network attacks, this paper propose an effective traffic features selection algorithm based on k-means++ clustering to deal with the problem of high dimensionality of traffic features which extracted from cyber-attacks samples. Firstly, this algorithm divide the original feature set into attack traffic feature set and background traffic feature set by the clustering. Then, we calculates the variation of clustering performance after removing a certain feature. Finally, evaluating the degree of distinctiveness of the feature vector according to the result. Among them, the effective feature vector is whose degree of distinctiveness exceeds the set threshold. The purpose of this paper is to select out the effective features from the extracted original feature set. In this way, it can reduce the dimensionality of the features so as to reduce the space-time overhead of subsequent detection. The experimental results show that the proposed algorithm is feasible and it has some advantages over other selection algorithms.

Orthophoto map is one in which the line-work and symbols of a conventional map are replaced by the orthophotographic image. The development of current technology has introduced the Unmanned Aerial Vehicles (UAV) as an alternative to conventional photogrammetry in creating orthophoto map. This paper aims to show the potential of UAV in orthophoto map creation. UAV photogrammetry describes a photogrammetric measurement platform, which operates remotely controlled, semi-autonomously, or autonomously, without a pilot sitting in the vehicle. The large format aerial camera in conventional photogrammetry is replaced with small format digital camera in UAV photogrammetry. For general area, the amount of forward overlap and side overlap is 75% and 60% respectively. For forest area, the amount of forward overlap and side overlap is 85% and 70% respectively as the UAV flight altitude will be higher. The whole workflow is introduced in the paper, where 461 images were collected with DJI Phantom 4 Pro with camera model FC6310. The flight mission was completed using PIX4D Capture. An orthophoto map covering an area of 250458 m2 was produced using Agisoft PhotoScan. The output is an orthophoto map of UTHM Pagoh Campus with good quality that can portray the overview of the area and benefits the users with the most updated information.

Detection of blood group is an essential factor in critical conditions before performing blood transfer. At presently tests are conducted by lab technicians manually in the laboratory. When the test is done by technicians with large samples it becomes monotonous to do and sometimes it leads to incorrect results and even its time consuming to get the result. The research survey proposal is to reduce the physical work to identify the blood group with a paper-based device. The paper is having a long thermometer with two ends. By using this we can detect the blood type by changing the color of the paper. Chemical reactions between dye, Bromo creosol green, and blood serum proteins, were performed to test the blood sample. The paper becomes teal or brown color, depending on whether the association of antibodies and antigens are present. It gives the result within 30 seconds, which is quicker than traditional detection system. The tested blood sample had a good accuracy rate.Detection of blood group is an essential factor in critical conditions before performing blood transfer. At presently tests are conducted by lab technicians manually in the laboratory. When the test is done by technicians with large samples it becomes monotonous to do and sometimes it leads to incorrect results and even its time consuming to get the result. The research survey proposal is to reduce the physical work to identify the blood group with a paper-based device. The paper is having a long thermometer with two ends. By using this we can detect the blood type by changing the color of the paper. Chemical reactions between dye, Bromo creosol green, and blood serum proteins, were performed to test the blood sample. The paper becomes teal or brown color, depending on whether the association of antibodies and antigens are present. It gives the result within 30 seconds, which is quicker than traditional detection system. The tested blood sample had a good accuracy rate.

Paper is focused on problem of robust feature detector performance in case when intensity distribution is uneven over image. Results of feature detector simulation in case of various intensity changes, equations for estimation rate of false alarms and probability of correct detection are presented in paper.Paper is focused on problem of robust feature detector performance in case when intensity distribution is uneven over image. Results of feature detector simulation in case of various intensity changes, equations for estimation rate of false alarms and probability of correct detection are presented in paper.

