M.Sc. Engg.

Browse

Now showing 1 - 13 of 13

Model Based ECG Denoising Using Discrete Bionic WaveletTransform
(Khulna University of Engineering & Technology (KUET), Khulna, Bangladesh., 2011-12) Awal, Md. Abdul; Ahmad, Prof. Dr. Mohiuddin.
Electrocardiogram (ECG) is a measurement of bio-electric potential produced by rhythmical cardiac activities, contraction and relaxation of the cardiac muscle produced at sinoatrial (SA) node. This electric potential associated with the cardiac cycle can be detected at the surface of the body, amplified and recorded as a time record of each cardiac cycle. Different cardiac function such as heart rate, abnormality of rhythm can easily be identified by ECG and it is a low cost tools in the medical diagnostic system. Therefore, ECG signal modeling and processing is one of the most significant topic in biomedical signal analysis. Most of the ECG models are complex and their computational time is high. In this research, a Gaussian wave-based model is proposed which can simulate ECG wave as well as its P, Q, R, S and T components individually. In addition, dynamically shifting baseline of the model reduces the preprocessing of ECG signal. The coefficient of the model is calculated by nonlinear least square technique using Gauss-Newton algorithm.The model fits well with real ECG by Normalized Root Mean Square error (NRMSE) of 0.0034 at the normal condition. Further analyses have been performed to evaluate the models ability of representing the different cardiac Dysrhythmias like atrial fibrillation,brachycardia and tachycardia successfully. For better model fitting denoised ECG plays a significant role. Bionic wavelet Transform (BWT) is based on auditory model but it is not efficient for ECG signal processing since ECG is generated from the heart. So for denoising ECG, a new adaptive wavelet transform is developed based on heart- arterial interaction model. Adaptability is adjusted instantaneous amplitude of the signal and its first-order difference. The automatically adjusted resolution is achieved by introducing the active control mechanism of the cardiac system into the wavelet transform. It is very hard to know what entropy function used in the bio-system. This is the problems of other transforms. But, the discrete BWT uses active control mechanism in the cardiac system to adjust the wavelet function rather than entropy function as criterion. Moreover, due to various oscillating behavior of different types of ECG signal constant Quality factor (Q) of wavelet is not as effective as variable Q. is changed with the instantaneous value of a signal and it will make BWT more adaptive compared to Tunable The variable Q Q wavelet transform (TQWT). As in variable Q-wavelet transform like DBWT which is he discrete version of BWT, is changed with the instantaneous value of a signal and its first order difference instead of Q-factor is tuned to a fixed value in TQWT. In addition, our proposed modified S-median thresholding technique has an adjustable factor and introduced in the system for better performance. In order to compare DBWT with other wavelet transform, experiments on traditional WT, multi- Q adaptive BWT, TQWT were conducted on both constructed signals and real ECG signals. The results show that novel DBWT performs better than these three wavelet transforms, and is appropriate for cardiac signal processing, especially over noisy environment.
Improvement of Medical Imaging Equipment Management System of Bangladesh
(Khulna University of Engineering & Technology (KUET), Khulna, Bangladesh, 2012-06) Hossain, Md. Anwar; Ahmad, Prof. Dr. Mohiuddin
This thesis deals with the development of a standard Medical Imaging Equipment Management System (MIEMS) in the public hospitals of Bangladesh. Medical imaging equipment management has become an important component of health services delivery in Bangladesh. One of the most serious issues of Hospitals, Medical Center, Clinics and the health Systems is the threat in diagnosing diseases due to improper medical imaging equipment management system in Bangladesh. Presently a proper equipment management set-up does not exist in public hospitals and the overall diagnostic equipment condition in various departments is very poor. To find the facts and reasons the author visited fifteen medical imaging equipment departments in public hospitals of Bangladesh. From the data it is seen that a big percentage of the required management steps were found to be lacking. Based on the findings of the studies and with a view to improving the existing MIEMS, it is proposed to improve the medical imaging equipment management system for public hospitals in Bangladesh. The major findings are insufficient Biomedical Engineering Team, negligible in-house and central MIEMS, weak bridge between users and biomedical engineering team. The Ministry of Health and Family welfare of Bangladesh Govt., maintains a separate Medical Imaging Equipment Management System, which is not consistent with this plan. The MIEMS is designed to assure selection of appropriate medical imaging equipment to support the medical care processes of the nation