A model to identify high crash road segments with the dynamic segmentation method.

Currently, high social and economic costs in addition to physical and mental consequences put road safety among most important issues. This paper aims at presenting a novel approach, capable of identifying the location as well as the length of high crash road segments. It focuses on the location of accidents occurred along the road and their effective regions. In other words, due to applicability and budget limitations in improving safety of road segments, it is not possible to recognize all high crash road segments. Therefore, it is of utmost importance to identify high crash road segments and their real length to be able to prioritize the safety improvement in roads. In this paper, after evaluating deficiencies of the current road segmentation models, different kinds of errors caused by these methods are addressed. One of the main deficiencies of these models is that they can not identify the length of high crash road segments. In this paper, identifying the length of high crash road segments (corresponding to the arrangement of accidents along the road) is achieved by converting accident data to the road response signal of through traffic with a dynamic model based on the wavelet theory. The significant advantage of the presented method is multi-scale segmentation. In other words, this model identifies high crash road segments with different lengths and also it can recognize small segments within long segments. Applying the presented model into a real case for identifying 10-20 percent of high crash road segment showed an improvement of 25-38 percent in relative to the existing methods.

Predicting termination of paroxysmal atrial fibrillation using empirical mode decomposition of the atrial activity and statistical features of the heart rate variability.

This paper presents an algorithm for predicting termination of paroxysmal atrial fibrillation (AF) attacks using features extracted from the atrial activity (AA) and heart rate variability (HRV) signals. First, AA signal was decomposed into a set of intrinsic mode functions (IMFs) using empirical mode decomposition method. Then, power spectrums of the AA and its IMFs (second, third, and forth components) were obtained, and the peak frequency of the power spectral densities were extracted. These features were complemented with three additional features consisting of mean, skewness, and kurtosis of the HRV signal. These seven features were then reduced to only two features by the generalized discriminant analysis technique. This not only reduces the number of the input features but also increases the classification accuracy by selecting most discriminating features. Finally, a linear classifier was used to classify AF episodes from AF termination database. This database consists of three types of AF episodes: N type (non-terminated AF episode), S type (terminated 1 min after the end of the record), and T type (terminated immediately after the end of the record). The obtained sensitivity, specificity, positive predictivity, and negative predictivity were 94, 97, 92, and 96 %, respectively. The important advantage of the proposed method comparing to the other existing approaches is that our algorithm can simultaneously discriminate three types of AF episodes with high accuracy.

Prediction of paroxysmal atrial fibrillation based on non-linear analysis and spectrum and bispectrum features of the heart rate variability signal.

In this paper, an effective paroxysmal atrial fibrillation (PAF) prediction algorithm is presented, which is based on analysis of the heart rate variability (HRV) signal. The proposed method consists of a preprocessing step for QRS detection and HRV signal extraction. In the next step, several features which can be used as markers for the prediction of PAF are extracted from the HRV signal. These features consist of spectrum features, bispectrum features, and non-linear features including sample entropy and Poincaré plot-extracted features. The spectrum features are able to discriminate the sympathetic and parasympathetic contents of the HRV signal, which are affected before PAF attacks. The bispectrum features are used in order to reveal information not presented on the spectral domain, and to detect quadratic phase coupled harmonics arising from non-linearities of the HRV signal. Moreover, the non-linear analysis can map the heart rate irregularities in the feature space and it leads to better understanding of the system dynamics before PAF attacks. In the final step, a support vector machine (SVM)-based classifier has been used for PAF prediction. The performance of the proposed method in prediction of PAF episodes was evaluated using the Atrial Fibrillation Prediction Database (AFPDB). The obtained sensitivity, specificity, and positive predictivity were 96.30%, 93.10%, and 92.86%, respectively. The proposed methodology presents better results than the other existing approaches. The other important advantage of the proposed method when compared to the other approaches is that we do not need the both records of a subject to specify which episode preceding PAF events.

Modeling of heart systolic murmurs based on multivariate matching pursuit for diagnosis of valvular disorders.

