Thanh-Nghia Nguyen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Thanh-Nghia Nguyen has a total of 39 co-authors. Here are the top co-authors by number of shared publications:
- Thanh-Hai Nguyen
Faculty of Electrical and Electronics Engineering, Ho Chi Minh City University of Technology and Education, Ho Chi Minh City 700000, Vietnam
10 shared publications - Ba-Viet Ngo
Faculty of Electrical and Electronics Engineering, Ho Chi Minh City University of Technology and Education, Ho Chi Minh City 700000, Vietnam
8 shared publications D. V. Binh
Ho Chi Minh City University of Technology and Education, Viet Nam
2 shared publications- Duc-Tan Tran
Faculty of Electrical and Electronic Engineering, Phenikaa School of Engineering, Phenikaa University, Hanoi, Vietnam
2 shared publications N. M. Hung
Ho Chi Minh City University of Technology and Education, Viet Nam
2 shared publicationsThanh-Tam Nguyen
Faculty of ElectricalElectronics Engineering, HCMC University of Technology and Engineering, Ho Chi Minh City, Vietnam
2 shared publications
Latest Publications
- Open Access
- 30 October 2026
- Published by Hanoi University of Science and Technology in Engineering and Technology For Sustainable Development
- Vol. 36 (5) , 103-115
A Cost-Optimized EEG-Based BCI System for Eye-Blink Behavior Recognition Using CBA-FusionNet
Electroencephalogram (EEG) signals in the field of Brain–Computer Interfaces (BCI) are increasingly being explored for research purposes, intelligent device control, behavior recognition, and supporting users in contactless interactions. To address the demands of BCI applications, this study proposes an eye-blink signal acquisition device based on EEG signals with key characteristics including compact size, low cost, and high accuracy. In parallel, a deep learning model based on a 1D Convolutional Neural Network (1D-CNN) integrated with the CBA-FusionNet architecture—combining CNN, BiLSTM, and an Attention mechanism—is developed to maximize the classification performance of eye-blink signals corresponding to specific behaviors such as left-eye blink, right-eye blink, simultaneous blinking of both eyes, and the neutral (no-blink) state. By integrating the signal acquisition hardware with CNN-based signal processing software, the proposed system forms a complete framework with potential applications in contactless control interfaces and user-assistive devices.
- 30 October 2026
- Published by Hanoi University of Science and Technology in Engineering and Technology For Sustainable Development
- Vol. 36 (5) , 103-115
Open AccessA Cost-Optimized EEG-Based BCI System for Eye-Blink Behavior Recognition Using CBA-FusionNet
Electroencephalogram (EEG) signals in the field of Brain–Computer Interfaces (BCI) are increasingly being explored for research purposes, intelligent device control, behavior recognition, and supporting users in contactless interactions. To address the demands of BCI applications, this study proposes an eye-blink signal acquisition device based on EEG signals with key characteristics including compact size, low cost, and high accuracy. In parallel, a deep learning model based on a 1D Convolutional Neural Network (1D-CNN) integrated with the CBA-FusionNet architecture—combining CNN, BiLSTM, and an Attention mechanism—is developed to maximize the classification performance of eye-blink signals corresponding to specific behaviors such as left-eye blink, right-eye blink, simultaneous blinking of both eyes, and the neutral (no-blink) state. By integrating the signal acquisition hardware with CNN-based signal processing software, the proposed system forms a complete framework with potential applications in contactless control interfaces and user-assistive devices.