and its ambulatory care facilities. The program is designed to assure effective preparation of staff responsible for the use or maintenance of the equipment. Finally, the program is designed to assure continuous availability of safe, effective equipment through a planned maintenance program, timely repair and evaluation of all events that could have an adverse impact on safety of patients or staff. The Clinical Biomedical Engineering Department staff has primary responsibility to maintain and repair all medical equipment including equipment under contract maintenance agreement by Vendors. All staff who uses medical equipment is required to learn and implement various general and specific procedures to ensure safe and reliable use of medical equipment. In addition, it will provide a proper guideline for all patient care staff, failures and user errors associated with the use of medical equipment. Regular maintenance and evaluation are necessary to assure that particular equipment delivers the expected performance within specified para
Clench Strength Prediction for Prosthesis Hand Using Surface Electromyogram
(Khulna University of Engineering & Technology (KUET), Khulna, Bangladesh., 2012-09) Mostafa, Sheikh Shanawaz; Ahmad, Prof. Dr. Mohiuddin
Hand is the main environmental manipulator for humans. Therefore, any accidental lost of hand does a great harm in amputee life. Only choice left that he or she is using prosthesis, which is dated back to ancient time. Due to their usefulness prosthesis is also used in modern days. Now-a-day robotic rehabilitation opens a new era in the field of prosthesis. They are more capable of doing things compared to their ancestors. However, to duplicate the complex work of hand, the robotic hand is needed to complete two basic operations first one is predict the angular displacement and another one is the estimation of force. Estimation of force is essential for instance when drinking from a glass, one needs to apply sufficient grip force to prevent the object from slipping Out of the hand. In addition, one needs to control the total torque exerted on the glass such that the glass remains vertical. Usually, the requirements for grip force stabilization allow for some laxity, while the requirements for total torque production are highly specified. The grip force needs only to be larger than the slip threshold and smaller than the force that would break the object. Different researchers used different approach to solve this problem like body-powered prostheses, force-sensitive resistors. Nevertheless, when it comes to rehabilitation of amputee they are not up to the mark in practical field.Using of Electromyography (EMG) in the robotic rehabilitation shows some hope. Muscle tissue conducts electrical potentials similar to the way nerves do and the name given to these electrical signals is the muscle action potential. A muscle is composed of bundles of specialized cells capable of contraction and relaxation. The primary function of these specialized cells is to generate forces, movements and the ability to communicate such as speech or writing or other modes of expression. The skeletal muscle tissue is attached to the bone and its contraction is responsible for supporting and moving the skeleton. The contraction of skeletal muscle is initiated by impulses in the neurons to the muscle and is usually under voluntary control. Skeletal muscle fibers are well-supplied with neurons for its contraction. This depolarization, accompanied by a movement of ions, generates an electric field near each muscle fiber. An EMG signal is the train of Motor Unit Action Potential (MUAP) showing the muscle response to neural stimulation. There is a clear relationship between force and EMG. The higher the muscle force, the higher EMG level is developed. Clench force estimation is highly desirable in the field of prosthesis hand. It is one of the most used postures among different types of postures. In this thesis, the author proposes to estimate the clench force using two types of Surface Electromyography (SEMG): rectified SEMG and integrated SEMG. An Artificial Neural Network (ANN) is used to estimate the force from SEMG. For weight adjustment of the estimator Levenberg-Marquardt (L-M) back propagation algorithm is used. The proposed network is trained and tested using SEMG recorded from five subjects. The estimation result clearly shows that integrated SEMG performed 3.53 times better than rectified SEMG in the case of cross correlation coefficient. So integrated SEMG is recommended for clench force estimation. A neural number based experiment is also done to find the optimal number of neurons in hidden layer, which are five neurons as previously described neural network. In addition to that, these result also compared with Support Vector Machine (S\TM). Same result is found for integrated SEMG. In the case of rectified SEMG ANN is more suitable than SVM.
Cardiac Function Evaluation of Healthy Young Adults by Consuming Energy Drinks
(Khulna University of Engineering & Technology (KUET), Khulna, Bangladesh., 2014-03) Uddin, Md. Bashir; Ahmad, Prof. Dr. Mohiuddin.