Heart murmurs are pathological sounds produced by turbulent blood flow due to certain cardiac defects such as valves disorders. Detection of murmurs via auscultation is a task that depends on the proficiency of physician. There are many cases in which the accuracy of detection is questionable. The purpose of this study is development of a new mathematical model of systolic murmurs to extract their crucial features for identifying the heart diseases. A high resolution algorithm, multivariate matching pursuit, was used to model the murmurs by decomposing them into a series of parametric time-frequency atoms. Then, a novel model-based feature extraction method which uses the model parameters was performed to identify the cardiac sound signals. The proposed framework was applied to a database of 70 heart sound signals containing 35 normal and 35 abnormal samples. We achieved 92.5% accuracy in distinguishing subjects with valvular diseases using a MLP classifier, as compared to the matching pursuit-based features with an accuracy of 77.5%.

Prediction of paroxysmal atrial fibrillation using recurrence plot-based features of the RR-interval signal.

Atrial fibrillation (AF) is the most common cardiac arrhythmia and increases the risk of stroke. Predicting the onset of paroxysmal AF (PAF), based on noninvasive techniques, is clinically important and can be invaluable in order to avoid useless therapeutic intervention and to minimize risks for the patients. In this paper, we propose an effective PAF predictor which is based on the analysis of the RR-interval signal. This method consists of three steps: preprocessing, feature extraction and classification. In the first step, the QRS complexes are detected from the electrocardiogram (ECG) signal and then the RR-interval signal is extracted. In the next step, the recurrence plot (RP) of the RR-interval signal is obtained and five statistically significant features are extracted to characterize the basic patterns of the RP. These features consist of the recurrence rate, length of longest diagonal segments (L(max )), average length of the diagonal lines (L(mean)), entropy, and trapping time. Recurrence quantification analysis can reveal subtle aspects of dynamics not easily appreciated by other methods and exhibits characteristic patterns which are caused by the typical dynamical behavior. In the final step, a support vector machine (SVM)-based classifier is used for PAF prediction. The performance of the proposed method in prediction of PAF episodes was evaluated using the Atrial Fibrillation Prediction Database (AFPDB) which consists of both 30 min ECG recordings that end just prior to the onset of PAF and segments at least 45 min distant from any PAF events. The obtained sensitivity, specificity, positive predictivity and negative predictivity were 97%, 100%, 100%, and 96%, respectively. The proposed methodology presents better results than other existing approaches.

Structures of the recurrence plot of heart rate variability signal as a tool for predicting the onset of paroxysmal atrial fibrillation.

This paper aims to propose an effective paroxysmal atrial fibrillation (PAF) predictor which is based on the analysis of the heart rate variability (HRV) signal. Predicting the onset of PAF, based on non-invasive techniques, is clinically important and can be invaluable in order to avoid useless therapeutic interventions and to minimize the risks for the patients. This method consists of four steps: Preprocessing, feature extraction, feature reduction, and classification. In the first step, the QRS complexes are detected from the electrocardiogram (ECG) signal and then the HRV signal is extracted. In the next step, the recurrence plot (RP) of HRV signal is obtained and six features are extracted to characterize the basic patterns of the RP. These features consist of length of longest diagonal segments, average length of the diagonal lines, entropy, trapping time, length of longest vertical line, and recurrence trend. In the third step, these features are reduced to three features by the linear discriminant analysis (LDA) technique. Using LDA not only reduces the number of the input features, but also increases the classification accuracy by selecting the most discriminating features. Finally, a support vector machine-based classifier is used to classify the HRV signals. The performance of the proposed method in prediction of PAF episodes was evaluated using the Atrial Fibrillation Prediction Database which consists of both 30-minutes ECG recordings end just prior to the onset of PAF and segments at least 45 min distant from any PAF events. The obtained sensitivity, specificity, and positive predictivity were 96.55%, 100%, and 100%, respectively.

A transformation based method to design ultrasound array.

