M. R. Homaeinezhad, A. Ghaffari, H. Najjaran Toosi, M. Tahmasebi, M. M. Daevaeiha,
Volume 7, Issue 1 (3-2011)
Abstract
In this study, a new long-duration holter electrocardiogram (ECG) major events detection-delineation algorithm is described which operates based on the false-alarm error bounded segmentation of a decision statistic with simple mathematical origin. To meet this end, first three-lead holter data is pre-processed by implementation of an appropriate bandpass finite-duration impulse response (FIR) filter and also by calculation of the Euclidean norm between corresponding samples of three leads. Then, a trous discrete wavelet transform (DWT) is applied to the resulted norm and an unscented synthetic measure is calculated between some obtained dyadic scales to magnify the effects of low-power waves such as P or T-waves during occurrence of arrhythmia(s). Afterwards, a uniform length window is slid sample to sample on the synthetic scale and in each slid, six features namely as summation of the nonlinearly amplified Hilbert transform, summation of absolute first order differentiation, summation of absolute second order differentiation, curve length, area and variance of the excerpted segment are calculated. Then all feature trends are normalized and superimposed to yield the newly defined multiple-order derivative wavelet based measure (MDWM) for the detection and delineation of ECG events. In the next step, a α-level Neyman-Pearson classifier (which is a false-alarm probability-FAP controlled tester) is implemented to detect and delineate QRS complexes. To show advantages of the presented method, it is applied to MIT-BIH Arrhythmia Database, QT Database, and T-Wave Alternans Database and as a result, the average values of sensitivity and positive predictivity Se = 99.96% and P+ = 99.96% are obtained for the detection of QRS complexes, with the average maximum delineation error of 5.7 msec, 3.8 msec and 6.1 msec for P-wave, QRS complex and T-wave, respectively showing marginal improvement of detection-delineation performance. In the next step, the proposed method is applied to DAY hospital high resolution holter data (more than 1,500,000 beats including Bundle Branch Blocks-BBB, Premature Ventricular Complex-PVC and Premature Atrial Complex-PAC) and average values of Se=99.98% and P+=99.97% are obtained for QRS detection. In summary, marginal performance improvement of ECG events detection-delineation process in a widespread values of signal to noise ratio (SNR), reliable robustness against strong noise, artifacts and probable severe arrhythmia(s) of high resolution holter data and the processing speed 163,000 samples/sec can be mentioned as important merits and capabilities of the proposed algorithm.
M. R. Homaeinezhad, E. Tavakkoli, A. Afshar, A. Atyabi, A. Ghaffari,
Volume 7, Issue 2 (6-2011)
Abstract
The paper addresses a new QRS complex geometrical feature extraction technique as well as its application for electrocardiogram (ECG) supervised hybrid (fusion) beat-type classification. To this end, after detection and delineation of the major events of ECG signal via a robust algorithm, each QRS region and also its corresponding discrete wavelet transform (DWT) are supposed as virtual images and each of them is divided into eight polar sectors. Then, the curve length of each excerpted segment is calculated and is used as the element of the feature space. To increase the robustness of the proposed classification algorithm versus noise, artifacts and arrhythmic outliers, a fusion structure consisting of three Multi Layer Perceptron-Back Propagation (MLP-BP) neural networks with different topologies and one Adaptive Network Fuzzy Inference System (ANFIS) were designed and implemented. To show the merit of the new proposed algorithm, it was applied to all MIT-BIH Arrhythmia Database records and the discrimination power of the classifier in isolation of different beat types of each record was assessed and as the result, the average accuracy value Acc=98.27% was obtained. Also, the proposed method was applied to 8 number of arrhythmias (Normal, LBBB, RBBB, PVC, APB, VE, PB, VF) belonging to 19 number of the aforementioned database and the average value of Acc=98.08% was achieved. To evaluate performance quality of the new proposed hybrid learning machine, the obtained results were compared with similar peer-reviewed studies in this area.
