Search

BACKGROUND: Snoring source analysis is essential for an appropriate surgical decision for both simple snorers and obstructive sleep apnea/hypopnea syndrome (OSAHS) patients. OBJECTIVE: As snoring sounds carry significant information about tissue vibrations within the upper airway, a new feature entitled compressed histogram of oriented gradients (CHOG) is proposed to recognize vibration patterns of the snoring source acoustically by compressing histogram of oriented gradients (HOG) descriptors via the multilinear principal component analysis (MPCA) algorithm. METHODS: Each vibration pattern corresponds to a sole or combinatorial vibration among the four upper airway soft tissues of soft palate, lateral pharyngeal wall, tongue base, and epiglottis. 1037 snoring events from noncontact sound recordings of 76 simple snorers or OSAHS patients during drug-induced sleep endoscopy (DISE) were evaluated. RESULTS: With a support vector machine (SVM) as the classifier, the proposed CHOG achieved a recognition accuracy of 89.8% for the seven observable vibration patterns of the snoring source categorized in our most recent work. CONCLUSION: The CHOG outperforms other single features widely used for acoustic analysis of sole vibration site. [ABSTRACT FROM AUTHOR]

Selective serotonin reuptake inhibitors (SSRIs) are most adopted therapeutics marketed for major depression, and the efficacy of which are greatly reduced by their delayed onset of action and undesirable side effects. 5-HT1A receptor partial agonist and SERT inhibitor (SPARI) was proposed as a novel strategy to overcome the shortage of efficacy by a negative feedback control of 5-HT1A receptor. However, only one SPARI (vilazodone) has been approved for clinical use, and none is currently in clinical trial, which demonstrates a strong need for searching more novel SPARIs to facilitate antidepressants discovery. This work applied a combinatorial virtual screening method (CVSM) by integrating multiple tools. Statistic analysis reveals that CVSM surpasses single virtual screening methods in terms of hit rates and enrichment factors. By adopting optimized CVSM, 91 promising dual target leads form 15 scaffolds were identified, and 40% of these scaffolds have already been reported to show antidepressant related therapeutic effects. In sum, CVSM is capable in identifying novel SPARIs from large chemical libraries with extremely low false hit rate. [ABSTRACT FROM AUTHOR]

Injury to the nervous system can lead to irreversible problems as nervous tissues have limited regenerative capability. Therefore it is imperative to find an objective, reliable, cheap, and easy-to-apply method that separates nervous fibers from muscles and blood vessels. The aim of this study is to determine structural differences that can aid in easy and reliable identification of nervous fibers. We analyzed light reflectance from these tissues from 230 nm to 1000 nm and found that in the range of 400 nm-600 nm nervous fibers have higher reflectance in comparison to others. Therefore, we generated distinct features in this range and utilized support vector machine to automatically classify samples. Classification performance demonstrated that light reflectance is a good candidate feature that can help to classify nervous tissue. [ABSTRACT FROM AUTHOR]

Support vector machine (SVM) is one of the most effective classification methods for cancer detection. The efficiency and quality of a SVM classifier depends strongly on several important features and a set of proper parameters. Here, a series of classification analyses, with one set of photoacoustic data from ovarian tissues ex vivo and a widely used breast cancer dataset- the Wisconsin Diagnostic Breast Cancer (WDBC), revealed the different accuracy of a SVM classification in terms of the number of features used and the parameters selected. A pattern recognition system is proposed by means of SVM-Recursive Feature Elimination (RFE) with the Radial Basis Function (RBF) kernel. To improve the effectiveness and robustness of the system, an optimized tuning ensemble algorithm called as SVM-RFE(C) with correlation filter was implemented to quantify feature and parameter information based on cross validation. The proposed algorithm is first demonstrated outperforming SVM-RFE on WDBC. Then the best accuracy of 94.643% and sensitivity of 94.595% were achieved when using SVM-RFE(C) to test 57 new PAT data from 19 patients. The experiment results show that the classifier constructed with SVM-RFE(C) algorithm is able to learn additional information from new data and has significant potential in ovarian cancer diagnosis. [ABSTRACT FROM AUTHOR]

Prediction of protein structural classes is of great significance to better understand protein folding patterns. An array of methods has been proposed to predict these structures based on sequences. However, the accuracy is strongly affected by the homology of sequences. In the present study, the features based on correlation coefficient of sequence and amino acid composition are extracted. Flexible neutral tree is employed as the classification model. To examine the performance of this method, four benchmark datasets are selected. Altogether, the results show that a higher prediction accuracy of alpha/beta can be achieved by the method compared to others. [ABSTRACT FROM AUTHOR]

