Giang Son Tran
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Giang Son Tran has a total of 33 co-authors. Here are the top co-authors by number of shared publications:
- Thi Phuong Nghiem
Department of Information and Communication Technology, University of Science and Technology of Hanoi, Vietnam Academy of Science and Technology, Hanoi, Vietnam
5 shared publications - Jean-Christophe Burie
La Rochelle Universite
3 shared publications - Ngoc Ha Pham
Information and Communication Technology Department, FPT University, Vietnam
3 shared publications - Antoine Doucet
University of La Rochelle, La Rochelle, France
2 shared publications - Chi Cuong Nguyen
University of Science and Technology of Hanoi, Vietnam
2 shared publications - Oanh Cuong Do
Thuyloi University
2 shared publications
Latest Publications
- Open Access
- 2 October 2026
- Published by Elsevier in SSRN Electronic Journal
Optimizing Wavelengths and Vegetation Indices for Estimating Relative Chlorophyll Content in Tea Bud Leaves using UAV-Based Hyperspectral Imagery
Accurate chlorophyll estimation in tea bud leaves supports plant health assessment, fertilization management, and yield optimization. UAV hyperspectral imagery provides detailed spectral information to capture changes in leaf chlorophyll content. However, traditional vegetation indices (VIs) often use fixed wavelengths and using many VIs in machine learning models can increase model complexity and decrease interpretability. Therefore, this study aimed to optimize wavelength combinations for individual VIs and identify the most informative indices for Soil Plant Analysis Development (SPAD) estimation in tea leaves. Hyperspectral UAV imagery spanning 400–1000 nm with 122 contiguous bands was collected over a tea field in Phu Tho Province, Vietnam. Optimal wavelength combinations were identified for each VI and evaluated using regression and machine learning models. The VIs were ranked based on their importance and SHAP analysis was applied to explain their contributions to XGBoost model. VIs using the same spectral region did not necessarily share the same optimal wavelengths. XGBoost using the optimized VIs achieved the best performance, with an R² of 0.630, RMSE of 0.027, and MAE of 0.021 using top 3 VIs. Feature ranking showed that only three indices, Visible–band Difference Vegetation Index (VDVI), The MERIS terrestrial chlorophyll index (MTCI), and Excess Green Index (EXG), were sufficient to achieve comparable prediction accuracy with 18 VIs while reducing the number of predictors.
- 2 October 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessOptimizing Wavelengths and Vegetation Indices for Estimating Relative Chlorophyll Content in Tea Bud Leaves using UAV-Based Hyperspectral Imagery
Accurate chlorophyll estimation in tea bud leaves supports plant health assessment, fertilization management, and yield optimization. UAV hyperspectral imagery provides detailed spectral information to capture changes in leaf chlorophyll content. However, traditional vegetation indices (VIs) often use fixed wavelengths and using many VIs in machine learning models can increase model complexity and decrease interpretability. Therefore, this study aimed to optimize wavelength combinations for individual VIs and identify the most informative indices for Soil Plant Analysis Development (SPAD) estimation in tea leaves. Hyperspectral UAV imagery spanning 400–1000 nm with 122 contiguous bands was collected over a tea field in Phu Tho Province, Vietnam. Optimal wavelength combinations were identified for each VI and evaluated using regression and machine learning models. The VIs were ranked based on their importance and SHAP analysis was applied to explain their contributions to XGBoost model. VIs using the same spectral region did not necessarily share the same optimal wavelengths. XGBoost using the optimized VIs achieved the best performance, with an R² of 0.630, RMSE of 0.027, and MAE of 0.021 using top 3 VIs. Feature ranking showed that only three indices, Visible–band Difference Vegetation Index (VDVI), The MERIS terrestrial chlorophyll index (MTCI), and Excess Green Index (EXG), were sufficient to achieve comparable prediction accuracy with 18 VIs while reducing the number of predictors.
- Open Access
- 5 March 2026
- Published by Association for Computing Machinery (ACM)
- p. 376-380
Chlorophyll Content Estimation in Rice Leaves using Spectral Reflectance with Gradient Boosted Multilayer Perceptron
- 5 March 2026
- Published by Association for Computing Machinery (ACM)
- p. 376-380
Open AccessChlorophyll Content Estimation in Rice Leaves using Spectral Reflectance with Gradient Boosted Multilayer Perceptron
- Open Access
- 5 March 2026
- Published by Association for Computing Machinery (ACM)
- p. 543-551
Optimizing Vision Transformers (ViT) for Pneumonia Diagnosis via CLAHE Preprocessing and GAN-Based Augmentation
Early identification of pneumonia through chest X-ray interpretation is a vital component of modern diagnostic workflows. While deep learning models—especially Vision Transformers (ViTs)—have demonstrated impressive results in medical image classification, their performance is highly dependent on access to large-scale and diverse datasets to prevent overfitting and ensure robustness. The commonly used Kermany dataset, however, presents challenges such as class distribution imbalance and limited image variability. To address these issues, we introduce a comprehensive classification framework that integrates Contrast Limited Adaptive Histogram Equalization (CLAHE) for image enhancement, data augmentation via a Deep Convolutional Generative Adversarial Network (DCGAN), and a fine-tuned ViT-B/16 model. Additionally, we incorporate a metaheuristic optimization strategy to systematically adjust critical hyperparameters such as learning rate, weight decay, and augmentation proportions. Experimental evaluations show that our approach achieves a classification accuracy of 96.6%, surpassing the baseline ViT and traditional convolutional neural network (CNN) models. An ablation study highlights the individual contributions of CLAHE, GAN-based augmentation, and hyperparameter tuning, demonstrating the effectiveness of the proposed system. This method offers a scalable and reliable solution suitable for clinical applications and deployment in healthcare-oriented AI systems.