Energy drink (ED) is a beverage that mainly contains some form of legal stimulants which are supposed to give consumers a short term boost in energy. It is a type of beverage containing caffeine as main stimulant, which is marketed as providing both mental and physical stimulation. With ED becoming a worldwide phenomenon, the short- and longterm effects of these beverages must be evaluated more closely in order to fully comprehend the psychological impact of these products. The market and degree of consumption of ED is increasing every year, but only few have global knowledge of their ingredients and actual physiological and psychological effects. Although ED have been sold worldwide for more than a decade, only a few published studies have examined their effects on health and well-being. The effects of ED consumption on physiological and psychological are more or less investigated but the effects of ED consumption on heart activity or cardiac function aren't well studied. The main aim of this study is to evaluate cardiac function of healthy young adults by consuming ED analyzing different cardiac signals. The modeling of cardiac signals with consuming ED is the partial objective of this study.The electrical activity of the heart over a period of time can be represented by the properties of Electrocardiogram (ECG) wave. Photo Plethysmogram (PPG) is the tactilearterial palpation of the heartbeat which is also known as pulse. ECG and PPG signals are used in this study as cardiac signals which are recorded by biopac accessories from healthy human subjects. ECG is recorded using electrode lead set connected to MP36 data acquisition unit. PPG (pulse) is also recorded using pulse transducer connected to the same MP36 data acquisition unit. ECG and PPG are recorded at both before and after the consumption of ED. The Laser Doppler Flowmetry (LDF) recording is also done for results verification for certain subjects. The consumption of ED affects heart activity that is determined in this study using electrocardiographic and photo plethysmographic parameters. The ECG parameters analysis show significant reduction in their corresponding amplitude as well as heart rate due to having ED. The amplitude of R wave of ECG increases little bit that may give short-term boost of energy. A notable decrement in peak to peak amplitude of PPG as well as pulse rate is observed due to having ED. The spectrum or frequency components for ECG as well as PPG signal decreases with a significant rate from the instant of having ED. That is, the spectrum parameters of cardiac activity decrease due to the consumption of energy drinks. The spectrum analysis of LDF signal also results similar type of decrement in their spectrum parameters for same type of energy drinks consumption. This LDF signal analysis validates our main experimental results. These results reflect adverse impacts of energy drinks consumption on cardiac activity. ECG is the main cardiac signal which represents complete cardiac function which is the measurement of bio-electric potential produced by rhythmical cardiac activities (contraction and relaxation) of the cardiac muscle. Different cardiac functions can be easily identified by ECG that's why ECG modeling is the most important with the consumption of energy drinks. Different techniques have been developed in the past for modeling of ECG. An ECG model is proposed in this study using peak amplitude based Gaussian function with some modifications in both before and after having ED. This model is best suited with practical ECG parameters. Using this ECG model it is possible to find out cardiac parameters which are needed for the evaluation of cardiac function with having ED. In this research, we have find out cardiac parameters using our ECG model and compared with real ECG parameters. The comparison results less error between real ECG and model ECG in the evaluation of cardiac function with consuming ED. Thus this ECG model is effective to evaluate cardiac function with the consumption of ED
Determination of Muscle Activity and Work-Done Through Emg Analysis During Human Movements in Salat
(Khulna University of Engineering & Technology (KUET), Khulna, Bangladesh., 2016-05) Khanam, Farzana; Ahmad, Prof. Dr. Mohiuddin.
In this thesis, human movements during Salat (the ritual prayer of Muslim) are taken in account to observe this muscle movement patterns as exercise. Since surface electromyography (sEMG) is a proven method to observe the muscle activity based on the generated action potential by the muscles, sEMG method is applied to measure the physical activities from different muscles of the different steps of Salat. From these measurements, it has been shown that muscle activity during Salat can perform better exercise than walking in treadmill based on muscle fatigue index. The results specify that during Salat, Biceps Brachii (BB) and Erector Spine (ES) produce improved EMG level for both male and female subjects in opposition to Treadmill exercise. Therefore, the aftermath of this work gives us a message that the person who performs Salat five times in a day is doing exercise of upper limb muscles especially the BB and ES muscles. Furthermore, a modified power spectrum analyzing method is developed. By this proposed method it is proved that mean frequency (MNF) based EMG power spectrum analyzing method is comparatively efficient method than the previously proposed to determine Salat associated muscle fatigue and indices. For finding the relation between work done and corresponding sEMG voltage a number of subjects (male and female) are studied with a known work done and their corresponding sEMG voltages are measured. As a result, some approximate mathematical functions are developed by appropriate curve fitting to relate the work done and corresponding EMG voltage. These mathematical functions are proposed for upper limb (Bicep Bracii) muscle and lower limb (Biceps Femoris and Medial Gastrocnemius) muscle. The estimated relationship between work done and sEMG voltage show minimum error when it actually follows second degree function. Based on the work, it is observed that the functions vary from muscle to muscle which means for upper limb and lower limb the relationships follow different mathematical functions. It is also found that the sEMG voltage generated from muscle undergoes saturation after increasing to a certain work done. At last it can be said so far that this work focuses on human movement in Salat to prove it as a fresh exercise (no muscle fati(Iue) mathematically.In addition with that, analytical techniques are developed to determine its muscle indices and fatigue, as well as mathematical relations are established between work done and corresponding sEMG voltage to observe the work done during Salat.