Optical Coherence Tomography (OCT) is an imaging technique used to localize the intra-retinal boundaries for the diagnostics of macular diseases. Due to speckle noise, low image contrast and accurate segmentation of individual retinal layers is difficult. Due to this, a method for retinal layer segmentation from OCT images is presented. This paper proposes a pre-processing filtering approach for denoising and segmentation methods for segmenting retinal layers OCT images using graph based segmentation technique. These techniques are used for segmentation of retinal layers for normal as well as patients with Diabetic Macular Edema. The algorithm based on gradient information and shortest path search is applied to optimize the edge selection. In this paper the four main layers of the retina are segmented namely Internal limiting membrane (ILM), Retinal pigment epithelium (RPE), Inner nuclear layer (INL) and Outer nuclear layer (ONL). The proposed method is applied on a database of OCT images of both ten normal and twenty DME affected patients and the results are found to be promising.

Data analysis is an important issue for companies as it provides deep business insights thus facilitating their performance. Usually enterprise applications and platforms for Big Data processing live in the Java ecosystem. Therefore it is important to investigate how Java-based platforms for data analytics and machine learning work and the functionality they provide. The paper analyzes the most widely used machine learning libraries in the Java world - Deeplearning4j and Spark MLlib. Using data lakes is a common practice for storing large amount of data for data analytic tasks. Hadoop has proved to be the common choice for building data lakes. The paper will review Hadoop ecosystem components, their advantages and potential problems concerning machine learning tasks.

DNA microarray technology is used to analyze thousands of gene expression data simultaneously and a very important task for drug development and test, function annotation, and cancer diagnosis. Various clustering methods have been used for analyzing gene expression data. However, when analyzing very large and heterogeneous collections of gene expression data, conventional clustering methods often cannot produce a satisfactory solution. Clustering algorithm has been used as an alternative approach to identify structures from gene expression data. In this paper, we introduce a transform technique based on Singular Value Decomposition (SVD) to identify normalized matrix of gene expression data followed by Partitioning Around Medoids (PAM) to cluster and then displaying the best cluster according to Davis Bouldin Index (DBI) based on Agglomerative Hierarchical Clustering (AHC). Experimental study on standard dataset demonstrated the effectiveness of the algorithm in gene expression data.DNA microarray technology is used to analyze thousands of gene expression data simultaneously and a very important task for drug development and test, function annotation, and cancer diagnosis. Various clustering methods have been used for analyzing gene expression data. However, when analyzing very large and heterogeneous collections of gene expression data, conventional clustering methods often cannot produce a satisfactory solution. Clustering algorithm has been used as an alternative approach to identify structures from gene expression data. In this paper, we introduce a transform technique based on Singular Value Decomposition (SVD) to identify normalized matrix of gene expression data followed by Partitioning Around Medoids (PAM) to cluster and then displaying the best cluster according to Davis Bouldin Index (DBI) based on Agglomerative Hierarchical Clustering (AHC). Experimental study on standard dataset demonstrated the effectiveness of the algorithm in gene expression data.

Based on nonsubsampled contourlet transform (NSCT) and modified pulse coupled neural network (MPCNN), the paper proposes an effective method of image fusion. Firstly, the paper decomposes the source image into the low-frequency components and high-frequency components using NSCT, and then processes the low-frequency components by regional statistical fusion rules. For high-frequency components, the paper calculates the spatial frequency (SF), which is input into MPCNN model to get relevant coefficients according to the fire-mapping image of MPCNN. At last, the paper restructures the final image by inverse transformation of low-frequency and high-frequency components. Compared with the wavelet transformation (WT) and the traditional NSCT algorithm, experimental results indicate that the method proposed in this paper achieves an improvement both in human visual perception and objective evaluation. It indicates that the method is effective, practical and good performance.Based on nonsubsampled contourlet transform (NSCT) and modified pulse coupled neural network (MPCNN), the paper proposes an effective method of image fusion. Firstly, the paper decomposes the source image into the low-frequency components and high-frequency components using NSCT, and then processes the low-frequency components by regional statistical fusion rules. For high-frequency components, the paper calculates the spatial frequency (SF), which is input into MPCNN model to get relevant coefficients according to the fire-mapping image of MPCNN. At last, the paper restructures the final image by inverse transformation of low-frequency and high-frequency components. Compared with the wavelet transformation (WT) and the traditional NSCT algorithm, experimental results indicate that the method proposed in this paper achieves an improvement both in human visual perception and objective evaluation. It indicates that the method is effective, practical and good performance.

