Showing papers presented at "Biomedical Engineering International Conference in 2015"

Stroke risk prediction model based on demographic data

Abstract: Nowadays stroke is the third leading cause of mortality of all life periods. The statistics from the Office of the National Economic and Social Development Board (NESDB) between 1994 and 2013 found that the stroke caused 255,307 cases mortality. Period of treatment in stroke patients depends on symptom and damage of organs. It seems to be beneficial if the data analysis method likes data mining can be used to predict stroke disease to reduce amount of risk patients before initial disease. In this study, three classification algorithms: Decision Tree, Naive Bayes and Neural Network are used for predicting stroke which are model-based, superior to general statistics, and got a proper model for identification. The scope of data use is the demographic information of patients. This work was initialized by attributes selection, grouping, and resampling before modeling. This study uses the accuracy and area under ROC curve (AUC) as the indicators for evaluation. Decision tree is the most accurate and Naive Bayes is the best in AUC. The further research should also include patients' diagnosis.

Hand postures for evaluating trigger finger using leap motion controller

Abstract: The purpose of this research reported here is to evaluate an abnormal finger motion of trigger finger using Leap Motion Controller to measure finger joint angles by the dot product equation. This sensor is 3D non-contact motion sensor which can detect finger bones. It is estimated positions and directions for fingers and bones. Four postures are used to evaluate an abnormal finger motion that has been used by physical therapist: 1) a flexion of thumb IP joint, 2) neutral position of finger PIP joint, 3) flexion of finger MP Joint, 4) thumb radial abduction. A finger goniometer is used as a reference method to measure the bent fingers' angle. The results showed that an average degree different value of a finger's angle measured by the finger goniometer and this system was 5.73 degrees.

Automated classification of malaria parasite species on thick blood film using support vector machine

Abstract: Malaria is a serious global health problem. It requires fast and effective diagnosis for detecting and classifying the type of infection. Proper treatment should be administered in a timely fashion to prevent an outbreak. Microscopic examination of thick blood films is one of the current standards for malaria diagnosis. However, inspecting a thick blood film is time-consuming and requires experienced technicians. Hence, for developing countries where most cases of malaria occur but microscopy expertise may not be available, a computerized system to aid such diagnosis is desirable. In this paper, an automated classification system operating on digitized images of thick blood film has been developed to classify between Plasmodium falciparum and Plasmodium vivax malaria parasite species. The system is fully automated. It is fast and can be handled by non-experts. We calculate five statistical features — mean, standard deviation, kurtosis, skewness and entropy — from four color channels (green, intensity, saturation, and value) of these images. The features are then projected onto a subspace representing image characteristics from both species. The projected features are used by the support vector machine for classification. It is found that the algorithm has acceptable training error and can classify test images with good accuracy.

A different approach to epilepsy risk level classification utilizing various distance measures as post classifiers

Abstract: The electrical activity of the brain can be studied thoroughly through the recordings of the Electroencephalography (EEG) signals and is considered as a vital tool for the analysis and diagnosis of neurological diseases like tumours of the brain, epilepsy and other cognitive disorders. Due to the continuous electrical discharges from the cortex of the cerebrum, epilepsy occurs which results in several severe consequences thereby making many vital changes in the EEG signal. In this paper, the epilepsy risk levels are classified by making use of Approximate Entropy as a Feature Extraction technique followed by Various Distance Measures such as Euclidean Distance Measure (EDM), City Block Distance Measure (CBDM) and Correlation Distance Measure (CDM) as Post Classifiers for the perfect classification of epilepsy risk levels from EEG signals. The validation parameters taken here are Performance Index (PI), Time Delay (TD), Quality Value (QV), Sensitivity, Specificity and Accuracy.