- 20 May 2026
- Published by Thai Nguyen University in TNU Journal of Science and Technology
- Vol. 231 (06) , 369-376
STUDENT TRANSPORTATION MONITORING SYSTEM BASED ON EMBEDDED SYSTEMS AND IOT: A CASE STUDY IN VIET NAM
Đảm bảo an toàn cho học sinh trong quá trình di chuyển giữa nhà và trường luôn là mối quan tâm hàng đầu của phụ huynh. Nghiên cứu này đề xuất một hệ thống giám sát và hỗ trợ khẩn cấp cho việc vận chuyển học sinh nhằm ngăn ngừa tình trạng học sinh bị bỏ quên trên xe buýt trường học. Hệ thống nhúng sử dụng hệ điều hành thời gian thực được sử dụng để lập trình và điều khiển hệ thống, trong khi phương pháp Internet vạn vật (IoT) được áp dụng để quản lý dữ liệu học sinh. Cụ thể, công nghệ RFID được sử dụng để theo dõi sự có mặt của học sinh, trong khi công nghệ GPS được sử dụng để giám sát vị trí và thực hiện cuộc gọi khẩn cấp. Giao diện người dùng được thiết kế để cung cấp nhiều chức năng khác nhau như quản lý điểm danh, giám sát và liên lạc cho ba nhóm người dùng: nhà trường (để quản lý), tài xế (để giám sát học sinh) và phụ huynh (để theo dõi con em mình). Kết quả thực nghiệm cho thấy hệ thống có thể giám sát thời gian lên và xuống xe của học sinh trong thời gian thực. Nhà trường có thể quản lý tuyến đường và thời gian di chuyển của tài xế, trong khi phụ huynh có thể theo dõi thời gian đến trường của con em mình.
- 20 May 2026
- Published by Thai Nguyen University in TNU Journal of Science and Technology
- Vol. 231 (06) , 369-376
STUDENT TRANSPORTATION MONITORING SYSTEM BASED ON EMBEDDED SYSTEMS AND IOT: A CASE STUDY IN VIET NAM
Đảm bảo an toàn cho học sinh trong quá trình di chuyển giữa nhà và trường luôn là mối quan tâm hàng đầu của phụ huynh. Nghiên cứu này đề xuất một hệ thống giám sát và hỗ trợ khẩn cấp cho việc vận chuyển học sinh nhằm ngăn ngừa tình trạng học sinh bị bỏ quên trên xe buýt trường học. Hệ thống nhúng sử dụng hệ điều hành thời gian thực được sử dụng để lập trình và điều khiển hệ thống, trong khi phương pháp Internet vạn vật (IoT) được áp dụng để quản lý dữ liệu học sinh. Cụ thể, công nghệ RFID được sử dụng để theo dõi sự có mặt của học sinh, trong khi công nghệ GPS được sử dụng để giám sát vị trí và thực hiện cuộc gọi khẩn cấp. Giao diện người dùng được thiết kế để cung cấp nhiều chức năng khác nhau như quản lý điểm danh, giám sát và liên lạc cho ba nhóm người dùng: nhà trường (để quản lý), tài xế (để giám sát học sinh) và phụ huynh (để theo dõi con em mình). Kết quả thực nghiệm cho thấy hệ thống có thể giám sát thời gian lên và xuống xe của học sinh trong thời gian thực. Nhà trường có thể quản lý tuyến đường và thời gian di chuyển của tài xế, trong khi phụ huynh có thể theo dõi thời gian đến trường của con em mình.
- Open Access
- 2 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 35511-35526
Dual-View Colorized Mammogram ROIs With EfficientNet-B7 and Attention Network for Breast Lesion Classification
Breast cancer diagnosis using mammography can be significantly enhanced by integrating information from multi-source fusion and complementary imaging perspectives. In this paper, we propose a novel deep learning framework with dataset fusion for breast lesion classification. In particular, our approach fuses mammogram datasets of both Medio-Lateral-Oblique (MLO) and Cranio-Caudal (CC) views in four breast lesion categories from two data sources to ensure comprehensive lesion representation with total 5,906 images, and a combined EfficientNet-B7 and Multi-Head Attention (MHA) network is applied to classify these breast lesion categories. Therefore, the largest Region of Interest (ROI) in each image is extracted and then colorized before inserted into the network for lesion classification. Two colorized MLO and CC ROIs are separately processed by two parallel EfficientNet-B7 networks to produce the combined high-level features. The combined feature vectors from the dual EfficientNet-B7 network are then concatenated to yield a unified feature representation, which is subsequently refined using the MHA network to prioritize the most salient patterns and suppress background noise before classified using a fully connected head. To select an optimal classification model for these breast lesion categories, we performed four different models for two cases with and without an MHA network. Therefore, extensive experiments on integrated multi-source mammogram datasets demonstrate that the proposed framework outperforms conventional single-view and non-attentive models, achieving around 93.9% accuracy for breast lesion classification. It is obvious that the use of MHA improves accuracy by about 2% compared to not using it. The results highlight the value of dataset fusion and MHA-based feature selection in advancing computer-aided mammography.