Forecasting soil pollution is an important part of environmental protection around the world. The work is devoted to the use of artificial neural networks for spatial prediction of the anomalously distributed chemical element Chromium (Cr). Generalized regression neural networks (GRNN) and multilayer perceptrons (MLP) are classes of neural networks widely used for continuous mapping of functions. Each network has its pros and cons; however, both demonstrated quick preparation and good modeling capabilities. In the paper, we studied and compared two neural networks: GRNN and MLP. The present study is based on real data sets on surface contamination of Cr at a specific location in subarctic Novy Urengoy, Russia, obtained during previous screening. The network structures were chosen during computer modeling, based on the minimization of the root mean squared error (RMSE). GRNN (anisotropy) showed the best predictive accuracy in comparison with MLP and kriging.Forecasting soil pollution is an important part of environmental protection around the world. The work is devoted to the use of artificial neural networks for spatial prediction of the anomalously distributed chemical element Chromium (Cr). Generalized regression neural networks (GRNN) and multilayer perceptrons (MLP) are classes of neural networks widely used for continuous mapping of functions. Each network has its pros and cons; however, both demonstrated quick preparation and good modeling capabilities. In the paper, we studied and compared two neural networks: GRNN and MLP. The present study is based on real data sets on surface contamination of Cr at a specific location in subarctic Novy Urengoy, Russia, obtained during previous screening. The network structures were chosen during computer modeling, based on the minimization of the root mean squared error (RMSE). GRNN (anisotropy) showed the best predictive accuracy in comparison with MLP and kriging.

The technique of obtaining a wider field-of-view of an image to get high resolution integrated image is normally required for development of panorama of a photographic images or scene from a sequence of part of multiple views. There are various image stitching methods developed recently. For image stitching five basic steps are adopted stitching which are Feature detection and extraction, Image registration, computing homography, image warping and Blending. This paper provides review of some of the existing available image feature detection and extraction techniques and image stitching algorithms by categorizing them into several methods. For each category, the basic concepts are first described and later on the necessary modifications made to the fundamental concepts by different researchers are elaborated. This paper also highlights about the some of the fundamental techniques for the process of photographic image feature detection and extraction methods under various illumination conditions. The Importance of Image stitching is applicable in the various fields such as medical imaging, astrophotography and computer vision. For comparing performance evaluation of the techniques used for image features detection three methods are considered i.e. ORB, SURF, HESSIAN and time required for input images feature detection is measured. Results obtained finally concludes that for daylight condition, ORB algorithm found better due to the fact that less tome is required for more features extracted where as for images under night light condition it shows that SURF detector performs better than ORB/HESSIAN detectors.

In the paper a solution proposal for indoor navigation and obstacle detection problem in hospital environment for a humanoid that is 150 cm tall was presented. As a medical assistant, the robot is expected to autonomously move between hospital rooms with extra care about people safety. The overview of existing solutions was made, mostly focused on Simultaneous Localization and Mapping (SLAM) applications including human-aware algorithms. Considering the robot size, a common approach for mobile robots utilizing single 2D laser scanner or RGB+Depth camera is not applicable as vertical field of view of scans is too narrow to detect all obstacles. We proposed a hybrid system combining laser scanner, multiple cameras and distance sensors. The hardware architecture and software design in Robot Operating System (ROS) were presented. The last part of the paper includes the future plans and proposals of improvements for the robot.

