Treesukon Treebupachatsakul
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Treesukon Treebupachatsakul has a total of 13 co-authors. Here are the top co-authors by number of shared publications:
- Suejit Pechprasarn
College of Biomedical Engineering, Rangsit University, Pathum Thani 12000, Thailand
4 shared publications - Siratchakrit Shinnakerdchoke
Department of Biomedical Engineering, School of Engineering, King Mongkut’s Institute of Technology Ladkrabang, Bangkok 10520, Thailand
2 shared publications - Apivitch Boosamalee
Department of Biomedical Engineering, School of Engineering, King Mongkut’s Institute of Technology Ladkrabang, Bangkok 10520, Thailand
1 shared publication Bhornsawan Thanathornwong
Department of General Dentistry, Faculty of Dentistry, Srinakharinwirot University, Bangkok, Thailand
1 shared publication- Chuchart Pintavirooj
Department of Biomedical Engineering, School of Engineering, King Mongkut’s Institute of Technology Ladkrabang, Bangkok 10520, Thailand
1 shared publication Dulyapong Rungraungrayabkul
Department of Oral Medicine and Periodontology, Faculty of Dentistry, Mahidol University, Bangkok, Thailand
1 shared publication
Latest Publications
- Open Access
- 22 August 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-12
Performance of deep learning models for the classification and object detection of different oral white lesions using photographic images
Computer vision adjunctive technology for oral lesion diagnoses has been developed to detect and identify Oral Potentially Malignant Disorders (OPMDs) and non-OPMDs. The early detection of OPMDs can reduce the risk of oral cancer development, improving the survival rate of the patients. This study aims to evaluate the computer vision technique in the white oral lesion domain within the scope of photographic images. Deep learning techniques for the classification of Convolution Neural Networks (CNNs) and transformer neural networks, and one-stage models of YOLOv7 and YOLOv8 were utilized to classify and detect five classes of OPMDs and non-OPMDs oral white lesions including oral leukoplakia, oral lichen planus, pseudomembranous candidiasis, oral ulcers covered with pseudomembrane and other white benign oral lesions. From the evaluation results of classification, the IFormerBase model achieves overperformance compared to CNN models with accuracy, precision, and F1 score of more than 80% on the test set. The best model for object detection is YOLOv7 with 84.5% mean Average Precision (mAP) at Intersection over Union (IoU) threshold of 0.3 and 74.5% at IoU of 0.5 on the test set. Object detection results reveal promising automatic oral lesion identification, which can be further developed to enhance the lesion screening system.
- 22 August 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-12
Open AccessPerformance of deep learning models for the classification and object detection of different oral white lesions using photographic images
Computer vision adjunctive technology for oral lesion diagnoses has been developed to detect and identify Oral Potentially Malignant Disorders (OPMDs) and non-OPMDs. The early detection of OPMDs can reduce the risk of oral cancer development, improving the survival rate of the patients. This study aims to evaluate the computer vision technique in the white oral lesion domain within the scope of photographic images. Deep learning techniques for the classification of Convolution Neural Networks (CNNs) and transformer neural networks, and one-stage models of YOLOv7 and YOLOv8 were utilized to classify and detect five classes of OPMDs and non-OPMDs oral white lesions including oral leukoplakia, oral lichen planus, pseudomembranous candidiasis, oral ulcers covered with pseudomembrane and other white benign oral lesions. From the evaluation results of classification, the IFormerBase model achieves overperformance compared to CNN models with accuracy, precision, and F1 score of more than 80% on the test set. The best model for object detection is YOLOv7 with 84.5% mean Average Precision (mAP) at Intersection over Union (IoU) threshold of 0.3 and 74.5% at IoU of 0.5 on the test set. Object detection results reveal promising automatic oral lesion identification, which can be further developed to enhance the lesion screening system.