Polyadenylation [poly(A)] of mRNA is a critical step during gene expression, which plays an important role in the termination of transcription. Prediction of poly(A) sites can help identify 3' ends of genes and improve genome annotation. Due to the limited knowledge of poly(A) signals in plants, predictive modeling of poly(A) sites in agricultural crops remains challenging. Recent studies have uncovered widespread occurrences of alternative poly(A) (APA) sites in intron and coding sequence (CDS), whereas the study on the prediction of these APA sites is scarce. In this study, four feature representation methods, involving a position weight matrix, the k-gram frequency, core hexamers, and a transition matrix, were adopted to characterize poly(A) signals surrounding APA sites. The classification model was built to predict each group of APA sites. Experimental results showed that this model was effective in the identification of APA sites located in different genomic regions, with a compromise between sensitivity and specificity higher than 87%. Compared with previous model PASS rice, accuracies for the prediction of APA sites in 3'-UTR, intron and CDS were enhanced by 5%, 7%, and 27%, respectively. This model will contribute to genetic engineering by enabling researchers to control poly(A) site selection. [ABSTRACT FROM AUTHOR]

Classification methods are a crucial direction in the current study of brain-computer interfaces (BCIs). To improve the classification accuracy for electroencephalogram (EEG) signals, a novel KF-PP-SVM (kernel fisher, posterior probability, and support vector machine) classification method is developed. Its detailed process entails the use of common spatial patterns to obtain features, based on which the within-class scatter is calculated. Then the scatter is added into the kernel function of a radial basis function to construct a new kernel function. This new kernel is integrated into the SVM to obtain a new classification model. Finally, the output of SVM is calculated based on posterior probability and the final recognition result is obtained. To evaluate the effectiveness of the proposed KF-PP-SVM method, EEG data collected from laboratory are processed with four different classification schemes (KF-PP-SVM, KF-SVM, PP-SVM, and SVM). The results showed that the overall average improvements arising from the use of the KF-PP-SVM scheme as opposed to KF-SVM, PP-SVM and SVM schemes are 2.49%, 5.83 % and 6.49 % respectively. [ABSTRACT FROM AUTHOR]

Silicosis remains one of the most harmful occupational respiratory diseases. It threatens the workers exposed to dust environment. Chest radiograph is the main available image source for silicosis diagnosis according to the diagnostic criteria of pneumoconiosis (DCP). Automatic detection and recognition of silicosis in chest radiograph has great importance on aiding the process of silicosis diagnosis. This paper proposes a multi-scale opacity detection approach to detect all suspected opacities from the chest radiograph. A support vector machine (SVM) based computer-aided silicosis diagnosis is proposed to recognize silicosis opacity from a large amount of candidate regions, and gives processing result for radiologist reference. Comprehensive experiments conducted on real world chest radiographs demonstrate that the proposed approach can reveal changes of silicosis pathology well, and it can be adopted as an effective tool for automatic silicosis diagnosis. [ABSTRACT FROM AUTHOR]

Plaque assaying, measurement of the number, diameter, and area of plaques in a Petri dish image, is a standard procedure gauging the concentration of phage in biology. This paper presented a novel and effective method for implementing automatic plaque assaying. The method was mainly comprised of the following steps: In the training stage, after preprocessing the images for noise suppression, an initial training set was readied by sampling positive (with a plaque at the center) and negative (plaque-free) patches from the training images, and extracting the HOG features from each patch. The linear SVM classifier was trained in a self-learnt supervised learning strategy to avoid possible missing detection. Specifically, the training set which contained positive and negative patches sampled manually from training images was used to train the preliminary classifier which exhaustively searched the training images to predict the label for the unlabeled patches. The mislabeled patches were evaluated by experts and relabeled. And all the newly labeled patches and their corresponding HOG features were added to the initial training set to train the final classifier. In the testing stage, a sliding-window technique was first applied to the unseen image for obtaining HOG features, which were inputted into the classifier to predict whether the patch was positive. Second, a locally adaptive Otsu method was performed on the positive patches to segment the plaques. Finally, after removing the outliers, the parameters of the plaques were measured in the segmented plaques. The experimental results demonstrated that the accuracy of the proposed method was similar to the one measured manually by experts, but it took less than 30 seconds. [ABSTRACT FROM AUTHOR]