Design and construction of system to control the movement of the robot arm

Abstract: The objective of this project was to study on the design and construction of a system to control the movement of the robot arm from remotely via wirelessly glove using flex sensors and the variable resistor. The designed and constructed of the project composed of 5 main parts : 1) The detector consists of a control unit with Flex Sensors and a Position Sensors (variable resistor), 2) the signal conditioner part includes Voltage Divider circuit and voltage differential amplifier circuit, 3) the signal processing part using microcontroller Arduino Mega 2560 with C-language program, 4) the signal transmitter and receiver part using Xbee series 2 module (Zigbee Wireless Standard) and 5) robot arms designed and construction with Cartia program and 3-D printer. The results of functional testing of the project showed that this system can control the movement of the fingers, wrist, elbow and shoulders complying with the objectives of the project.

Adaptive wavelet technique for EEG de-noising

Abstract: Recording the electrical current of the cortex is called electroencephalography (EEG). EEG signals can be affected by high and low frequency noises which are caused due to muscular activity (EMG), Power line interference, eye blinks and etc. In this paper, we introduce an adaptive wavelet method for elimination of high frequency noises from EEG. The desired noise is extracted from the raw signal by wavelet approach. Afterwards, the extracted noise is applied as the input for adaptive filter. The effectiveness of the proposed method is compared to the forth order Butterworth high pass filtering with cut-off frequency at 30 Hz and EMD approach. In order to evaluate the performance of the methods, signal to noise rate (SNR) and correlation coefficient between pure and filtered signals are calculated. The obtained results suggest the proposed method can reduce EEG noise more efficiently than tile other methods.

Prediction of preterm labor from EHG signals using statistical and non-linear features

Abstract: Prediction of preterm labor is of great importance to reduce neonatal death. Analysis of electrohysterogram (EHG) could be considered as a proper tool for this aim. In this paper, the statistical and non-linear features have been extracted from EHG signals and then Support Vector machine (SVM) has been applied for classification between term and preterm labor. The dataset of this research consists of 26 records from term delivery (duration of pregnancy ≥37 weeks) and 26 records from pre-term delivery (duration of pregnancy <37 weeks). The obtained results show the highest accuracy can be achieved by 4 statistical features from channel 1.

The automated skull stripping of brain magnetic resonance images using the integrated method

Abstract: Skull stripping is one of the significant steps in brain image processing. There are still a number of difficulties using those common methods such as the region growing method. Aforesaid methods were largely depended on shape or intensity of non-brain tissues. This led to a difficulty when those non-brain tissues and intracranial have approximately the same intensity values. This research proposed an automatic skull stripping method based on the combination of mathematical morphology, component labeling and segmentation by Object Attribute Threshold (OAT). With this proposed method: MLO that combined the morphology, labeling and object attribute threshold method together, the removing of non-cerebral tissues can be completed. The proposed method also performed well even for the case that both cerebral and non-cerebral values on the MRI brain images have similar intensity. We used 20 samples of T1-weighted MRI brain images in the experiments.

Medical image encryption based on DCT-DWT domain combining 2D-DataMatrix Barcode

Abstract: In this paper, an application for medical image encryption and watermarking to store patient information is proposed. By using scrambling algorithm, the 1st view patient information is hidden and require unique patient password to access a real image. For watermarking scene, a Discrete Cosine Transform (DCT) and Discrete Wavelet Transform (DWT) are applied to non-overlapping block size 8×8 image pixels. Later, a DWT and DCT coefficient value are compared in each block for position identification of embed message in medical image. To improve high capacity and imperceptibility in watermarking step, a Two Dimension Barcode ECC200 standard is used to convert message before embed to cover image. Experimental results demonstrate that the proposed scheme achieves better performance of imperceptibility.

Active control of cell with microbubbles for cellular immunotherapy by acoustic force

Abstract: Recently immunotherapy using therapeutic cells has attracted attention. Because the cells injected in human body disperse through the blood stream, there is a problem that small amount reaches to the diseased area. So we intend for in vivo cell delivery system. Microbubbles were attracted on the surface of cells for dynamic control of cells under ultrasound emission. The cells and microbubbles were observed under fluorescent excitation to be observed their behavior in a thin channel, where the cells were induced to boundary of the channel. Furthermore, mean velocity of the cells increased with maximum sound pressure of the ultrasound, where the highest velocity of 2.2 mm/s was confirmed when the maximum sound pressure of 500 kPa-pp was applied. Through the experiments there is a possibility of delivery of in vivo cells.