- Open Access
Sensing Mechanisms of Rough Plasmonic Surfaces for Protein Binding of Surface Plasmon Resonance Detection
Surface plasmon resonance (SPR) has been utilized in various optical applications, including biosensors. The SPR-based sensor is a gold standard for protein kinetic measurement due to its ultrasensitivity on the plasmonic metal surface. However, a slight change in the surface morphology, such as roughness or pattern, can significantly impact its performance. This study proposes a theoretical framework to explain sensing mechanisms and quantify sensing performance parameters of angular surface plasmon resonance detection for binding kinetic sensing at different levels of surface roughness. The theoretical investigation utilized two models, a protein layer coating on a rough plasmonic surface with and without sidewall coatings. The two models enable us to separate and quantify the enhancement factors due to the localized surface plasmon polaritons at sharp edges of the rough surfaces and the increased surface area for protein binding due to roughness. The Gaussian random surface technique was employed to create rough metal surfaces. Reflectance spectra and quantitative performance parameters were simulated and quantified using rigorous coupled-wave analysis and Monte Carlo simulation. These parameters include sensitivity, plasmonic dip position, intensity contrast, full width at half maximum, plasmonic angle, and figure of merit. Roughness can significantly impact the intensity measurement of binding kinetics, positively or negatively, depending on the roughness levels. Due to the increased scattering loss, a tradeoff between sensitivity and increased roughness leads to a widened plasmonic reflectance dip. Some roughness profiles can give a negative and enhanced sensitivity without broadening the SPR spectra. We also discuss how the improved sensitivity of rough surfaces is predominantly due to the localized surface wave, not the increased density of the binding domain.
Open AccessSensing Mechanisms of Rough Plasmonic Surfaces for Protein Binding of Surface Plasmon Resonance Detection
Surface plasmon resonance (SPR) has been utilized in various optical applications, including biosensors. The SPR-based sensor is a gold standard for protein kinetic measurement due to its ultrasensitivity on the plasmonic metal surface. However, a slight change in the surface morphology, such as roughness or pattern, can significantly impact its performance. This study proposes a theoretical framework to explain sensing mechanisms and quantify sensing performance parameters of angular surface plasmon resonance detection for binding kinetic sensing at different levels of surface roughness. The theoretical investigation utilized two models, a protein layer coating on a rough plasmonic surface with and without sidewall coatings. The two models enable us to separate and quantify the enhancement factors due to the localized surface plasmon polaritons at sharp edges of the rough surfaces and the increased surface area for protein binding due to roughness. The Gaussian random surface technique was employed to create rough metal surfaces. Reflectance spectra and quantitative performance parameters were simulated and quantified using rigorous coupled-wave analysis and Monte Carlo simulation. These parameters include sensitivity, plasmonic dip position, intensity contrast, full width at half maximum, plasmonic angle, and figure of merit. Roughness can significantly impact the intensity measurement of binding kinetics, positively or negatively, depending on the roughness levels. Due to the increased scattering loss, a tradeoff between sensitivity and increased roughness leads to a widened plasmonic reflectance dip. Some roughness profiles can give a negative and enhanced sensitivity without broadening the SPR spectra. We also discuss how the improved sensitivity of rough surfaces is predominantly due to the localized surface wave, not the increased density of the binding domain.
- Open Access
- 1 March 2023
- Published by MDPI AG in Fermentation
- Vol. 9 (3) , 239
Machine Learning Models Using Data Mining for Biomass Production from Yarrowia lipolytica Fermentation
In this paper, a database of biomass production from Yarrowia lipolytica fermentation is prepared and constructed using machine learning and data mining approaches. The database is curated from 15 publications and consists of 301 rows of data with 25 predictors and 1 label. The predictors include inoculum size, temperature, pH, and time, while the label is the corresponding biomass production. The database is then divided into training, validation, and test datasets and analyzed as a supervised machine learning task for regression. Twenty-six regression models are employed and compared for their performance in predicting biomass production. The best-performing model is the Matern 5/2 Gaussian process regression model, which has the lowest root-mean-squared error of 0.75 g/L, the highest R squared of 0.90, and the lowest mean absolute error of 0.52 g/L. The t-test is used to identify the most important predictors, and 14 predictors are sufficient for creating an accurate model. These 14 predictors are fermentation time, peptone, temperature, total Kjeldahl nitrogen, shaking rate, total nitrogen, inoculum size, yeast extract, crude glycerol, glucose, oil and grease, media pH, ammonium sulfate, and olive oil. This research demonstrates the application of machine learning and data mining to estimate biomass production and gives insight into which parameters are essential for Yarrowia lipolytica fermentation.
