Tretriluxana, Suradej

Preferred name

Alternative Name

Tretriluxana, S.

Main Affiliation

School of Engineering

suradej.tr@kmitl.ac.th

Now showing 1 - 10 of 18

Some of the metrics are blocked by your
consent settings
Effect of Class Weights on Imbalanced Classes in Bi-directional LSTM Training for Sleep Apnea Classification
(2024-01-01)
Jansri, Ukkrit
;
Tretriluxana, Suradej
Sleep apnea, which is defined as the repetitive cessations of breathing during sleep, is the common disorder worldwide. The cost and the process of sleep test to obtain the polysomnogram is not optimal for sleep apnea screening in the large population. A deep learning model was developed to classify the normal and apnea events in a single time-series signal exported from the US National Institute of Health (NIH) sponsored database. Our challenge was to train the model with imbalanced dataset between normal and abnormal respiratory events. Three different methods, Synthetic Minority Over-sampling Technique (SMOTE), Random Under-Sampling (RUS), and the Class Weights (CW) were chosen to improve the model performance over the original data on five selected signals from polysomnographic dataset. The binary classification outputs were evaluated by four metrics. Our results showed (1) Matthews Correlation Coefficient was highest (MCC= 0.1385) in the Class Weights method on the nasal airflow signal. (2) Cohen's Kappa score, was highest (k= 0.0819) in SMOTE technique on the abdominal signal, followed by the Class Weights method on the abdominal signal (k= 0.0687) and RUS technique on nasal airflow signal (k= 0.0441). (3) F1-score was highest (F1= 11.89%) in SMOTE technique on the abdominal signal, followed by the Class Weights method on nasal airflow signal (F1 = 11.17%) and RUS technique on nasal airflow signal (F1= 9.16%). The findings suggest that the Class Weights method on nasal airflow and the Class Weights method on abdominal signal were the two combinations to be used in the DL model.
Some of the metrics are blocked by your
consent settings
Evaluating Marker-Based and Markerless Motion Capture Systems in Reach-to-Grasp Task
(2024-01-01)
Liangsorn, Natsakorn
;
Tretriluxana, Suradej
This study evaluated the accuracy of marker-based versus markerless motion capture systems in tracking reach-to-grasp movements, using an electromagnetic tracking system as the gold standard. Reach-to-grasp tasks are fundamental activity in daily activities and often affected by neurological conditions like stroke, highlighting the importance of accurate measurement in rehabilitation research. The findings show that the marker-based system consistently outperformed the markerless system, with significantly lower root mean square errors (RMSE) across all axes. The marker-based system achieved an average RMSE of 4.55 millimeters (mm), demonstrating clinical-level accuracy, while the markerless system averaged 23.64 mm, indicating limitations in its ability to track movements with high accuracy.
Some of the metrics are blocked by your
consent settings
Effect of Resampling Techniques on Deep Learning Model Training in Sleep Apnea Classification
(2022-01-01)
Jansri, Ukkrit
;
Tretriluxana, Suradej
This study is using deep learning model to classify the respiratory events of Sleep Disordered Breathing (SDB) data. Our pilot results showed the missed identification in some classes even the total accuracy is high. This is the result of unbalanced training dataset given to the model. Two different resampling techniques; Synthetic Minority Over-sampling Technique (SMOTE) and Random Under-Sampling (RUS), were introduced to balance the data. One hundred overnight nasal airflow signals were randomly selected from NIH funded polysomnography database. They were used to train and test these two algorithms with Bi-directional Long Short-Term Memory (Bi-LSTM) model. The results showed greater agreement index when compared between with and without data resampling process. However, SMOTE in sum performed better than RUS (93.72% vs 70.01% in overall accuracy and 0.91 vs 0.55 in Cohen's kappa). It demonstrates that the over-sampling technique is more powerful than under-sampling one. Other resampling techniques will be investigated to make the robust conclusion.
Some of the metrics are blocked by your
consent settings
A design of configurable ECG recorder module
(2011-12-01)
Punapung, Anucha
;
Tretriluxana, Suradej
;
Chitsakul, Kitiphol
A modular design of Electrocardiogram (ECG) acquisition system is presented. As a crucial part of the system, the Front-End module is made of a programmable, mixed-signal Application Specific Integrated Circuit (ASIC). This module is designed to satisfy the various forms of bioelectrical measurement with no hardware modification. Controlled by a 32-bit microcontroller, the prototype module recorded the simulated signal in two different standard ECG recording configurations, ten and six electrodes placements. Our study shows a good performance signal recording and the module is open to further developed. © 2011 IEEE.
Some of the metrics are blocked by your
consent settings
Reach-to-Grasp Motion Tracking System Using Single Smart Phone Camera
(2022-01-01)
Thongprasan, Sahawatchara
;
Liangsorn, Natsakorn
;
Raden-Ahmad, Chafik
;
Kongkaew, Kotchakorn
;
Tretriluxana, Jarugool
Assessment of human movement is necessary for physical therapy management. This article presents a development of motion tracking system for human Upper Extremity (UE) function analysis. We proposed the optical motion capture system made by a single smart phone camera. It was used to capture the Reach-to-Grasp (RTG) movement of participants in sitting position. Image processing were used to detect color markers placed on chosen hand anatomical landmarks. With our simple camera calibration technique, the 3D coordinates of hand movement were obtained. Two clinical parameters, grasp aperture and hand transport velocity were computed. These results were compared with the outputs, collected at the same time, from the higher accuracy Electromagnetic Motion (EM) tracking system. Qualitatively, the result patterns from two systems were parallel to each other. Our ongoing work is to improve the algorithms according to the feedback from clinicians. This system may provide implication for physical therapist to assess the clients' movement in the clinical setting.