Brain complexity analysis of functional near infrared spectroscopy for working memory study

Abstract: Working memory is an important brain function for memorizing information in everyday life. This study proposed an alternative approach for analyzing visual working memory based on brain complexity to identify brain state condition for memorizing new scenes. Multi-scale entropy (MSE) was used for analyzing the complexity of function Near Infrared Spectroscopy (fNIRS) data measuring the hemodynamic response of brain during a cognitive experiment. The results revealed the distinctive entropy between remembered and forgotten cases in premotor cortex area at FC3 position. The entropy of remembered case is higher than that of forgotten case; this could indicate that the brain requires more activity and, then, more hemodynamic responses for good memorizing. Hence, the results indicated the potential of using the complexity of fNIRS for characterizing the brain state for working memory.

A novel combination of code converters and sparse representation classifiers for an efficient epilepsy risk level classification

Abstract: The aim of this paper is to give a performance analysis by considering the advantage of Code Converters as a feature extraction technique and Sparse Representation Classifier (SRC) as a post classifier for the classification of the epilepsy risk levels obtained from Electroencephalography (EEG) signals. A group of related or similar disorders which is generally characterized by the occurrence of frequency and recurrent seizures is termed as epilepsy. There are different types of epilepsy and seizures. To control the seizures, epilepsy drugs are usually prescribed and if medication is ineffective, then surgery can be done to control or prevent the occurrence of the seizures. This paper aims at the classification of epilepsy risk levels using Sparse Representation Classifier. The Performance Index (PI) and Quality Values (QV) are the two parameters that are used to assess the performance of the code converter and the sparse representation classifiers.

Osteoblast proliferation on graphene oxide eletrodeposited on anodized titanium

Abstract: A bone-implant surface of anodized titanium (ATi) with graphene oxide (GO) composites showed the potentials of biocompatibility and antibacterial property. The aim of this study was to fabricate anodized titanium, coated with graphene oxide (ATiGO) alone, using anodization and anodic electrophoretic deposition, respectively. Scanning electron microscopy (SEM) was performed to investigate surface topography. The physiochemical properties of samples were evaluated by energy-dispersive X-ray spectroscope (EDX) and X-ray diffractrometer (XRD). Cell proliferation of mouse osteoblastic cell line (MC3T3-E1) was investigated using MTT assay on ATiGO, GO coated on titanium (TiGO), ATi, and commercially pure titanium (Ti). The preliminary results in this study suggest that GO coatings promote osteoblast proliferation after 3 and 5 days of cultures when compared with ATi and pure Ti.

A fast depixelation method of fiber bundle image for an embedded system

Abstract: By using an fiber bundle endomicroscope to transmit the image, image quality is typically degraded by pixelation artifacts due to the spacing between individual fiber core in the liber bundle. Generally7 a personal computer is required to remove these artifacts in real-time which making the overall system bulky and cost In this work, we proposed a method of fast image processing to eliminate pixelation artifacts and enhancing signal to noise of acquire images. The computation was implemented on an embedded system in order to prototype a truly affordable portable imaging system for a remote clinic. The calculation technique to implement real-time display involves a convolution of a square point spread function separated into two one-dimensional arrays with no floating point An averaged processing time of 13 millisecond is achieved from this method, which is foster than a standard Gaussian filtering method of an OpenCV library version 3.0 by a factor of 50.

Evaluation of a head movement artifact removal method for EEG considering real-time prosessing

Abstract: The present study proposes the method of removing head movement artifact for considering real-time processing and improving reliability of EEG under unconstrained measurement. In the proposed method, independent component analysis and high-pass filtering were used with a hybrid accelerometer, which was used for detecting components containing head movement artifact. To evaluate the proposed method, we measured EEG which included the artifact and P300 components. As a result, although EEG amplitude decreased, original P300 waveform appeared clearly. This result indicates that the proposed method is effective for removing head movement artifact while keeping EEG information.