In medical ultrasound imaging, the desired lateral field distribution at each focal distance can be obtained by optimal apodization. On the other hand, the lateral field is a function of focal distance. Hence, finding the optimal apodization is a very arduous process. To overcome this, we have introduced a suboptimal method by which optimal apodization can be calculated in any distance through a nonlinear transformation by the knowledge of the optimal one at a distance. This transformation is established on a fact that the lateral field distribution at focal distance can be expressed as the Fourier transform of a nonlinear function of the aperture weighting, instead of direct expression as the Fourier transform of the above. We have applied this method to map the apodization which obtains the desired beam pattern into the apodization which maintains the same properties on the lateral field distribution. For example, applying this method on a 50-elements lambda/2 spaced linear array with length D has resulted in apodization that is optimal at distances D or D/2 by precision better than 9%. This method is useful especially in optimization problems, having no explicit constraint on the main lobe width, such as minimizing the sidelobe levels or minimizing main lobe width constrained to a predetermined value of sidelobe level. However, as the results show, this technique provides acceptable results in other cases.

Content-based medical image classification using a new hierarchical merging scheme.

Automatic medical image classification is a technique for assigning a medical image to a class among a number of image categories. Due to computational complexity, it is an important task in the content-based image retrieval (CBIR). In this paper, we propose a hierarchical medical image classification method including two levels using a perfect set of various shape and texture features. Furthermore, a tessellation-based spectral feature as well as a directional histogram has been proposed. In each level of the hierarchical classifier, using a new merging scheme and multilayer perceptron (MLP) classifiers (merging-based classification), homogenous (semantic) classes are created from overlapping classes in the database. The proposed merging scheme employs three measures to detect the overlapping classes: accuracy, miss-classified ratio, and dissimilarity. The first two measures realize a supervised classification method and the last one realizes an unsupervised clustering technique. In each level, the merging-based classification is applied to a merged class of the previous level and splits it to several classes. This procedure is progressive to achieve more classes. The proposed algorithm is evaluated on a database consisting of 9100 medical X-ray images of 40 classes. It provides accuracy rate of 90.83% on 25 merged classes in the first level. If the correct class is considered within the best three matches, this value will increase to 97.9%.

A feedback retina model for improving medical images fusion.

Image fusion has become a powerful technique for increasing the interpretation quality of images in medical applications. The diagnostic potential of brain positron emission tomography (PET) imaging is limited by low spatial resolution. For solving this problem, the high-frequency part of the MRI, which would be unrecoverable by the set PET acquisition system, is extracted and added to the PET image. The procedure has the potential of increasing the diagnostic value of a PET image. This paper presents a feedback retina model technique to reduce the spectral distortion and preserve high spatial resolution. Visual and statistical analyses show that the proposed feedback retina model significantly improves the fusion quality compared to non-feedback retina model.

Detection of atrial fibrillation episodes using SVM.

This paper explains an atrial fibrillation (AF) detection algorithm, which consists of a linear discriminant analysis (LDA) based feature reduction scheme and a support vector machine (SVM) based classifier. Initially nine features were extracted from the input episodes each containing 32 RR intervals by linear and nonlinear methods. Next, to improve the learning efficiency of the classifier and to reduce the learning time, these features are reduced to 4 features by LDA. The performance of the proposed method in discriminating AF episodes was evaluated using MIT-BIH arrhythmia database. The obtained sensitivity, specificity and positive predictivity were 99.07%, 100% and 100%, respectively.

A hybrid algorithm for medical image registration.

Registration is a process to align different acquired images of the same subject. A major problem in this field is to register images captured by different imaging systems. These images have different gray values so simple methods like correlation are not applicable. In this paper automated registration of CT and MR head images is studied. It is assumed that images are only of relative translation and rotation. The proposed method includes two stages. First, feature extraction of CT and MR images, (should be extractable and constant in both imaging systems). Second, feature alignment which is moment-based. The results confirm the accuracy and robustness of the proposed method.