Some of the metrics are blocked by your
consent settings
Comparison of Heart Rate statistical parameters from Photoplethysmographic signal in resting and exercise conditions
(2015-08-17)
Sengthipphany, T.
;
Tretriluxana, S.
;
Chitsakul, K.
The aim of this article is to investigate the possibility of computing Heart Rate Variability (HRV) indices from the finger Photoplethysmogram. To indicate that whether Pulse Rate Variability (PRV) from PPG signal can be used as an alternative to Heart Rate Variability (HRV) from ECG signal. The Photoplethsmographic (PPG) and Electrocardiographic (ECG) signals from 33 participants were recorded simultaneoussly during resting and exercise conditions. The peaks of PPG and R-waves of ECG were detected and reconstructed to the Peak-to-Peak interval (PPI) and R-to-R interval (RRI) waveforms respectively. In both conditions, the mean and standard deviation (SDNN) of the intervals over 5 minutes were computed from the two waveforms. Results from cross correlation between PPI and RRI show that the average correlation coefficients (r) are higher in resting than the 'r' in exercise. In regression analysis of the statistical parameters between RRI and PPI, the determination coefficients (R<sup>2</sup>) of the means are close to one in both conditions, whereas the R<sup>2</sup> of SDNN in exercise is lower than the one in resting. This finding suggests that the HRV indices can be evaluated from PPG with reliability during the rest.
Some of the metrics are blocked by your
consent settings
Brainwave spectrum analysis during paced breathing control: A pilot study
(2017-02-21)
Pakoktom, Nipawan
;
Tretriluxana, Suradej
;
Chitsakul, Kitiphol
Investigation of Electroencephalogram (EEG) has been made during the spontaneous breathing and paced breathing control (9 breaths/minute) conditions in eight young healthy college students (Male=4, Female=4). Their selected 5-minute EEG segments were transformed into 3 frequency bands (Delta Wave: 0.5-4Hz, Theta Wave: 4-8Hz and Alpha Wave: 8-13Hz). The power in each band were compared between two conditions. Our statistical tests show no significant differences between the two conditions in all three spectrums. This pilot study needs to be further analyzed to make a firm conclusion and a larger sample size is suggested.
Some of the metrics are blocked by your
consent settings
Analysis of Heart Rate Variability and Breath to Breath Interval in frequency domain
(2014-01-20)
Sengthipphany, Tick
;
Tretriluxana, Suradej
;
Chitsakul, Kitiphol
Respiratory is known to be a confounding factor of Heart Rate Variability (HRV) analysis. This article introduces a Breath-to-Breath Interval (BBI) spectral computation to investigate the insight between respiration and the HRV. Six males and five female volunteers (age 20-25 years) underwent the Electrocardiogram (ECG) and respiratory chest movement recordings for 5 minutes while sitting in resting condition. Auto Regressive Moving Average (ARMA) model was employed to the R-wave to R-wave interval (RRI) and BBI signals for the spectral analysis. The results from all participants demonstrate that the peak amplitude in high frequency (0.15-0.40 Hz) band are higher than the ones in low frequency (0.04-0.15 Hz) band in both RRI and BBI frequency plots. It suggests that respiratory plays a major role in HRV oscillation. Our further study is to develop a mathematical model to explain this finding.
Some of the metrics are blocked by your
consent settings
Differential effects of feedback in the virtual reality environment for arm rehabilitation after stroke
(2016-02-04)
Tretriluxana, Suradej
;
Tretriluxana, Jarugool
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.
Some of the metrics are blocked by your
consent settings
Automatic Sleep Data Scoring by Artificial Intelligence: A Pilot Study in Thai Population
(2021-04-01)
Jansri, Ukkrit
;
Chirakalwasan, Naricha
;
Chaitusaney, Busarakum
;
Busayakanon, Supasuta
;
Khongjui, Thamonwan
Sleep apnea, a sleep-disordered breathing (SDB), is defined as repeatedly intermittent cessation of breathing during sleep. It causes various life-threatening diseases. The American Academy of Sleep Medicine (AASM) releases the manual for sleep data scoring. Patients with SDB are prescribed to be monitored at the sleep clinic where several physiological data are recorded, called polysomnogram (PSG). The massive PSG data must be scored by the well-trained expert before being diagnosed by the physician. Our research is to use the Artificial Intelligence (AI) in sleep data scoring, particularly in respiratory events detection. Three ready-made Convolution Neural Networks (CNN); AlexNet, ResNet-50, and VGG-16, with transfer learning were applied to classify 5 overnight PSG data from Chulalongkorn hospital. Our preliminary results showed that all networks provide higher classification result in European Data Format (EDF) than in the text (ASCII) formats (71% vs 54%). The ResNet-50 model structure performed better than the other two networks on both data formats. As expected, the visualized (EDF) data is better than the unconditioned (ASCII) data. Our future development is modifying learning model to increase the scoring performance from more recruited PSG data.

Tretriluxana, Suradej

Filters

Settings

Sort By

Results per page