Detection of Neural Activity for Cerebrovascular Disease using MRI and EEG
(Khulna University of Engineering & Technology (KUET), Khulna, Bangladesh., 2016-08) Rahman, G. M. Mahmudur; Shahjahan, Prof. Dr. Md.
(CVD) such as stroke is the leading cause of long-term disability and third most common reason of death in the world. Due to the neuronal deficiency occurred by impaired blood flow to the brain, half of the stroke patients survive a severe cognitive impairment such as impaired speech, numbness in limbs, immobilization of limbs, visual impairment and many other observable symptoms. In the recent days, due to the gradual increase of hypertensive and diabetic patient, the risk of developing stroke is growing tremendously. The evaluation of actual neuronal status in stroke patient is thus very important to determine the indication of neurosurgical treatment. To diagnosis the type, source and location of stroke in the brain, generally, cerebral blood flow (CBF) test, neuroimaging, and electrical activity test are being used. Among all of these tests, electroencephalogram (EEG) is a useful tool for acute stroke detection and monitoring affected tissue owing to its relatively cheap and completely hazardless for quantitative and statistical analysis. Although a large number of studies have been performed on the ischemic stroke using EEG, no investigation has been carried out on the interplay between the infarcted cerebral hemisphere with other healthy part. Moreover, earlier studies are also limited only age-related changes in EEG activity or memory performance compared younger people between the ages of about 20 to 30 years with older participants between the ages of about 60 to 80 years. There is also no study at all for the child below the age of 20 years to show the change and compare the neuronal deterioration of the infarcted part of brain. This work focused on the correlation between the extent of infarction and the clinical effect to monitor the degree of hypo-activity of the affected part through the EEG analysis. An indication for the assessment of neuronal activity and the degree of severity of stroke patient has been proposed. A parallel study has also been carried out on healthy volunteers of under fifteen years to find the comparison of neural activity between different age group. From the analysis it is found that delta activities of EEG are highly unique of brain pathophysiology, and preservation of alpha and beta frequencies following stroke is evidently indicative of neuronal survival and a good prognosis. In order to assess brain vi pathophysiology in supratentorial brain lesion patients the delta/alpha ratio (DAR) and deltaplus- theta to alpha-plus-beta ratio (DTABR) have also been used. It is observed that the DAR and DTABR values of the left cerebral hemisphere of the patient are much higher than the right cerebral hemisphere. It is also higher in both cerebral hemispheres than the control. A threshold value of ~3.7 for DAR and ͂ 3.5 for DTABR has been obtained. It is observed that DAR and DTABR of left hemisphere of patient are greater than 3.7and 3.5 respectively for all the measurements indicating severe ischemic stroke in the fronto-temporal region of left hemisphere. The findings have been further confirmed by a neuroimaging technique such as magnetic resonance imaging (MRI).This study also show that DAR and DTABR of the healthier child is~1.It is also found that delta and DAR indices of the old age are two times more than the child indicating the diminishing of neuronal activity of old age is half of the child. These results could be important for stroke diagnosis, prognosis, re-habitation strategies, and proper neurosurgical treatment.
Study of Stimulation Effects on Different Bands of EEG Signal
(Khulna University of Engineering & Technology (KUET), Khulna, Bangladesh., 2016-09) Ghosh, Tarun Kanti; Ahmad, Prof. Dr. Mohiuddin
Brain–computer interfaces (BCIs) are the communication bridge between the human brain and a computer which may be implemented on the basis of steady-state evoked potentials (SSEPs). A brain–computer interface is a direct communication pathway between the brain and an external device. BCIs are often directed at assisting, augmenting, or repairing human cognitive or sensory-motor functions. The field of BCI research and development has since focused primarily on neuroprosthetics applications that aim at restoring damaged hearing, sight and movement. The objectives of the research is to check the different stimulation effects of human brain, to improve the accuracy, higher information transfer rate (ITR), desired bandwidth (BW), and signal to noise ratio (SNR) of BCIs and to identify Power and Energy of different Stimulation Effects on Alpha and Beta bands of EEG signal. In this research the power and energy of Alpha and Beta bands of EEG signal for different stimulation were determined and signal analysis, signal processing, Fast Fourier Transform (FFT), Statistical parameter methods were used. The performance of stimulator depends on many factors such as size and shape of stimulator, frequency of stimulation, luminance, color, and subject attention. Information Transfer Rate (ITR) varies with the change of frequency and size of the visual stimuli. In this research, a circular repetitive visual stimulator (CRVS) of different diameter (2", 2.5"and 3"), color (RGB), frequency (10, 15 and 20 Hz) was used. When the size of the stimulator changes from 2" to 2.5" a greater increase of Alpha wave (58.18%) is observed than Beta wave (13.68%). But a further increase in size from 2" to 2.5" a greater decrease of Alpha wave (45%) is observed than Beta wave (36.59%).