H. Rezghian Moghadam, H. R. Kobravi, M. Homam,
Volume 14, Issue 2 (6-2018)
Abstract
The tremor injury is one of the common symptoms of Parkinson's disease. The patients suffering from Parkinson's disease have difficulty in controlling their movements owing to tremor. The intensity of the disease can be determined through specifying the range of intensity values of involuntary tremor in Parkinson patients. The level of disease in patients is determined through an empirical range of 0-5. In the early stages of Parkinson, resting tremor can be very mild and intermittent. So, diagnosing the levels of disease is difficult but important since it has only medication therapy. The aim of this study is to quantify the intensity of tremor by the analysis of electromyogram signal. The solution proposed in this paper is to employ a polynomial function model to estimate the Unified Parkinson's Disease Rating Scale (UPDRS) value. The algorithm of Fast Orthogonal Search (FOS), which is based on identification of orthogonal basic functions, was utilized for model identification. In fact, some linear and nonlinear features extracted from wrist surface electromyogram signal were considered as the input of the model identified by FOS, and the model output was the UPDRS value. In this research, the proposed model was designed based on two different structures which have been called the single structure and parallel structure. The efficiency of designed models with different structures was evaluated. The evaluation results using K-fold cross validation approach showed that the proposed model with a parallel structure could determine the tremor severity of the Parkinson's disease with accuracy of 99.25% ±0.41, sensitivity of 97.17% ±1.9 and specificity of 99.72% ±0.18.
S. Mirzakuchaki, Z. Paydar,
Volume 14, Issue 4 (12-2018)
Abstract
In this study a method has been introduced to map the features extracted from the recorded electromyogram signals from the forearm and the force generated by the fingers. In order to simultaneously record of sEMG signals and the force produced by fingers, 9 requested movements of fingers conducted by 10 healthy people. Estimation was done for 6 degrees of freedom (DoF) and generalized regression neural network (GRNN) was selected for system training. The optimal parameters, including the length of the time windows, the parameters of the neural network, and the characteristics of the sEMG signal were calculated to improve the performance of the estimate. The performance was obtained based on R2 criterion. The Total value of R2 for 6 DoF was 92.8±5.2% that obtained by greedy looking system parameters in all the subjects. The result shows that proposed method can be significant in simultaneous myoelectric control.
G. Das, R. Panda, L. Samantaray , S. Agrawal,
Volume 18, Issue 2 (6-2022)
Abstract
Multilevel optimal threshold selection is important and comprehensively used in the area of image processing. Mostly, entropic information-based threshold selection techniques are used. These methods make use of the entropy of the distribution of the grey levels of an image. However, entropy functions largely depend on spatial distribution of the image. This makes the methods inefficient when the distribution of the grey information of an image is not uniform. To solve this problem, a novel non-entropic method for multilevel optimal threshold selection is proposed. In this contribution, simple numbers (pixel counts), explicitly free from the spatial distribution, are used. A novel non-entropic objective function is proposed. It is used for multilevel threshold selection by maximizing the partition score using the adaptive equilibrium method. A new theoretical derivation for the fitness function is highlighted. The key to the achievement is the exploitation of the score among classes, reinforcing an improvised threshold selection process. Standard test images are considered for the experiment. The performances are compared with state-of-the-art entropic value-based methods used for multilevel threshold assortment and are found better. It is revealed that the results obtained using the suggested technique are encouraging both qualitatively and quantitatively. The newly proposed method would be very useful for solving different real-world engineering optimization problems.
Srinivas Babu N, Shashikiran S, M Jayanthi, Rajani N, K M Palaniswamy, M R Kushalatha,
Volume 20, Issue 0 (12-2024)
Abstract
Tuberculosis (TB) is a dangerous disease caused by mycobacterium leads to mortality. Early detection and identification of tuberculosis is crucial for managing tuberculosis infections. Recent technological improvements use a machine learning-based SVM and Modified CNN to identify specific diseases more accurately, as demonstrated in this research. The modified CNN's improved feature extraction and classification accuracy are maintained throughout construction. To obtain good performance a TBX11K publicly accessible dataset is used it consists of 11000 images of which 4600 chest x-ray (CXR) images are considered in this research, and the suggested model is verified. This approach significantly increases the accuracy of categorizing TB symptoms. The PCA in this system locates the elements and extracts a large amount of variance technique applied to the full chest radiograph for pulmonary tuberculosis identification accuracy using SVM is 93.14% and modified CNN 96.72% respectively. When it comes to helping radiologists diagnose patients and public health professionals screen for tuberculosis in places where the disease is endemic, the proposed system SVM and modified CNN perform better than existing methods.
Abolfazl Karimiyan Abdar, Ali Esteki, Mohsen Sheikh Hassani,
Volume 20, Issue 1 (3-2024)
Abstract
The impact of cognitive tasks on human movement is of practical significance; we hereby aim to demonstrate that a significant relationship exists between the dual task’s cognitive demand and the disruption caused in hand movement, with the hope to extend this experiment to subjects with disorders (MS, CP, stroke patients) in future studies. By doing so, we hope to be able to develop a metric for evaluating their disease levels using our method and minimize clinical interventions. While previous research has predominantly focused on lower body activities, this study explores the effect of dual tasks on hand movements in healthy individuals.