- 2 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 35511-35526
Open AccessDual-View Colorized Mammogram ROIs With EfficientNet-B7 and Attention Network for Breast Lesion Classification
Breast cancer diagnosis using mammography can be significantly enhanced by integrating information from multi-source fusion and complementary imaging perspectives. In this paper, we propose a novel deep learning framework with dataset fusion for breast lesion classification. In particular, our approach fuses mammogram datasets of both Medio-Lateral-Oblique (MLO) and Cranio-Caudal (CC) views in four breast lesion categories from two data sources to ensure comprehensive lesion representation with total 5,906 images, and a combined EfficientNet-B7 and Multi-Head Attention (MHA) network is applied to classify these breast lesion categories. Therefore, the largest Region of Interest (ROI) in each image is extracted and then colorized before inserted into the network for lesion classification. Two colorized MLO and CC ROIs are separately processed by two parallel EfficientNet-B7 networks to produce the combined high-level features. The combined feature vectors from the dual EfficientNet-B7 network are then concatenated to yield a unified feature representation, which is subsequently refined using the MHA network to prioritize the most salient patterns and suppress background noise before classified using a fully connected head. To select an optimal classification model for these breast lesion categories, we performed four different models for two cases with and without an MHA network. Therefore, extensive experiments on integrated multi-source mammogram datasets demonstrate that the proposed framework outperforms conventional single-view and non-attentive models, achieving around 93.9% accuracy for breast lesion classification. It is obvious that the use of MHA improves accuracy by about 2% compared to not using it. The results highlight the value of dataset fusion and MHA-based feature selection in advancing computer-aided mammography.
- Open Access
- 6 October 2025
- Published by Engineering, Technology & Applied Science Research in Engineering, Technology & Applied Science Research
- Vol. 15 (5) , 26935-26942
A Safe 2D Grid Map for Indoor Wheelchair Navigation Using the LiDAR Camera System
Mobile robots are widely used in localization, mapping, and path planning, with map construction being essential for safe navigation. This study presents a method for creating a safe 2D grid map for indoor electric wheelchair navigation using a LiDAR-camera system. The process converts a 3D environmental map into a 2D grid map, helping the wheelchair detect obstacles at unsafe heights. First, a 3D environmental map is generated using the LiDAR-camera system of an iPhone 12 Pro, capturing 3D point clouds. The Iterative Closest Point (ICP) algorithm merges these point clouds into a complete 3D map. Accuracy is verified by comparing the distances between the real environment and the generated map, with deviations less than 2%. To enhance safety, points exceeding the wheelchair's height limit are removed, forming a refined 3D height map. This map is then projected into a 2D representation and overlaid with a grid, ensuring each cell matches the wheelchair's dimensions. A pixel density approximation refines the grid. The Q-learning method is applied for autonomous wheelchair navigation, proving the system's effectiveness in real indoor environments.