System analysis and synthesis of mechatronic testbench for testing modules and control systems of transfemoral prostheses

Abstract: The ability to move independently in the space of environment predetermines many other possibilities for the invalid. Key to this are lower limb prostheses and, in particular, transfemoral. Provide lot of quality characteristics of transfemoral prostheses is possible already at the stage of synthesis. In this paper describes the design process of specialized mechatronic testbench for testing modules and control systems of transfemoral prostheses based on the methods of system analysis and synthesis. The testbench represents device of quasi-human-machine type (bio-inspired mechatronic stand — "amputee" and phisical model of transfemoral prosthesis with controlled artificial knee mechanism) with conditionally independent control systems. The testbench simulation performed in MapleSim. It will be used to developing new designs of controlled transfemoral prostheses.

Elements of vision based respiratory monitoring

Abstract: Traditional respiration sensing in neonatal intensive care often suffers from signal quality issues such as drop-out, presenting difficulties in emerging applications such as automated control of blood oxygen saturation. We are developing a non-contact vision based respiratory monitoring system, suitable for use in a neonatal intensive care unit, to monitor the presence of apnoea, and to provide estimates of respiratory rate and tidal volume as a supplementary system complimenting traditional respiratory monitoring methods. This paper reports on progress relating to data capture and logging and motion detection with this system. The system comprises a Microsoft Kinect™ webcam and PC running a custom Lab VIEW capture and logging program that takes images of neonates lying in intensive care cribs. Image acquisition methods have been developed allowing capturing of camera images and streaming images to disk at up to 30Hz. Image processing methods haw been developed to automatically identify neonates and to isolate areas of respiratory movement. This basic system will be evaluated in an upcoming pilot study.

Game-based virtual rehabilitation system for upper extremity using low-cost camera

Abstract: In health care, rehabilitation treatment provides interventions that go beyond medical therapy to help those with injuries and illness to re-establish themselves as productive and socially-integrated citizens. However, in developing countries, rehabilitation becomes a big challenge for the demands of rehabilitation from a great amount of patients with very limited medical services. Along with the development of technology, scholars in the rehabilitation field and medical care are trying to integrate Virtual Reality (VR) technology to perform cost reduction, home based activities, easy to deploy and maintain for patients. The paper aims at upper extremities rehabilitation for patients who lost their motor function caused by disconnection among their brain neurons or stroke. We propose a system using a low-cost camera to track movement of patient's hand and design game-based exercises to verify the functionality of the system on rehabilitation. Our experiment results have proved the effectiveness of this rehabilitation training task and also exploit topics related to VR technology.

Simplified pseudo-dynamic receive beamforming for FPGA implementation

Abstract: An algorithm to reduce computational time for pseudo-dynamic receive beamforming using FPGA is presented. This includes a method to store and retrieve the predetermined delays and other parameters. In addition, we present a simplified implementation method to compute the scanline one sample at a time. The addresses of the echo signal data from different piezoelectric elements are parallel calculated using the delays. Subsequently, the data are loaded and summed, resulting in one sample on the scanline. This method is easy to be modified for delay error compensation and for different window lengths. This algorithm is implemented on an FPGA (Virtex-4, Xilinx, inc., San Jose, CA) and providing fast computational time of 0.998 ms per scanline for 8192 samples at 40-MHz sampling with comparable image qualities to dynamic receive beamforming.

Design and construction of median cubital vein transillumination device by using LED

Abstract: The purpose of the research aims to design and construct the median cubital vein transillumination device by using N1R LED. This research is designed based on the principle of the light absorbance of the blood vessels, transillumination and electronics. The design and construction of median cubital vein transillumination device by using LED was composed of 2 main parts: 1) Power supply was composed of the 12 Volt, 1 Amperes of an electrical transformer, bridge rectifier circuit, and IC voltage regulator No. 7812 and 2) Light Source was composed of LED that have a near infrared wavelength for the suitable looking through the vein, LED board and LED circuit. The result of functional testing found that the median cubital vein transillumination device by using LED can help facilitate a clear vision for the blood vessels to penetrate compared to standard methods.