Motor Imagery Movement and Neurological Disorder Classification using Salient Features of EEG Signal
(Khulna University of Engineering & Technology (KUET), Khulna, Bangladesh., 2017-07) Rasshid, Md. Mamun Or; Ahmad, Prof. Dr. Mohiuddin.
Electroencephalography (EEG) measurement plays a significant role in the clinical and scientific research of brain studies. EEG signals are very important, particularly for classification and treatment of neurological diseases and brain computer interface (BCI) applications. The aim of this dissertation is to develop methods for the analysis and classification of different categories of motor imagery (MI) movements, epileptic EEG signals, and human alertness states. A novel method is also developed for continuous alertness monitoring. EEG signals of MI movements are classified for right hand and left hand (two class) and right hand, left hand, and feet (three class) movements. Nowadays, MI is a highly prescribed method for the disabled patients to give them hope to control machine or computer by interfacing with brain or mind. This dissertation proposes a classification method between imagery left and right hands movement using Daubechies wavelet of discrete wavelet transform (DWT) and Levenberg-Marquardt back propagation training algorithm of artificial neural network (ANN). DWT decomposes the raw EEG data to extract significant features that provide feature vectors precisely. ANN classifies the two class and three class trials data. Classification accuracy varies with respect to the subject. This method can be used to design a well-organized BCI system with better accuracy. Results from classifier can be used to design brain machine interface (BMI) for better performance that requires high precision and accuracy scheme. Neurological disorder i.e. epilepsy detection is enough time consuming and requires thorough observation to determine epilepsy type and locate the responsible area of the cerebral cortex. The dissertation proposes an effortless epilepsy classification method for epilepsy detection and investigates the classification accuracy of multiclass EEG signal during epilepsy. For accomplishing the proposed research work we use DWT to obtain responsible features to accumulate feature vectors. Afterward feature vectors are given in the input layer of the ANN classifiers to differentiate normal, interictal, and ictal EEG periods. Accuracy rate is calculated based on the confusion matrix. Proposed method can be utilized to monitor and detect epilepsy type incorporating with an alarm system. It is tiresome for human to concentrate constantly, though several works require continuous alertness like efficient driving, learning, etc. A practical method is applied to investigate the concentration state of human brain by EEG acquisition. This research work proposes continuous alertness state classification method based on two different types of mental tasks with respect to the resting state (resting with eyes open and eyes close). To conduct this research work, some participants were involved and they performed several tasks such as alphabet counting, virtual motor driving, resting with eyes open and eyes close. During the performances of the tasks, 9 channel EEG data has been acquired from their scalps. The data acquisition is performed by B-Alert (BIOPAC) system. The acquired data are filtered by IIR filter and responsible channels are selected by the statistical method. The features of the signals were extracted by using principal component analysis (PCA) and DWT algorithms. The alert states of our brain are classified by ANN. In addition, the specific relative power (RP) of the responsible frequency band of EEG signals is calculated for alertness monitoring. Within the RP range of resting and active state, a threshold value is proposed for monitoring the alertness state of the participants. This work will be helpful to classify the epileptic states with more accuracy as well as this works is also a well guide to classify the motor imagery movements. In addition, the proposed method based on continuous alertness monitoring will be remarkable approach to design machines for monitoring driving or learning.