- 6 October 2025
- Published by Engineering, Technology & Applied Science Research in Engineering, Technology & Applied Science Research
- Vol. 15 (5) , 26935-26942
Open AccessA Safe 2D Grid Map for Indoor Wheelchair Navigation Using the LiDAR Camera System
Mobile robots are widely used in localization, mapping, and path planning, with map construction being essential for safe navigation. This study presents a method for creating a safe 2D grid map for indoor electric wheelchair navigation using a LiDAR-camera system. The process converts a 3D environmental map into a 2D grid map, helping the wheelchair detect obstacles at unsafe heights. First, a 3D environmental map is generated using the LiDAR-camera system of an iPhone 12 Pro, capturing 3D point clouds. The Iterative Closest Point (ICP) algorithm merges these point clouds into a complete 3D map. Accuracy is verified by comparing the distances between the real environment and the generated map, with deviations less than 2%. To enhance safety, points exceeding the wheelchair's height limit are removed, forming a refined 3D height map. This map is then projected into a 2D representation and overlaid with a grid, ensuring each cell matches the wheelchair's dimensions. A pixel density approximation refines the grid. The Q-learning method is applied for autonomous wheelchair navigation, proving the system's effectiveness in real indoor environments.
- Open Access
- 4 June 2025
- Published by Engineering, Technology & Applied Science Research in Engineering, Technology & Applied Science Research
- Vol. 15 (3) , 23499-23505
A Robust Approach for Breast Cancer Classification from DICOM Images
The number of breast cancer patients is rapidly increasing worldwide, with Asia currently accounting for 45% of global breast cancer cases. In addition, the number of breast cancer cases is expected to increase by 21.0%, and the mortality rate is projected to increase by 27.8% during the 2020-2030 period. This paper proposes a method for classifying breast cancer from Digital Imaging and Communications in Medicine (DICOM) images. In particular, an image segmentation technique was developed for extracting the breast region from DICOM images of varying sizes. The extracted images were then enhanced using multiple augmentation techniques to improve classification performance. Finally, a deep learning network was applied to classify breast cancer from the processed DICOM images. The VinDr-Mammo dataset was used to evaluate the effectiveness of the proposed method, and the experimental results showed an accuracy of 81.45%, demonstrating that the proposed approach is highly suitable for breast cancer detection and classification.
- 4 June 2025
- Published by Engineering, Technology & Applied Science Research in Engineering, Technology & Applied Science Research
- Vol. 15 (3) , 23499-23505
Open AccessA Robust Approach for Breast Cancer Classification from DICOM Images
The number of breast cancer patients is rapidly increasing worldwide, with Asia currently accounting for 45% of global breast cancer cases. In addition, the number of breast cancer cases is expected to increase by 21.0%, and the mortality rate is projected to increase by 27.8% during the 2020-2030 period. This paper proposes a method for classifying breast cancer from Digital Imaging and Communications in Medicine (DICOM) images. In particular, an image segmentation technique was developed for extracting the breast region from DICOM images of varying sizes. The extracted images were then enhanced using multiple augmentation techniques to improve classification performance. Finally, a deep learning network was applied to classify breast cancer from the processed DICOM images. The VinDr-Mammo dataset was used to evaluate the effectiveness of the proposed method, and the experimental results showed an accuracy of 81.45%, demonstrating that the proposed approach is highly suitable for breast cancer detection and classification.
- 26 March 2025
- Published by Thai Nguyen University in TNU Journal of Science and Technology
- Vol. 230 (06) , 183-190
A SIGN LANGUAGE IDENTIFICATION SYSTEM TO SUPPORT DISABILITY LANGUAGE USING SUPPORT VECTOR MACHINE
Giao tiếp giữa những người bị khuyết tật ngôn ngữ và những người không khuyết tật thường phải đối mặt với những thách thức đáng kể. Hiện nay, có rất nhiều người khuyết tật ngôn ngữ trên toàn thế giới và tại Việt Nam, do đó cần một giải pháp hữu ích để giúp cho người khuyết tật ngôn ngữ có thể giao tiếp dễ dàng hơn. Bài báo này đề xuất một hệ thống giúp những người khuyết tật ngôn ngữ giao tiếp dễ dàng hơn với những người không khuyết tật. Hệ thống được đề xuất bao gồm một găng tay cảm biến để đo tín hiệu chuyển động ngón tay. Các tín hiệu đo được sẽ được tiền xử lý để loại bỏ các nhiễu và artifacts. Sau đó, các tín hiệu đã xử lý được sử dụng để nhận dạng các chữ cái tương ứng với các cử chỉ. Cuối cùng, âm thanh tương ứng với các chữ cái nhận dạng được sẽ được phát qua loa. Hệ thống đạt được độ chính xác là 99,67% trong việc nhận dạng ngôn ngữ ký hiệu. Những kết quả đầy hứa hẹn này cho thấy hệ thống có thể được áp dụng trong các tình huống thực tế để hỗ trợ những người khuyết tật ngôn ngữ.