Development of a triangular micro well for single cell trapping — Computational study

Abstract: The aim of this study is to find appropriate parameters of a trapping device, i.e. equilateral triangular microwell array, using a computational simulation. This simulation was divided into nine cases of microwell with various depths and widths to help analyze the effects of them on velocity and vorticity distribution induced inside the microwell. The microwells with equilateral length of 40, 60 and 80μm, and each length consisted of three depths, i.e. 15, 30 and 45 μm, were studied. The results suggested that the equilateral triangular 40-μm microwell with the depth of 30 μm should be able to effectively trap a 10–20 μm particle.

Physiological and biomechanical model of patient specific lung motion based on 4D CT images

Abstract: Prediction of respiratory motion has the potential to substantially improve cancer radiation therapy. Tumor motion during irradiation reduces the target coverage and increases dose to healthy tissues. In this paper, we have developed a new dynamic biomechanical model of the respiratory system, permitting the simulation of a complete cycle of respiratory motion, based on the finite element method (FEM), including the real boundary conditions of the organs (the diaphragm, the thorax, mediastinum and skin behaviors) to predict the lung tumor displacement and deformation. The model is monitored by two muscles: the diaphragm and the rib kinematics. We validate our approach with real 4D CT images. The results demonstrate that the proposed approach is able to predict the respiratory motion with an average error less than 2.0 mm in the different lobes.

Airflow effect on microwave ablation in lung model

Abstract: Pulmonary microwave ablation has been trailed and became dramatically recognized as an alternative surgical method for lung cancer treatment due to its minimal invasive technique and destroyable only a small part of malignant tissue. Even though, lung is realized as a porous tissue in which filled of humid air and a capillary network that rather difficult to control an ablation region, then a prediction of lung destructive region by using MWA simulation seem to be a reliable investigation and numerical tool for supporting this medical maneuver. In this paper, we propose a simulation of microwave thermal ablation for lung cancer treatment in a simple lung model. Microwave applicator was designed as an opened-tip coaxial antenna of 2.45 GHz at 10 Watts. To study an effect of air flow in pulmonary bronchus, a small tube with air flow was also placed close to the microwave applicator then a temperature distribution and coagulation volume at 60 °c were analyzed. The in silico results show a predictable ablation shape which affected by a convection heat transfer of a humid air. However, before a real ablation with swine tissue, an amount of pulmonary blood flow in small blood vessel should be also included in the next investigation inorder to obtain more reasonable results.

Investigation of shear stress and cell survival in a microfluidic chip for a single cell study

Abstract: Shear stress has emerged as a significant player in cell viability because it directly affects cell physical and biological properties. This draws a great attention of cell survival in the field of microfluidics. Recently, the spiral microchannel technique has widely been used in a process of cell separation because it is a size-based separation technique and does not require an external field, however, the microchannel including its accessories generate the shear stress that causes cell damage! In order to address this issue, the reduction of shear stress in microfluidic devices needs to take into our consideration. This project presents the investigation of shear stress in the setup including syringe, silicone tube, needle, spiral microchannel device, straight channel and outlets by using computational simulation as well as finding the possibly survival cells on the devices using a microscopy with vital dye. According to our work, computational software potentially identifies the critical areas inside the device where cells would be damaged due to shear stresses.

Realization of robust real time robotic arm control system based on EMG signal

Abstract: Vast variety of EMG signal applications have been proposed and practiced on EMG control system such as rehabilitation and prosthetic hand and so on. In this paper, a realization of a robust real time robotic arm control system is proposed. First, the root mean square (RMS) of EMG signal for a hand movement is measured and the measured EMG signal is transformed to a complex number. Secondly, the transformed signal is expand Discrete Fourier series (DFS) and the coefficients of DFS is defined the individual feature of hand movement. Thirdly, a hand movement can be recognized by evaluating the Euclidian distance between the coefficients of DFS for unknown hand movement and the reference feature which has been evaluated in the training process. Finally, the experimental results show that the proposed method is robust and faster than the conventional methods.

Using ultrasonography method for monitoring ovarian status in freshwater stingray (Potamotrygon sp.)

