In 2015, stroke was the number one cause of death in Indonesia. The majority type of stroke is ischemic. The standard tool for diagnosing stroke is CT-Scan. For developing countries like Indonesia, the availability of CT-Scan is very limited and still relatively expensive. Because of the availability, another device that potential to diagnose stroke in Indonesia is EEG. Ischemic stroke occurs because of obstruction that can make the cerebral blood flow (CBF) on a person with stroke has become lower than CBF on a normal person (control) so that the EEG signal have a deceleration. On this study, we perform the ability of ID Convolutional Neural Network (1DCNN) to construct classification model that can distinguish the EEG and EOG stroke data from EEG and EOG control data. To accelerate training process our model we use Batch Normalization. Involving 62 person data object and from leave one out the scenario with five times repetition of measurement we obtain the average of accuracy 0.86 (F-Score 0.861) only at 200 epoch. This result is better than all over shallow and popular classifiers as the comparator (the best result of accuracy 0.69 and F-Score 0.72). The feature used in our study were only 24 handcrafted feature with simple feature extraction process.
Stroke diminishes someone’s quality of life since post-stroke symptoms can make some limbs do not function normally-in most cases, feet-thus the stroke patients’ mobility will be limited. However, the brain wave of stroke patients was observed as normal, thus in this study, the normal brainwaves were utilized to rehabilitate post-stroke patient and were expected to help their mobility through wheelchairs with brain waves control so that post-stroke patients can improve their quality of life. Brain Computer Interfaces (BCI) is a technology that allows taking action on a computer based on brain waves. Brain waves are recorded by electroencephalography so they can be processed by a computer. There have been many studies using BCI including analyzing brain waves in humans, many things that can be utilized using BCI make a lot of researchers use them to make smart wheelchairs that use brain control. and developments continue to be made to create the most optimal system. In this project we use one of the topics that is still being developed, that is Motor Imagery, where we record and analyze the brain waves while imagining motor activities such as moving hands, walking, running and so on. This record will serve as data reference to trigger the process on the computer to move the actuator. The purpose of this project is to control the wheelchair based on motor movements obtained by BCI using the Neurosky Mindwave Mobile 2 headset. This headset has one electrode where the signals from one electrode are analyzed by concentration and meditation values from the case of imagery motors, which in this project are more portable than using conventional EEG data acquisition devices that are not portable and use many channels. This headset is able to record data wirelessly via Bluetooth to a PC (Personal Computer), so the obtained signal can be processed and classified into five movement classes using the Matlab GUI, where the classes are default/motionless, move forward, move backward, turn right and turn left. The method used for the wheelchair was replacement of the default joystick in the electric wheelchair with an self-made controller module which was based on brain waves signals obtained from the headset will be processed and classified by Matlab GUI and then forwarded to Arduino Uno to control the motor in the wheelchair. The average success rates of the five classes from five trials were: the first class with a success rate of 82.22 %, the second class with a success rate of 70 %, the third class with a success rate of 73.33 %, the fourth class with a success rate of 46.67 % and the fifth class with a success rate of 17.78 %. The results of this study indicate that Neurosky Mindwave mobile 2 headset can be a possible choice for this project.
1 June 2021 in Journal of Physics: Conference Series
Indonesia has high level of seismic activity, so determining magnitude of an earthquake is important in the Earthquake Early Warning System. In the Earthquake Early Warning System, the parameter magnitude must be estimated earlier, so that warnings can be disseminated before the S and surface waves arrive. In previous studies machine learning technology can be used to recognized earthquake events and extract hidden information with massive datasets. This study was a preliminary, proposed the alternative methods to calculate the earthquake magnitude as fast as possible, the data was 1s before and 3 seconds after the P wave from the 3-component single station raw seismogram historical data and developed with a classification deep neural network (DNN) model, classical machine learning random forest (RF) algorithm and the regression deep neural network (DNN). Results from the statistical analysis show that the waveform can be modelled by deep neural network (DNN) models. Classification DNN Model that we constructed reaches good pattern which final loss of 0.63. If it benchmarked to another model such as Random forest (RF), Classification DNN was a better model than RF which is determined by final loss of RF. Our recommendation related to estimate the magnitude from seismic raw modelling are better using Classification DNN with larger dataset. In our study, with relatively small dataset, modelling using RF algorithm can be another option. Another suggestion related this work was utilizing the Regression DNN, that resulting best alternative related to estimation of magnitude.
20 January 2023 in Engineering Applications of Artificial Intelligence
In Indonesia, based on the result of Basic Health Research 2013, the number of stroke patients had increased from 8.3 ‰ (2007) to 12.1 ‰ (2013). These days, some researchers are using electroencephalography (EEG) result as another option to detect the stroke disease besides CT Scan image as the gold standard. A previous study on the data of stroke and healthy patients in National Brain Center Hospital (RS PON) used Brain Symmetry Index (BSI), Delta-Alpha Ratio (DAR), and Delta-Theta-Alpha-Beta Ratio (DTABR) as the features for classification by an Extreme Learning Machine (ELM). The study got 85% accuracy with sensitivity above 86 % for acute ischemic stroke detection. Using EEG data means dealing with many data dimensions, and it can reduce the accuracy of classifier (the curse of dimensionality). Principal Component Analysis (PCA) could reduce dimensionality and computation cost without decreasing classification accuracy. XGBoost, as the scalable tree boosting classifier, can solve real-world scale problems (Higgs Boson and Allstate dataset) with using a minimal amount of resources. This paper reuses the same data from RS PON and features from previous research, preprocessed with PCA and classified with XGBoost, to increase the accuracy with fewer electrodes. The specific fewer electrodes improved the accuracy of stroke detection. Our future work will examine the other algorithm besides PCA to get higher accuracy with less number of channels.
Noninvasive and nondestructive mobility assessment of dental implants is very important and useful for dental implantation diagnostic-aids. The development of implant movement (IM) checker based on microcontroller is presented in this paper. Data acquisition system and bender-type piezoelectric probe were used to improve measurement quality to the original tooth mobility (TM) tester. The adoption of a microcontroller and the use of a dental drill-sized measuring probe were sufficient in the reproducibility and reliability of the IM checker. When the implant was subjected to a constant force and amplitude, the acceleration of the model was detected using the measuring probe. The data acquisition system controlled for obtaining the appropriate acceleration signals based on the preload detection during measurement. Dental implant models of Molteno and Rigolac were made at different stiffness and were used to verify the reliability and validity of measurements. The values of measurements obtained by the IM checker were reliable and precise. The maximum error for perpendicular measurements was less than 12% measured by a new operator and decreased to 2% by an experienced operator. The IM checker was applied to monitor the stability of dental implantation, which compared the relative IM score of the new 5[see text] implant with the adjacent old 6 [see text] implant that had been used functionally for 3 years.