The human hand gesture can record and reproduce the posture and action information of the hand, which is of great significance to people’s production and life. This paper has improved the existing data gloves based on micro inertial technology, and integrates the magnetic field localization method into finger gesture measurements. The strap down inertial navigation technology and the magnetic localization technology are combined in this paper to make the advantages complement each other, and a low cost and high degree of freedom of data gloves are put forward to realize the way of data gloves in the past.

In order to solve the problem of how to reconstruct the original image under the condition of unknown sparse basis, this paper proposes an image reconstruction method based on blind compressed sensing model. In this model, the image signal is regarded as the product of a sparse coefficient matrix and a dictionary matrix. Based on the existing blind compressed sensing theory, the optimal solution is solved by the alternative minimization method. The proposed method solves the problem that the sparse basis in compressed sensing is difficult to represent, which restrains the noise and improves the quality of reconstructed image. This method ensures that the blind compressed sensing theory has a unique solution and can recover the reconstructed original image signal from a complex environment with a stronger self-adaptability. The experimental results show that the image reconstruction algorithm based on blind compressed sensing proposed in this paper can recover high quality image signals under the condition...

Image segmentation methods are essential in computer vision applications. This paper proposes a new algorithm to detect a hand in an image and then to create a convex hull around it. A hand thus detected, via a stationary video camera placed in front of the user, could serve to control robots of various kinds. This goal is achieved by proposing a new process of hand segmentation and recognition utilizing Open CV image processing framework. The algorithm is able to identify hand in the picture and prepare the separated object for gesture recognition. Final implementation was written and verified in the C# language.Image segmentation methods are essential in computer vision applications. This paper proposes a new algorithm to detect a hand in an image and then to create a convex hull around it. A hand thus detected, via a stationary video camera placed in front of the user, could serve to control robots of various kinds. This goal is achieved by proposing a new process of hand segmentation and recognition utilizing Open CV image processing framework. The algorithm is able to identify hand in the picture and prepare the separated object for gesture recognition. Final implementation was written and verified in the C# language.

Iris is regarded as the most unique biometric identification. This paper proposes a new feature extraction based on iris texture patterns with Principal Component Analysis (PCA). PCA is used to store computing process in classification process. The focus of this paper was to compare the accuracy result of classification methods of distance measurement such as Euclidean Distance, City Block Distance, Chebyshev Distance, Canberra Distance and Bray-Curtis Distance. Accuracy test shows that PCA can be used in various classification methods that use distance measurement.Iris is regarded as the most unique biometric identification. This paper proposes a new feature extraction based on iris texture patterns with Principal Component Analysis (PCA). PCA is used to store computing process in classification process. The focus of this paper was to compare the accuracy result of classification methods of distance measurement such as Euclidean Distance, City Block Distance, Chebyshev Distance, Canberra Distance and Bray-Curtis Distance. Accuracy test shows that PCA can be used in various classification methods that use distance measurement.

While the prediction of spatial distribution of potential landslide occurrences is a primary interest in landslide hazard mitigation, it remains a challenging task. To overcome the scarceness of complete, sufficiently detailed geomorphological attributes and environmental conditions, various machine-learning techniques are increasingly applied to effectively map landslide susceptibility for large regions. Nevertheless, limited review papers are devoted to this field, particularly on the various domain specific applications of machine learning techniques. Available literature often report relatively good predictive performance, however, papers discussing the limitations of each approaches are quite uncommon. The foremost aim of this paper is to narrow these gaps in literature and to review up-to-date machine learning and ensemble learning techniques applied in landslide susceptibility mapping. It provides new readers an introductory understanding on the subject matter and researchers a contemporary review ...

This research paper aims to propose a hybrid of ant colony optimization (ACO) and k-nearest neighbor (KNN) algorithms as feature selections for selecting and choosing relevant features from customer review datasets. Information gain (IG), genetic algorithm (GA), and rough set attribute reduction (RSAR) were used as baseline algorithms in a performance comparison with the proposed algorithm. This paper will also discuss the significance test, which was used to evaluate the performance differences between the ACO-KNN, IG-GA, and IG-RSAR algorithms. This study evaluated the performance of the ACO-KNN algorithm using precision, recall, and F-score, which were validated using the parametric statistical significance tests. The evaluation process has statistically proven that this ACO-KNN algorithm has been significantly improved compared to the baseline algorithms. The evaluation process has statistically proven that this ACO-KNN algorithm has been significantly improved compared to the baseline algorithms. In ...