In the neural science society, multi-subject brain decoding is of great interest. However, due to the variability of activation patterns across brains, it is difficult to build an effective decoder using fMRI samples pooled from different subjects. In this paper, a hierarchical model is proposed to extract robust features for decoding. With feature selection for each subject treated as a separate task, a novel multi-task feature selection method is introduced. This method utilizes both complementary information among subjects and local correlation between brain areas within a subject. Finally, using fMRI samples pooled from all subjects, a linear support vector machine (SVM) classifier is trained to predict 2-D stimuli-related images or 3-D stimuli-related images. The experimental results demonstrated the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Varieties of neurophysiological measures have been utilized in visual attention studies. The linear parameters like power spectrum are the most commonly used features in the existing studies. In this paper, however, nonlinear parameters including approximate entropy, sample entropy and multiscale entropy were tested. All subjects were instructed to perform tasks with three different attention levels (i.e. attention, no attention and rest) in two experiments. Nonlinear features were extracted from the EEG signals. Then, statistical analyses and classification with support vector machine (SVM) were performed. A comparison between the classification results based on the linear feature / and the sample entropy was performed for further analysis. The results suggest that sample entropy stands out in the dynamical parameters with the accuracies of 76.19% and 85.24% in recognition of three levels of attention for the two experiments respectively. And the further comparison shows that the sample entropy performs even better than the / power ratio. It is suggested that nonlinear dynamical parameters may be indispensable for a robust attention recognition system. [ABSTRACT FROM AUTHOR]

In this paper, we propose a novel method for the detection of microcalcifications using mathematical morphology and a support vector machine (SVM). First, the contrast in the original mammogram was improved by gamma correction and two carefully designed structural elements were used to enhance any microcalcifications. Next, the potential regions were extracted using our proposed dual-threshold technique. Finally, a SVM classifier was used to reduce the number of false positives. The performance of the proposed method was evaluated using the MIAS database. The experimental results demonstrated the efficiency and effectiveness of our method. [ABSTRACT FROM AUTHOR]

The classification of EEG tasks has drawn much attention in recent years. In this paper, a novel classification model based on independent component analysis (ICA) and Extreme learning machine (ELM) is proposed to detect lying. Firstly, ICA and its topography information were used to automatically identify the P300 ICs. Then, time and frequency-domain features were extracted from the reconstructed P3 waveforms. Finally, two classes of feature samples were used to train ELM, Back-propagation network (BPNN) and support vector machine (SVM) classifiers for comparison. The optimal number of P3 ICs and the values of classifier parameter were optimized by the cross-validation procedures. Experimental results show that the presented method (ICA_ELM) achieves the highest training accuracy of 95.40% with extremely less training and testing time on detecting P3 components for the guilty and the innocent subjects. The results indicate that the proposed method can be applied in lie detection. [ABSTRACT FROM AUTHOR]

Automated human larynx carcinoma (HEp-2) cell classification is critical for medical diagnosis. In this paper, we propose a sparse coding-based unsupervised transfer learning method for HEp-2 cell classification. First, the low level image feature is extracted for visual representation. Second, a sparse coding scheme with the Elastic Net penalized convex objective function is proposed for unsupervised feature learning. At last, a Support Vector Machine classifier is utilized for model learning and predicting. To our knowledge, this work is the first to transfer the human-crafted visual feature, sensitive to the variation of appearance and shape during cell movement, to the high level representation which directly denotes the correlation of one sample and the bases in the learnt dictionary. Therefore, the proposed method can overcome the difficulty in discriminative feature formulation for different kinds of cells with irregular and changing visual patterns. Large scale comparison experiments will be conducted to show the superiority of this method. [ABSTRACT FROM AUTHOR]

Correlation-based feature selection (CFS) using neighborhood mutual information (NMI) and particle swarm optimization (PSO) are combined into an ensemble technique in this paper. Based on this observation, an efficient gene selection algorithm, denoted by NMICFS-PSO, is proposed. Several cancer recognition tasks are gathered for testing the proposed technique. Moreover, support vector machine (SVM), integrated with leave-one-out cross-validation and served as a classifier, is employed for six classification profiles to calculate the classification accuracy. Experimental results show that the proposed method can reduce the redundant features effectively and achieve superior performance. The classification accuracy obtained by our method is higher in five out of the six gene expression problems as compared with that of other classifi cation methods. [ABSTRACT FROM AUTHOR]