- 1 March 2023
- Published by MDPI AG in Fermentation
- Vol. 9 (3) , 239
Open AccessMachine Learning Models Using Data Mining for Biomass Production from Yarrowia lipolytica Fermentation
In this paper, a database of biomass production from Yarrowia lipolytica fermentation is prepared and constructed using machine learning and data mining approaches. The database is curated from 15 publications and consists of 301 rows of data with 25 predictors and 1 label. The predictors include inoculum size, temperature, pH, and time, while the label is the corresponding biomass production. The database is then divided into training, validation, and test datasets and analyzed as a supervised machine learning task for regression. Twenty-six regression models are employed and compared for their performance in predicting biomass production. The best-performing model is the Matern 5/2 Gaussian process regression model, which has the lowest root-mean-squared error of 0.75 g/L, the highest R squared of 0.90, and the lowest mean absolute error of 0.52 g/L. The t-test is used to identify the most important predictors, and 14 predictors are sufficient for creating an accurate model. These 14 predictors are fermentation time, peptone, temperature, total Kjeldahl nitrogen, shaking rate, total nitrogen, inoculum size, yeast extract, crude glycerol, glucose, oil and grease, media pH, ammonium sulfate, and olive oil. This research demonstrates the application of machine learning and data mining to estimate biomass production and gives insight into which parameters are essential for Yarrowia lipolytica fermentation.
- Open Access
- 4 March 2022
- Published by MDPI AG in Biosensors
- Vol. 12 (3) , 159
Cuff-Less Blood Pressure Prediction from ECG and PPG Signals Using Fourier Transformation and Amplitude Randomization Preprocessing for Context Aggregation Network Training
This research proposes an algorithm to preprocess photoplethysmography (PPG) and electrocardiogram (ECG) signals and apply the processed signals to the context aggregation network-based deep learning to achieve higher accuracy of continuous systolic and diastolic blood pressure monitoring than other reported algorithms. The preprocessing method consists of the following steps: (1) acquiring the PPG and ECG signals for a two second window at a sampling rate of 125 Hz; (2) separating the signals into an array of 250 data points corresponding to a 2 s data window; (3) randomizing the amplitude of the PPG and ECG signals by multiplying the 2 s frames by a random amplitude constant to ensure that the neural network can only learn from the frequency information accommodating the signal fluctuation due to instrument attachment and installation; (4) Fourier transforming the windowed PPG and ECG signals obtaining both amplitude and phase data; (5) normalizing both the amplitude and the phase of PPG and ECG signals using z-score normalization; and (6) training the neural network using four input channels (the amplitude and the phase of PPG and the amplitude and the phase of ECG), and arterial blood pressure signal in time-domain as the label for supervised learning. As a result, the network can achieve a high continuous blood pressure monitoring accuracy, with the systolic blood pressure root mean square error of 7 mmHg and the diastolic root mean square error of 6 mmHg. These values are within the error range reported in the literature. Note that other methods rely only on mathematical models for the systolic and diastolic values, whereas the proposed method can predict the continuous signal without degrading the measurement performance and relying on a mathematical model.
- 4 March 2022
- Published by MDPI AG in Biosensors
- Vol. 12 (3) , 159
Open AccessCuff-Less Blood Pressure Prediction from ECG and PPG Signals Using Fourier Transformation and Amplitude Randomization Preprocessing for Context Aggregation Network Training
This research proposes an algorithm to preprocess photoplethysmography (PPG) and electrocardiogram (ECG) signals and apply the processed signals to the context aggregation network-based deep learning to achieve higher accuracy of continuous systolic and diastolic blood pressure monitoring than other reported algorithms. The preprocessing method consists of the following steps: (1) acquiring the PPG and ECG signals for a two second window at a sampling rate of 125 Hz; (2) separating the signals into an array of 250 data points corresponding to a 2 s data window; (3) randomizing the amplitude of the PPG and ECG signals by multiplying the 2 s frames by a random amplitude constant to ensure that the neural network can only learn from the frequency information accommodating the signal fluctuation due to instrument attachment and installation; (4) Fourier transforming the windowed PPG and ECG signals obtaining both amplitude and phase data; (5) normalizing both the amplitude and the phase of PPG and ECG signals using z-score normalization; and (6) training the neural network using four input channels (the amplitude and the phase of PPG and the amplitude and the phase of ECG), and arterial blood pressure signal in time-domain as the label for supervised learning. As a result, the network can achieve a high continuous blood pressure monitoring accuracy, with the systolic blood pressure root mean square error of 7 mmHg and the diastolic root mean square error of 6 mmHg. These values are within the error range reported in the literature. Note that other methods rely only on mathematical models for the systolic and diastolic values, whereas the proposed method can predict the continuous signal without degrading the measurement performance and relying on a mathematical model.