Abstract: Freshwater stingray is one of the ornamental fish trade in Thailand. The information of reproductive management has been limited. No studies have been carried out on ovarian status. Therefore, the main objective of this study was to apply and develop techniques for monitoring ovarian status in freshwater stingray by using ultrasonography. The ovarian fecundity varied among species and only the left ovary seemed to be functional. The stingrays were placed in fitted tank with water in ventral recumbency position. The imaging was done in B-mode. The suitable frequency was 6.5 MHz, covered all size of mature female stingrays. There was no statistically significant relationship between body depth and ovarian size (p>0.05). In the TLV view, it was easy to manipulate and differentiate type and appearance of follicles or corpora lutea, meanwhile in the LDV view, estimate amount of the structures can be provided.

Differential effects of feedback in the virtual reality environment for arm rehabilitation after stroke

Abstract: Two types of augmented feedback (FB) were compared during reaching training in the Virtual Reality (VR) environment for chronic stroke patients. Six participants were assigned to receive either the Knowledge of Result (KR) or the Knowledge of Performance (KP) FB. They went through 12 training sessions, of which there were 75 Reach-to-Target trials in the VR. They were instructed to perform the actions as fast and accurate as possible. KR group were given audio feedback whereas KP group could see the hand path of their movement. To evaluate, the Wolf Motor Function Test (WMFT) were performed before (Baseline), after (Post training) and 1 week after (Post lweek) the training. Also the kinematics data, Total Movement Time (TMT), Peak Transport Velocity (VMax), and Relative timing of VMax (RP), were analyzed. Results show that only in KP group, the dexterity tasks of WMFT was improved after the training and maintained for at least 1 week. Additionally, VMax in KP group was higher and occurred later than that in KR group. These preliminary outcomes imply that different strategy of movement recovery after stroke result from different types of FB. More participants will be recruited in the future to confirm this finding.

Redundant feature identification and redundancy analysis for causal feature selection

Abstract: High dimensional data can lead to low accuracy of classification and take a long time to calculate because it contains irrelevant features and redundant features. To overcome this problem, dimension of data has to be reduced. Causal feature selection is one of methods for feature reduction but it cannot identify redundant features. This paper presents Parent-Children based for Causal Redundant Feature Identification (PCRF) algorithm to identify and remove redundant features. The accuracy of classification and number of feature reduced by PCRF algorithm are compared with correlation feature selection. According to the results, PCRF algorithm can identify redundant feature but has lower accuracy of classification than correlation feature selection.

A simple laboratory test of music therapy for insufficient sleep

Abstract: This article studies the behavior of human sleep found the insufficient sleep problem due to the brain performance. This experiment uses devices such as sleeptracker to save the results of a sleep and electroencephalogram (EEG) to detect brainwaves while enjoying music. However, this study considers only a general sleep and sleep event that a relaxed with the music. The findings indicated the music can help to reduce the problems of insufficient sleep. As a result, volunteers are developing the sleep as sufficiently and efficiently due to get the best quality of sleep in the life. It is also available to deploy to the general person in order to solve the problem of insufficient sleep.

Photopolymerization of hydrogels for cartilage tissue engineering

Abstract: Regeneration of damaged tissues by cell-based strategies has attracted much interest in various tissue engineering applications. For the repair of small cartilage defects, in situ encapsulation of stem cells in injectable, photopolymerizable hydrogels can be used to facilitate new cartilage formation with minimal invasive surgery. In this study, poly(ethylene glycol) diacrylate (PEGDA) hydrogels embedded with Wharton's Jelly mesenchymal stromal cells (WJMSCs) were developed by UV exposure and served as an in vitro model. We demonstrated that the cell viability was more than 60% on day 1 and remained relatively similar on day 8. The decrease in water uptake ability of PEGDA hydrogels without cells was noticeable after incubation in aqueous environment for 2 weeks, suggesting a loss of gel matrix. Taken together, the photopolymerization approach can potentially be applied to deliver stem cells locally at defect sites in VIVO to stimulate tissue regeneration.