Protein-protein interaction (PPI) is essential to understand the fundamental processes governing cell biology. The mining and curation of PPI knowledge are critical for analyzing proteomics data. Hence it is desired to classify articles PPI-related or not automatically. In order to build interaction article classification systems, an annotated corpus is needed. However, it is usually the case that only a small number of labeled articles can be obtained manually. Meanwhile, a large number of unlabeled articles are available. By combining ensemble learning and semi-supervised self-training, an ensemble self-training interaction classifier called EST_IACer is designed to classify PPI-related articles based on a small number of labeled articles and a large number of unlabeled articles. A biological background based feature weighting strategy is extended using the category information from both labeled and unlabeled data. Moreover, a heuristic constraint is put forward to select optimal instances from unlabeled data to improve the performance further. Experiment results show that the EST_IACer can classify the PPI related articles effectively and efficiently. [ABSTRACT FROM AUTHOR]

Five well-known arrhythmia classification algorithms were compared in this paper based on the recommendations in AAMI standard. They are C4.5, k-Nearest Neighbor, Multilayer Perceptron, PART, and Support Vector Machine, respectively, with inputs related to heartbeat intervals and ECG morphological features. They were evaluated on three independent datasets, including the MIT-BIH arrhythmia database, a collection of ECG signals acquired from healthy subjects by the wireless Body Sensor Network (BSN) nodes, and a third dataset captured also by the BSN nodes. Results showed the overall accuracy on the MIT-BIH arrhythmia database was approximately 99.04%, with high sensitivity, specificity, and selectivity. When tested with ECG signals acquired from the human subjects, which were partially deteriorated due to several factors, e.g., motion artifacts and data transmission problems, the overall accuracy of 94.19% and that of 81.22% were obtained for static activities and dynamic activities, respectively. In addition, the effects of the signal quality from these human subjects on false alarms were investigated. When false alarms occurring in signal segments with low quality were excluded, the number of false detections reduced from 14.17% to 8.65%. When evaluated on signals generated by the patient simulator, which included several types of premature ventricular contraction without artifacts from body movements, a high classification accuracy was also observed. [ABSTRACT FROM AUTHOR]

Varieties of neurophysiological measures have been utilized in visual attention studies. The linear parameters like power spectrum are the most commonly used features in the existing studies. In this paper, however, nonlinear parameters including approximate entropy, sample entropy and multiscale entropy were tested. All subjects were instructed to perform tasks with three different attention levels (i.e. attention, no attention and rest) in two experiments. Nonlinear features were extracted from the EEG signals. Then, statistical analyses and classification with support vector machine (SVM) were performed. A comparison between the classification results based on the linear feature / and the sample entropy was performed for further analysis. The results suggest that sample entropy stands out in the dynamical parameters with the accuracies of 76.19% and 85.24% in recognition of three levels of attention for the two experiments respectively. And the further comparison shows that the sample entropy performs even better than the / power ratio. It is suggested that nonlinear dynamical parameters may be indispensable for a robust attention recognition system. [ABSTRACT FROM AUTHOR]

The classification of EEG tasks has drawn much attention in recent years. In this paper, a novel classification model based on independent component analysis (ICA) and Extreme learning machine (ELM) is proposed to detect lying. Firstly, ICA and its topography information were used to automatically identify the P300 ICs. Then, time and frequency-domain features were extracted from the reconstructed P3 waveforms. Finally, two classes of feature samples were used to train ELM, Back-propagation network (BPNN) and support vector machine (SVM) classifiers for comparison. The optimal number of P3 ICs and the values of classifier parameter were optimized by the cross-validation procedures. Experimental results show that the presented method (ICA_ELM) achieves the highest training accuracy of 95.40% with extremely less training and testing time on detecting P3 components for the guilty and the innocent subjects. The results indicate that the proposed method can be applied in lie detection. [ABSTRACT FROM AUTHOR]

Correlation-based feature selection (CFS) using neighborhood mutual information (NMI) and particle swarm optimization (PSO) are combined into an ensemble technique in this paper. Based on this observation, an efficient gene selection algorithm, denoted by NMICFS-PSO, is proposed. Several cancer recognition tasks are gathered for testing the proposed technique. Moreover, support vector machine (SVM), integrated with leave-one-out cross-validation and served as a classifier, is employed for six classification profiles to calculate the classification accuracy. Experimental results show that the proposed method can reduce the redundant features effectively and achieve superior performance. The classification accuracy obtained by our method is higher in five out of the six gene expression problems as compared with that of other classifi cation methods. [ABSTRACT FROM AUTHOR]