- 26 March 2025
- Published by Thai Nguyen University in TNU Journal of Science and Technology
- Vol. 230 (06) , 183-190
A SIGN LANGUAGE IDENTIFICATION SYSTEM TO SUPPORT DISABILITY LANGUAGE USING SUPPORT VECTOR MACHINE
Giao tiếp giữa những người bị khuyết tật ngôn ngữ và những người không khuyết tật thường phải đối mặt với những thách thức đáng kể. Hiện nay, có rất nhiều người khuyết tật ngôn ngữ trên toàn thế giới và tại Việt Nam, do đó cần một giải pháp hữu ích để giúp cho người khuyết tật ngôn ngữ có thể giao tiếp dễ dàng hơn. Bài báo này đề xuất một hệ thống giúp những người khuyết tật ngôn ngữ giao tiếp dễ dàng hơn với những người không khuyết tật. Hệ thống được đề xuất bao gồm một găng tay cảm biến để đo tín hiệu chuyển động ngón tay. Các tín hiệu đo được sẽ được tiền xử lý để loại bỏ các nhiễu và artifacts. Sau đó, các tín hiệu đã xử lý được sử dụng để nhận dạng các chữ cái tương ứng với các cử chỉ. Cuối cùng, âm thanh tương ứng với các chữ cái nhận dạng được sẽ được phát qua loa. Hệ thống đạt được độ chính xác là 99,67% trong việc nhận dạng ngôn ngữ ký hiệu. Những kết quả đầy hứa hẹn này cho thấy hệ thống có thể được áp dụng trong các tình huống thực tế để hỗ trợ những người khuyết tật ngôn ngữ.
- Open Access
- 13 February 2025
- Published by MDPI AG in Electronics
- Vol. 14 (4) , 734
Vision-Based Hand Gesture Recognition Using a YOLOv8n Model for the Navigation of a Smart Wheelchair
Electric wheelchairs are the primary means of transportation that enable individuals with disabilities to move independently to their desired locations. This paper introduces a novel, low-cost smart wheelchair system designed to enhance the mobility of individuals with severe disabilities through hand gesture recognition. Additionally, the system aims to support low-income individuals who previously lacked access to smart wheelchairs. Unlike existing methods that rely on expensive hardware or complex systems, the proposed system utilizes an affordable webcam and an Nvidia Jetson Nano embedded computer to process and recognize six distinct hand gestures—“Forward 1”, “Forward 2”, “Backward”, “Left”, “Right”, and “Stop”—to assist with wheelchair navigation. The system employs the “You Only Look Once version 8n” (YOLOv8n) model, which is well suited for low-spec embedded computers, trained on a self-collected hand gesture dataset containing 12,000 images. The pre-processing phase utilizes the MediaPipe library to generate landmark hand images, remove the background, and then extract the region of interest (ROI) of the hand gestures, significantly improving gesture recognition accuracy compared to previous methods that relied solely on hand images. Experimental results demonstrate impressive performance, achieving 99.3% gesture recognition accuracy and 93.8% overall movement accuracy in diverse indoor and outdoor environments. Furthermore, this paper presents a control circuit system that can be easily installed on any existing electric wheelchair. This approach offers a cost-effective, real-time solution that enhances the autonomy of individuals with severe disabilities in daily activities, laying the foundation for the development of affordable smart wheelchairs.