- Open Access
Analysis of Effects of Surface Roughness on Sensing Performance of Surface Plasmon Resonance Detection for Refractive Index Sensing Application
This paper provides a theoretical framework to analyze and quantify roughness effects on sensing performance parameters of surface plasmon resonance measurements. Rigorous coupled-wave analysis and the Monte Carlo method were applied to compute plasmonic reflectance spectra for different surface roughness profiles. The rough surfaces were generated using the low pass frequency filtering method. Different coating and surface treatments and their reported root-mean-square roughness in the literature were extracted and investigated in this study to calculate the refractive index sensing performance parameters, including sensitivity, full width at half maximum, plasmonic dip intensity, plasmonic dip position, and figure of merit. Here, we propose a figure-of-merit equation considering optical intensity contrast and signal-to-noise ratio. The proposed figure-of-merit equation could predict a similar refractive index sensing performance compared to experimental results reported in the literature. The surface roughness height strongly affected all the performance parameters, resulting in a degraded figure of merit for surface plasmon resonance measurement.
Open AccessAnalysis of Effects of Surface Roughness on Sensing Performance of Surface Plasmon Resonance Detection for Refractive Index Sensing Application
This paper provides a theoretical framework to analyze and quantify roughness effects on sensing performance parameters of surface plasmon resonance measurements. Rigorous coupled-wave analysis and the Monte Carlo method were applied to compute plasmonic reflectance spectra for different surface roughness profiles. The rough surfaces were generated using the low pass frequency filtering method. Different coating and surface treatments and their reported root-mean-square roughness in the literature were extracted and investigated in this study to calculate the refractive index sensing performance parameters, including sensitivity, full width at half maximum, plasmonic dip intensity, plasmonic dip position, and figure of merit. Here, we propose a figure-of-merit equation considering optical intensity contrast and signal-to-noise ratio. The proposed figure-of-merit equation could predict a similar refractive index sensing performance compared to experimental results reported in the literature. The surface roughness height strongly affected all the performance parameters, resulting in a degraded figure of merit for surface plasmon resonance measurement.
- Open Access
- 5 March 2021
- Published by MDPI AG in Healthcare
- Vol. 9 (3) , 285
Multi-Parameter Vital Sign Telemedicine System Using Web Socket for COVID-19 Pandemics
Telemedicine has become an increasingly important part of the modern healthcare infrastructure, especially in the present situation with the COVID-19 pandemics. Many cloud platforms have been used intensively for Telemedicine. The most popular ones include PubNub, Amazon Web Service, Google Cloud Platform and Microsoft Azure. One of the crucial challenges of telemedicine is the real-time application monitoring for the vital sign. The commercial platform is, by far, not suitable for real-time applications. The alternative is to design a web-based application exploiting Web Socket. This research paper concerns the real-time six-parameter vital-sign monitoring using a web-based application. The six vital-sign parameters are electrocardiogram, temperature, plethysmogram, percent saturation oxygen, blood pressure and heart rate. The six vital-sign parameters were encoded in a web server site and sent to a client site upon logging on. The encoded parameters were then decoded into six vital sign signals. Our proposed multi-parameter vital-sign telemedicine system using Web Socket has successfully remotely monitored the six-parameter vital signs on 4G mobile network with a latency of less than 5 milliseconds.
- 5 March 2021
- Published by MDPI AG in Healthcare
- Vol. 9 (3) , 285
Open AccessMulti-Parameter Vital Sign Telemedicine System Using Web Socket for COVID-19 Pandemics
Telemedicine has become an increasingly important part of the modern healthcare infrastructure, especially in the present situation with the COVID-19 pandemics. Many cloud platforms have been used intensively for Telemedicine. The most popular ones include PubNub, Amazon Web Service, Google Cloud Platform and Microsoft Azure. One of the crucial challenges of telemedicine is the real-time application monitoring for the vital sign. The commercial platform is, by far, not suitable for real-time applications. The alternative is to design a web-based application exploiting Web Socket. This research paper concerns the real-time six-parameter vital-sign monitoring using a web-based application. The six vital-sign parameters are electrocardiogram, temperature, plethysmogram, percent saturation oxygen, blood pressure and heart rate. The six vital-sign parameters were encoded in a web server site and sent to a client site upon logging on. The encoded parameters were then decoded into six vital sign signals. Our proposed multi-parameter vital-sign telemedicine system using Web Socket has successfully remotely monitored the six-parameter vital signs on 4G mobile network with a latency of less than 5 milliseconds.