Five well-known arrhythmia classification algorithms were compared in this paper based on the recommendations in AAMI standard. They are C4.5, k-Nearest Neighbor, Multilayer Perceptron, PART, and Support Vector Machine, respectively, with inputs related to heartbeat intervals and ECG morphological features. They were evaluated on three independent datasets, including the MIT-BIH arrhythmia database, a collection of ECG signals acquired from healthy subjects by the wireless Body Sensor Network (BSN) nodes, and a third dataset captured also by the BSN nodes. Results showed the overall accuracy on the MIT-BIH arrhythmia database was approximately 99.04%, with high sensitivity, specificity, and selectivity. When tested with ECG signals acquired from the human subjects, which were partially deteriorated due to several factors, e.g., motion artifacts and data transmission problems, the overall accuracy of 94.19% and that of 81.22% were obtained for static activities and dynamic activities, respectively. In addition, the effects of the signal quality from these human subjects on false alarms were investigated. When false alarms occurring in signal segments with low quality were excluded, the number of false detections reduced from 14.17% to 8.65%. When evaluated on signals generated by the patient simulator, which included several types of premature ventricular contraction without artifacts from body movements, a high classification accuracy was also observed. [ABSTRACT FROM AUTHOR]

A sample entropy (SampEn)-based emotion recognition approach was presented. The SampEn results of notable EEG channels screened by K-S test were fed to the support vector machine (SVM)-weight classifier for training, after which it was applied to two emotion recognition tasks. One is to distinguish positive and negative emotion with high arousal and the other genitive emotion with different arousal status. Results showed that channels related to emotions were mostly located on the prefrontal region, i.e., F3, CP5, FP2, FZ, and FC2. And they were applied to form the input vectors of SVM-weight classifier. The accuracies of the present algorithm for the two tasks were 80.43% and 79.11%, respectively indicated by the leave-one-person-out validation procedure, demonstrating that the present algorithm had a reasonable generalization capability. [ABSTRACT FROM AUTHOR]

Osteosarcoma is the most common malignant bone tumor among children and adolescents. In this study, image texture analysis was made to extract texture features from bone CR images to evaluate the recognition rate of osteosarcoma. To obtain the optimal set of features, Sym4 and Db4 wavelet transforms and gray-level co-occurrence matrices were applied to the image, with statistical methods being used to maximize the feature selection. To evaluate the performance of these methods, a support vector machine algorithm was used. The experimental results demonstrated that the Sym4 wavelet had a higher classification accuracy (93.44%) than the Db4 wavelet with respect to osteosarcoma occurrence in the epiphysis, whereas the Db4 wavelet had a higher classification accuracy (96.25%) for osteosarcoma occurrence in the diaphysis. Results including accuracy, sensitivity, specificity and ROC curves obtained using the wavelets were all higher than those obtained using the features derived from the GLCM method. It is concluded that, a set of texture features can be extracted from the wavelets and used in computer-aided osteosarcoma diagnosis systems. In addition, this study also confirms that multi-resolution analysis is a useful tool for texture feature extraction during bone CR image processing. [ABSTRACT FROM AUTHOR]

Automated human larynx carcinoma (HEp-2) cell classification is critical for medical diagnosis. In this paper, we propose a sparse coding-based unsupervised transfer learning method for HEp-2 cell classification. First, the low level image feature is extracted for visual representation. Second, a sparse coding scheme with the Elastic Net penalized convex objective function is proposed for unsupervised feature learning. At last, a Support Vector Machine classifier is utilized for model learning and predicting. To our knowledge, this work is the first to transfer the human-crafted visual feature, sensitive to the variation of appearance and shape during cell movement, to the high level representation which directly denotes the correlation of one sample and the bases in the learnt dictionary. Therefore, the proposed method can overcome the difficulty in discriminative feature formulation for different kinds of cells with irregular and changing visual patterns. Large scale comparison experiments will be conducted to show the superiority of this method. [ABSTRACT FROM AUTHOR]

In this paper, we propose a novel method for the detection of microcalcifications using mathematical morphology and a support vector machine (SVM). First, the contrast in the original mammogram was improved by gamma correction and two carefully designed structural elements were used to enhance any microcalcifications. Next, the potential regions were extracted using our proposed dual-threshold technique. Finally, a SVM classifier was used to reduce the number of false positives. The performance of the proposed method was evaluated using the MIAS database. The experimental results demonstrated the efficiency and effectiveness of our method. [ABSTRACT FROM AUTHOR]