Detection of Angina Pectoris Using ECG Signals
(Khulna University of Engineering & Technology (KUET), Khulna, Bangladesh, 2018-06) Islam, Md. Merajul; Hossain, Prof. Dr. A. B. M. Aowlad
Angina pectoris due to ischemia is very crucial to detect because, if it can be detected earlier, doctor can provide the patient proper medication to cure. Sometimes from a long ECG data it is tiresome to calculate and differentiate the normal and angina pectoris affected ECG peak to decide about the condition of the patient. In addition, the remote areas of lower and midlower income countries often face lack of experienced doctors or highly cost devices like ecocardiogram, MRI to detect angina. In these regions, automatic identification of ECG features can be a fruitful solution. Therefore, only way is computerized and efficient ECG analyzing algorithm development that can be able to detect angina pectoris. In this work, an efficient algorithm is developed to detect angina pectoris from ECG signal. This algorithm consists of several steps to take decision on ECG signal. First of all this algorithm removes baseline wandering from ECG signal by baseline wandering path finding algorithm. After that it removes other noises from ECG signal by Gaussian weighted moving average window method. In this consequence, QRS complex was detected by very well-known method First and Second Derivative (FS2) algorithm and gradually other important points like S, J, K, and T were detected by possible range maxima-minima criterion. Besides, the isoelectric line of ECG signal is estimated and eventually the statistical features of J-K points of normal and abnormal ECG peaks are compared with that isoelectric line by the algorithm, Finally, this algorithm takes the decision whether the patient is suffering from angina or not. This algorithm is applied on MIT arrhythmia Database to detect angina pectoris. From the result provided by this algorithm, we have found 94% (average) accuracy which is noticeable. In addition with that the sensitivity and specificity of our proposed algorithm have also been calculated which are found 91% and 89%, respectively. Since the previous work is based on the single feature, it may prove inappropriate for all the time. Therefore with the help of multiple features machine learning based approach k-nearest neighbor (kNN) method has been deployed in this research work to make it more accurate and acceptable. Although kNN based prediction method also provide almost similar results found by the previous methodologies. Therefore our proposed approach for angina pectoris detection has been testified by both statistical and kNN based method. It is expected that the proposed algorithm will be helpful for computerized angina pectoris detection from ECG signals.
Classification Accuracy Enhancement of fNIR based Imagery Movement by Modified Common Spatial Pattern
(Khulna University of Engineering & Technology (KUET), Khulna, Bangladesh, 2019-01) Kabir, Md. Faisal; Islam, Prof. Dr. Sheikh Md. Rabiul
Motor imagery event classification from functional near-infrared spectroscopy (fNIRS) is one of the most interesting problems of current brain-computer interfaces (BCIs) challenges because it needs no additional data guiding visual or listening protocol. A vital step of the fNIRS signal classification by machine learning approach is feature extraction. The feature extraction from multiple channel fNIRS signal is always challenging due to its high dimensionality. There exist several conventional feature extraction procedures like principal component analysis, nonlinear principal component analysis, independent component analysis, norm analysis, spectral norm analysis, etc. This research work studies such existing feature extraction method to classify the hand movement events of fNIRS signals. The accuracies of these methods have been found less than the expectation. Therefore, some more accurate method is needed. In this regard, usually common spatial pattern (CSP) is used to reduce the dimensional reduction and improving the classification accuracy. The conventional CSP method can be proven also ineffective for the motor imagery fNIRS signal due to its high level of trial to trial variations. The present research work proposes an algorithm named by standardized common spatial pattern (SCSP) based feature extraction method for fNIRS based motor imagery classification which can perform well in the context of the trial to trial significant variation. The classification results corresponding to the proposed feature extraction method reveal that the proposed SCSP algorithm outperformed the conventional CSP method and channel-wise method for classifying the two motor imagery event classifications. For classification accuracy measurement, four well-known classifiers: artificial neural network (ANN), k-nearest neighbor (kNN), support vector machine (SVM), and linear discriminant analysis (LDA) have been used. We have utilized both the fNIRS data of oxidized hemoglobin (HbO) and deoxidized hemoglobin (HbR) for classifying the motor imagery fNIRS data with the conventional and proposed methods. From the comparisons, we have found that in both cases, the proposed method outperforms the conventional methods in the context of the classification accuracies. To validate the classification accuracies, the sensitivities and specificities of the classifier are also calculated in this work. We believe that the proposed SCSP method will contribute in the field of feature extraction method for other types of fNIRS based BCI system, effectively. In addition, this method may be applied for the feature extraction to the other multidimensional signals so far.