A simulated finger-to-nose test combined with a standard reverse counting task, featuring four difficulty levels, was conducted. Utilizing SVM and decision tree classifiers, we analyzed various features to discern the impact of cognitive tasks on hand movements, including completed cycles and idle time at markers. Our findings reveal that features such as entropy and kurtosis effectively distinguish between task difficulty levels and hand movement disruption. The classifiers achieved accuracies of 70% and 74% for decision tree and SVM, respectively. We hope extending this research to diseased subjects could potentially provide a more accurate assessment of disease severity through the measurement of hand movements during cognitive tasks, offering a non-clinical alternative for disease evaluation.
Aws Alazawi, Huda Jameel, Mohammed Mohsen,
Volume 20, Issue 2 (6-2024)
Abstract
This study explores the use of distortion product otoacoustic emission (DPOAE) as a hearing screening modality for newborns and adults with hearing impairment. The goal is to improve cochlear response by developing digital filter characteristics to make it consistent for specialists to make accurate diagnoses. To accomplish this, the proposed system consists of a DPOAE ER-10C as stimulation and cochlear response probe, a digital signal processor, an oscilloscope, PC, and audio cables. Real-time distortion product frequency components were extracted using a signal processor of TMS320C6713. To validate the system, a senior medical physicist at Baghdad Medical City in Iraq conducted a study with five hearing-normal volunteers ages 38 and 55 at the center for hearing and communication. The results showed an ability to extract distortion product components in real-time implementation, with the superiority of shape parameters greater than 0.5. In addition, the quantization of filter coefficients was compared for both floating-point arithmetic and fixed-point arithmetic. Noisy environment-based noise reduction techniques have to be investigated by considering the implementation of robust digital signal processing techniques. Finally, the proposed system would contribute to advancements in hearing screening and treatment for those with hearing impairment.
Biswapriyo Sen, Maharishi Kashyap, Jitendra Singh Tamang, Sital Sharma, Rijhi Dey,
Volume 20, Issue 2 (6-2024)
Abstract
Cardiovascular arrhythmia is indeed one of the most prevalent cardiac issues globally. In this paper, the primary objective was to develop and evaluate an automated classification system. This system utilizes a comprehensive database of electro- cardiogram (ECG) data, with a particular focus on improving the detection of minority arrhythmia classes.
In this study, the focus was on investigating the performance of three different supervised machine learning models in the context of arrhythmia detection. These models included Support Vector Machine (SVM), Logistic Regression (LR) and Random Forest (RF). An analysis was conducted using real inter-patient electrocardiogram (ECG) records, which is a more realistic scenario in a clinical environment where ECG data comes from various patients.
The study evaluated the models’ performances based on four important metrics: accuracy, precision, recall, and f1-score. After thorough experimentation, the results highlighted that the Random Forest (RF) classifier outperformed the other methods in all of the metrics used in the experiments. This classifier achieved an impressive accuracy of 0.94, indicating its effectiveness in accurately detecting arrhythmia in diverse ECG signals collected from different patients.
Sandra D’souza, Niranjan Reddy S, Saikonda Krishna Tarun, Sohan P, Aneesha Acharya K,
Volume 20, Issue 4 (11-2024)
Abstract
The incidence of heart-related illnesses is on the rise worldwide. Heart diseases are primarily caused by a multitude of parameters, including high blood pressure, diabetes, and excessive cholesterol, which are controlled by poor dietary and lifestyle choices. The growth in cardiovascular diseases (CVD) is mostly due to several other behaviors, such as smoking, drinking, and sleeplessness. In the research, machine learning-based prediction methods work on the audio recordings of heartbeats known as phonocardiograms (PCG) to develop an algorithm that differentiates a normal healthy heart from an abnormal heart based on the heart sounds. The data set consists of 831 normal and 260 abnormal data, and the duration of each sample is 5 seconds. Features extracted from the data are up-sampled and applied to the logistic regression and random forest classification models. The developed models record a classification accuracy of 71% for logistic regression and 94% for the random forest model. Further, artificial neural networks (ANN) and Deep learning networks have been trained to improve performance and demonstrated an accuracy of 94.5%.