With the ever increasing of internet applications and social networking sites, people nowadays can easily express their feelings towards any products and services. These online reviews act as an important source for further analysis and improved decision making. These reviews are mostly unstructured by nature and thus, need processing like sentiment analysis and classification to provide a meaningful information for future uses. In text analysis tasks, the appropriate selection of words/features will have a huge impact on the effectiveness of the classifier. Thus, this paper explores the effect of the pre-processing strategies in the sentiment analysis of online movie reviews. In this paper, supervised machine learning method was used to classify the reviews. The support vector machine (SVM) with linear and non-linear kernel has been considered as classifier for the classification of the reviews. The performance of the classifier is critically examined based on the results of precision, recall, f-measure,...

This paper investigates a habituation based approach in detecting visual changes using video surveillance systems in a passive environment. Various techniques have been introduced for dynamic environment such as motion detection, object classification and behaviour analysis. However, in a passive environment, most of the scenes recorded by the surveillance system are normal. Therefore, implementing a complex analysis all the time in the passive environment resulting on computationally expensive, especially when using a high video resolution. Thus, a mechanism of attention is required, where the system only responds to an abnormal event. This paper proposed a novelty detection mechanism in detecting visual changes and a habituation based approach to measure the level of novelty. The objective of the paper is to investigate the feasibility of the habituation based approach in detecting visual changes. Experiment results show that the approach are able to accurately detect the presence of novelty as deviatio...

This paper investigates applications of non-negative matrix factorization as feature selection method to select the features from gray level co-occurrence matrix. The proposed approach is used to classify dementia using MRI data. In this study, texture analysis using gray level co-occurrence matrix is done to feature extraction. In the feature extraction process of MRI data, we found seven features from gray level co-occurrence matrix. Non-negative matrix factorization selected three features that influence of all features produced by feature extractions. A Naive Bayes classifier is adapted to classify dementia, i.e. Alzheimer’s disease, Mild Cognitive Impairment (MCI) and normal control. The experimental results show that non-negative factorization as feature selection method able to achieve an accuracy of 96.4% for classification of Alzheimer’s and normal control. The proposed method also compared with other features selection methods i.e. Principal Component Analysis (PCA).This paper investigates applications of non-negative matrix factorization as feature selection method to select the features from gray level co-occurrence matrix. The proposed approach is used to classify dementia using MRI data. In this study, texture analysis using gray level co-occurrence matrix is done to feature extraction. In the feature extraction process of MRI data, we found seven features from gray level co-occurrence matrix. Non-negative matrix factorization selected three features that influence of all features produced by feature extractions. A Naive Bayes classifier is adapted to classify dementia, i.e. Alzheimer’s disease, Mild Cognitive Impairment (MCI) and normal control. The experimental results show that non-negative factorization as feature selection method able to achieve an accuracy of 96.4% for classification of Alzheimer’s and normal control. The proposed method also compared with other features selection methods i.e. Principal Component Analysis (PCA).