In Silico Characterization and Homology Modeling of Histamine Receptors
(Khulna University of Engineering & Technology (KUET), Khulna, Bangladesh, 2019-02) Zobayer, Md. Nayem; Hossain, Prof. Dr. A. B. M. Aowlad
Histamine plays vital role in molecular mechanism of allergic reactions. Therefore, characterization and homology modeling of Histamine receptor is of great importance to design effective vaccines. In this thesis, different methods are applied to analyse biomolecular features of histamine receptors and design best models of these receptors. In addition to this, the study tried to identify potential B cell and T cell epitope based vaccine of an allergen and consequently, emphasized on to develop a B cell prediction tool. Identified four histamine receptors, such as Histamine H1, Histamine H2, Histamine H3 and Histamine H4 have been analysed through ProtParam to extract physiochemical properties and ClastalW algorithm has been applied to identify conserved regions. Motif and Transmembrane regions have been identified through MEME suit and TMHMM servers, respectively. For homology modeling, I-tasser has been used and generated models have been validated through RAMPAGE, ERRAT and PROCHECK. Targeted api m3 allergen then rendered through self-optimized prediction method with alignment for physiochemical feature extraction. NetCTL 1.2 has been applied to identify preliminary T cell epitope candidates and then scrutinized by Stabilized Matrix Base Method, relative to IC50 values. Predicted T cell epitopes have been further analysed for conservancy and population coverage via IEDB tools. B cell epitopes of api m3 allergen have been predicted through, BCPREDS, ABCpred, Bepipred and Bcepred. In addition, classifier based single interface B cell epitope prediction and/or validation tool has been developed through establishing efficient MATLAB algorithms to classify beta turn regions, hydrophilic regions, surface accessible regions and antigenic regions. Lastly, with superimposing graphical representation of these four criteria in a single interface graph plotted to identify B cell epitopes via this tool. Extracted results denotes that, Histamine receptors possess molecular weight around 55.7 KDa, theoretical pI 9.33-9.62, instability index 34.93-47.00, aliphatic index (AI) was above 90 and the receptors were hydrophobic except histamine H1 receptor. Moderately conserved region was found in 75-94 amino acid position. A profound motif has been identified from 84-149 amino acid position for four histamine receptors with significantly lower E-value. It has been identified that, these receptors are seven pass transmembrane protein and a gap between transmembrane helix number five and six was found in each histamine receptor except Histamine H2 receptor, which can be potential drug target candidate. Generated 3D models have been passed through every spheres of validation. Api m3 allergen has been found relative thermostable nature and only 10.46% of the overall secondary structure consisted of beta turn region. Five MHC class I T-cell epitopes were identified and scrutinized and YTEESVSAL found out as the best epitope. For MHC class II T-cell epitopes YPKDPYLYYDFYPLE and GGPLLRIFTKHMLDV have been found as most prominent T-cell epitopes of api m3 allergen. This study also revealed that, GDRIPDEKN and PHVPEYSSS, as the most effective B-cell epitopes of api m3. The proposed tool efficiently identified B cell epitopes and provided result in a single interface. The tool can aid in B cell research and vaccine development. Finally, the suggested potential drug targets can be applied in designing more sustainable antihistamines and relevant drugs in treating allergic diseases. Predicted T-cell and B-cell epitopes of api m3 allergen could help the researchers to test these vaccines further for immunoreactivity applying in vivo analysis. As still there is no report of T-cell and B-cell epitopes of Apis mellifera, this study can be the pioneer in finding effective vaccine against allergens of honeybee. This research also predicted potential B cell epitope regions from an antigenic protein. The most exciting feature of this part of the study is, it presents results of potential B cell epitopes on a single interface, so that, researchers don‘t need to search for every feature (e.g., hydrophilicity, antigenicity, beta turn, surface accessibility etc.) separately. Finally, the study can certainly aid in B cell epitope-based vaccine design research.