- 13 February 2025
- Published by MDPI AG in Electronics
- Vol. 14 (4) , 734
Open AccessVision-Based Hand Gesture Recognition Using a YOLOv8n Model for the Navigation of a Smart Wheelchair
Electric wheelchairs are the primary means of transportation that enable individuals with disabilities to move independently to their desired locations. This paper introduces a novel, low-cost smart wheelchair system designed to enhance the mobility of individuals with severe disabilities through hand gesture recognition. Additionally, the system aims to support low-income individuals who previously lacked access to smart wheelchairs. Unlike existing methods that rely on expensive hardware or complex systems, the proposed system utilizes an affordable webcam and an Nvidia Jetson Nano embedded computer to process and recognize six distinct hand gestures—“Forward 1”, “Forward 2”, “Backward”, “Left”, “Right”, and “Stop”—to assist with wheelchair navigation. The system employs the “You Only Look Once version 8n” (YOLOv8n) model, which is well suited for low-spec embedded computers, trained on a self-collected hand gesture dataset containing 12,000 images. The pre-processing phase utilizes the MediaPipe library to generate landmark hand images, remove the background, and then extract the region of interest (ROI) of the hand gestures, significantly improving gesture recognition accuracy compared to previous methods that relied solely on hand images. Experimental results demonstrate impressive performance, achieving 99.3% gesture recognition accuracy and 93.8% overall movement accuracy in diverse indoor and outdoor environments. Furthermore, this paper presents a control circuit system that can be easily installed on any existing electric wheelchair. This approach offers a cost-effective, real-time solution that enhances the autonomy of individuals with severe disabilities in daily activities, laying the foundation for the development of affordable smart wheelchairs.
- 24 December 2024
- Published by Springer Nature
- p. 273-285
The EMG Signal Acquisition System for DC Motor Control Based on Threshold Method
- 24 December 2024
- Published by Springer Nature
- p. 273-285
The EMG Signal Acquisition System for DC Motor Control Based on Threshold Method
- Open Access
- 1 July 2024
- Published by Institute of Advanced Engineering and Science in Computer Science and Information Technologies
- Vol. 5 (2) , 186-194
Acoustic echo cancellation system based on Laguerre method and neural network
Acoustic echo cancellation (AEC) is a fundamental requirement of signal processing to increase the quality of teleconferences. In this paper, a system that combines the Laguerre method with neural networks is proposed for AEC. In particular, the signal is processed using the Laguerre method to effectively handle nonlinear transmission line system. The results after applying the Laguerre method are then fed into a neural network for training and acoustic echo cancellation. The proposed system is tested on both linear and nonlinear transmission lines. Simulation results show that combining the Laguerre method with neural networks is highly effective for AEC in both linear and nonlinear transmission lines system. The AEC results obtained by the proposed method achieves a significant improvement in nonlinear transmission lines and it is the basis for building a practical echo cancellation system.
- 1 July 2024
- Published by Institute of Advanced Engineering and Science in Computer Science and Information Technologies
- Vol. 5 (2) , 186-194
Open AccessAcoustic echo cancellation system based on Laguerre method and neural network
Acoustic echo cancellation (AEC) is a fundamental requirement of signal processing to increase the quality of teleconferences. In this paper, a system that combines the Laguerre method with neural networks is proposed for AEC. In particular, the signal is processed using the Laguerre method to effectively handle nonlinear transmission line system. The results after applying the Laguerre method are then fed into a neural network for training and acoustic echo cancellation. The proposed system is tested on both linear and nonlinear transmission lines. Simulation results show that combining the Laguerre method with neural networks is highly effective for AEC in both linear and nonlinear transmission lines system. The AEC results obtained by the proposed method achieves a significant improvement in nonlinear transmission lines and it is the basis for building a practical echo cancellation system.
- 19 October 2023
- Published by Springer Nature
- p. 857-867
Fabrication of Ferritin Nanoparticle Patterns for Controlling Neuron Adhesion and Neurite Outgrowth
- 19 October 2023
- Published by Springer Nature
- p. 857-867
Fabrication of Ferritin Nanoparticle Patterns for Controlling Neuron Adhesion and Neurite Outgrowth