Malaria is a global health problem, particularly in Africa and south Asia where it causes countless deaths and morbidity cases. Efficient control and prompt of this disease require early detection and accurate diagnosis due to the large number of cases reported yearly. To achieve this aim, this paper proposes an image segmentation approach via unsupervised pixel segmentation of malaria parasite to automate the diagnosis of malaria. In this study, a modified clustering algorithm namely enhanced k-means (EKM) clustering, is proposed for malaria image segmentation. In the proposed EKM clustering, the concept of variance and a new version of transferring process for clustered members are used to assist the assignation of data to the proper centre during the process of clustering, so that good segmented malaria image can be generated. The effectiveness of the proposed EKM clustering has been analyzed qualitatively and quantitatively by comparing this algorithm with two popular image segmentation techniques namely Otsu’s thresholding and k-means clustering. The experimental results show that the proposed EKM clustering has successfully segmented 100 malaria images of P. vivax species with segmentation accuracy, sensitivity and specificity of 99.20%, 87.53% and 99.58%, respectively. Hence, the proposed EKM clustering can be considered as an image segmentation tool for segmenting the malaria images.Malaria is a global health problem, particularly in Africa and south Asia where it causes countless deaths and morbidity cases. Efficient control and prompt of this disease require early detection and accurate diagnosis due to the large number of cases reported yearly. To achieve this aim, this paper proposes an image segmentation approach via unsupervised pixel segmentation of malaria parasite to automate the diagnosis of malaria. In this study, a modified clustering algorithm namely enhanced k-means (EKM) clustering, is proposed for malaria image segmentation. In the proposed EKM clustering, the concept of variance and a new version of transferring process for clustered members are used to assist the assignation of data to the proper centre during the process of clustering, so that good segmented malaria image can be generated. The effectiveness of the proposed EKM clustering has been analyzed qualitatively and quantitatively by comparing this algorithm with two popular image segmentation techniques nam...

At present, there are many relevant researches on data processing using classical pattern matching and its improved algorithm, but few researches on statistical data of link-layer bit stream. This paper adopts a frequent statistical method of link-layer bit stream data based on AC-IM algorithm for classical multi-pattern matching algorithms such as AC algorithm has high computational complexity, low efficiency and it cannot be applied to binary bit stream data. The method’s maximum jump distance of the mode tree is length of the shortest mode string plus 3 in case of no missing? In this paper, theoretical analysis is made on the principle of algorithm construction firstly, and then the experimental results show that the algorithm can adapt to the binary bit stream data environment and extract the frequent sequence more accurately, the effect is obvious. Meanwhile, comparing with the classical AC algorithm and other improved algorithms, AC-IM algorithm has a greater maximum jump distance and less time-consuming.

The novel concept of comprehensive weld inspection with use of three sources of data (radiogram image, image composed of laser profilometer 3D data, CCD camera image) and weld images fusion is presented in the paper. The algorithms of data processing aimed at detection and visualization of weld flaws are roughly described. The special attention was paid at presentation and explanation of algorithm for aligning and spatial correction of images composed of data delivered by three independent sources. The results yielded by proposed algorithms for preprocessing of weld images, weld detection, feature extraction, shape extraction, shape matching, filtering, distortion correction, etc. are illustrated with several figures included in the paper. The concise description of designed and implemented laboratory stands is included in the paper too.

This paper mainly introduces the research of human body state recognition technology based on multi-layer perceptron. By using the application environment of ASUS Xtion pro live infrared camera, this paper focuses on the daily behavior recognition of human body in small area. First is the depth image analysis, this paper extracts the characteristics of the postures that need to be classified and we need a large number of samples. Then the collected samples are trained and classified by the multi-layer perceptron model. Finally, the human state collected by the camera can be identified in real time.

This paper realizes the uncalibrated visual servoless control of the Six-DOF manipulator, where the intrinsic and extrinsic parameters of the camera and the robot manipulator model are not known, nor real-time computes the inverse of image Jacobian. And the neural network was used to optimize the trajectory of manipulator end in Cartesian space. Firstly, a biaxial parallel layoutt are proposed to establish the relationship between the moving component along the coordinate axis and the change of single image feature when the end of the robot moves in the Cartesian space. The neural network was used to fit the relationship established above, then the optimization model was established, and the parameter of PID controller were optimized by neural network. Finally, the simulation and comparison experiments are investigated, and the results indicate the controller posed in this paper is effective and stable.