Development of Multiple Class Human-Computer Interaction System using Machine Learning Algorithm for Eyeball Movement
(Khulna University of Engineering & Technology (KUET), Khulna, Bangladesht, 2019-08) Chowdhury, Mubtasim Rafid; Mollah, Prof. Dr. Md. Nurunnabi
Modern technologies in the field of Biomedical Engineering are flourishing astonishingly in recent times. Human-computer interaction (HCI) is one of the newest additions in this field. It is the study of the way people interact with computers and how the computers are or are not developed for interacting with human successfully. Electrooculography (EOG) machine can be used as HCI device. It is a technique for measuring the corneo-retinal standing potential which is present between the front and the back of the human eye. Pairs of electrodes are generally attached either above and below the eye or to the left and right of the eye to detect the eye movement. A potential difference occurs between the electrodes. Considering that the resting potential is constant, the recorded potential is a measure of the eye's position. EOG device can pick up these resting potentials while moving the eyeball in different directions. Data classification is important for HCI systems. It is to identify a new observation which belongs to a set of categories. Classification is done based on a training dataset having observations whose category membership is familiar. Various algorithms can be used to classify bio-signal data like ECG, EEG and EOG. These algorithms are called machine learning algorithm. It is a set of mathematical approaches to teaching computers to train based on large amount of data without step by step human instruction. In this research, EOG data are classified using machine learning algorithm to develop a multiple class HCI system. EOG data for different directional eyeball movement is acquired with the help of Biopac MP3X Acquisition unit. By placing the disposable surface electrodes on the right position of the skull and connecting all the leads and wires to the proper channel, the setup is ready to pick up the EOG data. Subjects are instructed to follow the LED sequence in the navigational setup. The data of 7 subjects aged between 22 to 48 years are taken for this experiment. The data is then saved using Biopac Student Lab Software and then preprocessed to prepare an EOG dataset for classification. With Weka 3.9.2, the classification procedure is done on the prepared dataset. Six classification algorithms i.e. naïve bayes, support vector machine, logistic regression, k-nearest neighbor, random forest and bagging are applied on the dataset. Comparison is shown among the algorithms based on different parameters. In the EOG dataset, features are also added which can be correlated with the classes. This correlation method is performed in IBM SPSS Statistics 25 software to find the most significant features related to the class. From the classification result, the accuracy of the different classifiers are obtained. The accuracy of Naïve Bayes is 30.7692%, SVM is 30.7692%, Logistic regression is 53.8462%, KNN is 7.6923%, Random forest is 84.61% and Bagging is 92.31% respectively. From the comparison among the classifiers based on different parameters, bagging has the highest and KNN has the lowest accuracy among them. The proposed method is then compared with other researches where it is seen that other methods applied only two or three algorithms but in the proposed method six machine learning algorithms are used. It is observed that bagging is more suiTable algorithm for EOG data than other algorithms used in the mentioned related works. As for the correlation, only the chi-square test is performed as Fisher’s exact test can be performed for 2x2 matrix whereas there are 9 classes in the dataset. From the chi-square test result, it is seen that the mean of channel 1 (horizontal channel) and channel 2 (vertical channel) used to acquire EOG data for eyeball movement are the most significant features. These two featuress are directly related to the classes.
Prediction on Ischemic Heart Disease using Machine Learning Approaches
(Khulna University of Engineering & Technology (KUET), Khulna, Bangladesh, 2019-10) Raihan, M.; Islam, Dr. Muhammad Muinul
Ischemic heart disease (IHD) is a terrible experience that occurs when the flow of blood severely reduced or cut off due to plaque deposited on the inner wall of arteries that brings oxygen to the heart muscle, leads to the ischemic heart attack (IHA). Atherosclerosis i.e. plaque deposition on the inner wall of arteries is a silent process, has no critical symptoms to get a warning before IHD. For this reason, early detection is very important for the proper management of patients prone to IHD. In this thesis work, it was tried to predict IHD on the basis of patient history, symptoms and pathological findings of patients with heart disease using computational intelligence. Total 506 patient’s data with a maximum of 151 features including historic, symptomatic and pathologic findings were collected from AFC Fortis Escort Heart Institute, Khulna, Bangladesh. First, it was tried to identify the significant risk factors of IHD i.e. the features which are significantly correlated with IHD by applying different feature selection techniques. Then IHD was predicted using significant risk factors by applying different classifier algorithms. The significant risk factors of IHD were determined by using Chi-Square correlation, Ranking the features based on information gain and Best First Search techniques. Among 151 collected features only 28 features showed high correlations with IHD based on 0.05 significance level and information gain 1% or above. 10-fold cross-validation technique was applied with different classification algorithms e.g. Artificial Neural Network (ANN), Bagging, Logistic Regression, and Random Forest to predict IHD using the most significant 28 risk factors. IHD prediction accuracy was observed ranges from 95.85% to 97.63% with different classifier algorithm. Random Forest showed the best prediction performance with an accuracy of 97.63%. The same processing technique and classification algorithms were applied to the Cleveland hospital dataset to validate our prediction approach. The observed IHD prediction accuracy was 80.46-83.77% without applying the proposed processing techniques, but the accuracy degraded to 79.80-81.46% applying the proposed processing techniques. The Cleveland hospital data contains 303 patients’ data with only 13 features whereas the collected dataset contains 506 patient’s data with 28 nicely correlated IHD risk factors. This is why the proposed method is not suitably applicable to Cleveland dataset.