This paper presents the concept and the modern advances in personalized medicine that rely on technology and review the existing tools for early detection of breast cancer. The breast cancer types and distribution worldwide is discussed. It is spent time to explain the importance of identifying the normality and to specify the main classes in breast cancer, benign or malignant. The main purpose of the paper is to propose a conceptual model for early detection of breast cancer based on machine learning for processing and analysis of medical big dataand further knowledge discovery for personalized treatment. The proposed conceptual model is realized by using Naive Bayes classifier. The software is written in python programming language and for the experiments the Wisconsin breast cancer database is used. Finally, the experimental results are presented and discussed.

Spectrum prediction technology is an effective way to solve the problems of processing latency, spectrum access, spectrum collision and energy consumption in cognitive radio networks. Spectral prediction technology is divided into three categories according to its nature, namely, spectral prediction method based on regression analysis, spectrum prediction method based on Markov model and spectrum prediction method based on machine learning. By analyzing and comparing the three kinds of prediction models, the author hopes to provide some reference for the later researchers. In this paper, the development situation, practical application and existent problems of three kinds of forecasting models are analyzed and summarized. On this basis, this paper discusses the development trend of the next step.

The aim of this paper is a comparison of the existing classification algorithms with different parameters, and selection those ones, which allows solving the problem of primary diagnosis of cervical intraepithelial neoplasia (CIN), as it characterizes the condition of the body in the precancerous stage. The paper describes a feature selection process, as well as selection of the best models for a multiclass classification. © 2017 Author(s).

This paper presents an image classification algorithm based on spatial sparse coding model and random forest. Firstly, SIFT feature extraction of the image; and then use the sparse encoding theory to generate visual vocabulary based on SIFT features, and using the visual vocabulary of SIFT features into a sparse vector; through the combination of regional integration and spatial sparse vector, the sparse vector gets a fixed dimension is used to represent the image; at last random forest classifier for image sparse vectors for training and testing, using the experimental data set for standard test Caltech-101 and Scene-15. The experimental results show that the proposed algorithm can effectively represent the features of the image and improve the classification accuracy. In this paper, we propose an innovative image recognition algorithm based on image segmentation, sparse coding and multi instance learning. This algorithm introduces the concept of multi instance learning, the image as a multi instance bag...

Traditional supervised classifiers neglect a large amount of data which not have sufficient follow-up information, only work with labeled data. Consequently, the small sample size limits the advancement of design appropriate classifier. In this paper, a transductive learning method which combined with the filtering strategy in transductive framework and progressive labeling strategy is addressed. The progressive labeling strategy does not need to consider the distribution of labeled samples to evaluate the distribution of unlabeled samples, can effective solve the problem of evaluate the proportion of positive and negative samples in work set. Our experiment result demonstrate that the proposed technique have great potential in cancer prediction based on gene expression.Traditional supervised classifiers neglect a large amount of data which not have sufficient follow-up information, only work with labeled data. Consequently, the small sample size limits the advancement of design appropriate classifier. In this paper, a transductive learning method which combined with the filtering strategy in transductive framework and progressive labeling strategy is addressed. The progressive labeling strategy does not need to consider the distribution of labeled samples to evaluate the distribution of unlabeled samples, can effective solve the problem of evaluate the proportion of positive and negative samples in work set. Our experiment result demonstrate that the proposed technique have great potential in cancer prediction based on gene expression.

In recent years, with the rapid development of deep learning, it has been widely used in the field of natural language processing. In this paper, I use the method of deep learning to achieve Chinese word segmentation, with large-scale corpus, eliminating the need to construct additional manual characteristics. In the process of Chinese word segmentation, the first step is to deal with the corpus, use word2vec to get word embedding of the corpus, each character is 50. After the word is embedded, the word embedding feature is fed to the bidirectional LSTM, add a linear layer to the hidden layer of the output, and then add a CRF to get the model implemented in this paper. Experimental results show that the method used in the 2014 People’s Daily corpus to achieve a satisfactory accuracy.