- 5 March 2026
- Published by Association for Computing Machinery (ACM)
- p. 543-551
Open AccessOptimizing Vision Transformers (ViT) for Pneumonia Diagnosis via CLAHE Preprocessing and GAN-Based Augmentation
Early identification of pneumonia through chest X-ray interpretation is a vital component of modern diagnostic workflows. While deep learning models—especially Vision Transformers (ViTs)—have demonstrated impressive results in medical image classification, their performance is highly dependent on access to large-scale and diverse datasets to prevent overfitting and ensure robustness. The commonly used Kermany dataset, however, presents challenges such as class distribution imbalance and limited image variability. To address these issues, we introduce a comprehensive classification framework that integrates Contrast Limited Adaptive Histogram Equalization (CLAHE) for image enhancement, data augmentation via a Deep Convolutional Generative Adversarial Network (DCGAN), and a fine-tuned ViT-B/16 model. Additionally, we incorporate a metaheuristic optimization strategy to systematically adjust critical hyperparameters such as learning rate, weight decay, and augmentation proportions. Experimental evaluations show that our approach achieves a classification accuracy of 96.6%, surpassing the baseline ViT and traditional convolutional neural network (CNN) models. An ablation study highlights the individual contributions of CLAHE, GAN-based augmentation, and hyperparameter tuning, demonstrating the effectiveness of the proposed system. This method offers a scalable and reliable solution suitable for clinical applications and deployment in healthcare-oriented AI systems.
- 11 June 2024
- Published by Thai Nguyen University in TNU Journal of Science and Technology
- Vol. 229 (07) , 133-140
RICE GRAIN TRAIT ESTIMATION USING COLOR SPACE CONVERSION AND DEEP LEARNING-BASED IMAGE SEGMENTATION
Việc trích chọn đặc điểm kiểu hình của hạt lúa một cách chính xác là việc rất quan trọng trong việc quản lý và ước lượng năng suất trồng lúa một cách hiệu quả, đồng thời mang lại những hiểu biết quý giá để cải tiến các phương pháp nông nghiệp. Tuy nhiên, thực hiện thủ công các công việc này là rất tốn công sức, tốn thời gian và dễ gây sai sót. Nghiên cứu này đề xuất một phương pháp mới sử dụng máy ảnh kỹ thuật số giá rẻ và công nghệ học sâu để đếm và trích chọn các đặc điểm kiểu hình của hạt lúa. Nghiên cứu của chúng tôi giới thiệu một bước tiền xử lý để phân tách các vùng hạt lúa từ nền ảnh đầu vào bằng cách chuyển đổi không gian màu. Sau đó, một mô hình phân đoạn ảnh dựa trên học sâu dùng YOLOv8 được sử dụng để đếm số lượng và trích chọn các đặc điểm kiểu hình của hạt lúa. Độ chính xác của phương pháp đề xuất được thử nghiệm trên 88 giống lúa khác nhau được cung cấp bởi Trung tâm Tài nguyên Thực vật tại Hà Nội. Kết quả thực nghiệm cho thấy phương pháp đề xuất có độ chính xác cao và có khả năng xử lý lượng lớn ảnh màu với chi phí thấp để trích chọn đặc điểm kiểu hình của hạt lúa. Kết quả này có tiềm năng trong việc hỗ trợ các chương trình lai tạo giống lúa và xác định các gene chức năng của các giống lúa.
- 11 June 2024
- Published by Thai Nguyen University in TNU Journal of Science and Technology
- Vol. 229 (07) , 133-140
RICE GRAIN TRAIT ESTIMATION USING COLOR SPACE CONVERSION AND DEEP LEARNING-BASED IMAGE SEGMENTATION
Việc trích chọn đặc điểm kiểu hình của hạt lúa một cách chính xác là việc rất quan trọng trong việc quản lý và ước lượng năng suất trồng lúa một cách hiệu quả, đồng thời mang lại những hiểu biết quý giá để cải tiến các phương pháp nông nghiệp. Tuy nhiên, thực hiện thủ công các công việc này là rất tốn công sức, tốn thời gian và dễ gây sai sót. Nghiên cứu này đề xuất một phương pháp mới sử dụng máy ảnh kỹ thuật số giá rẻ và công nghệ học sâu để đếm và trích chọn các đặc điểm kiểu hình của hạt lúa. Nghiên cứu của chúng tôi giới thiệu một bước tiền xử lý để phân tách các vùng hạt lúa từ nền ảnh đầu vào bằng cách chuyển đổi không gian màu. Sau đó, một mô hình phân đoạn ảnh dựa trên học sâu dùng YOLOv8 được sử dụng để đếm số lượng và trích chọn các đặc điểm kiểu hình của hạt lúa. Độ chính xác của phương pháp đề xuất được thử nghiệm trên 88 giống lúa khác nhau được cung cấp bởi Trung tâm Tài nguyên Thực vật tại Hà Nội. Kết quả thực nghiệm cho thấy phương pháp đề xuất có độ chính xác cao và có khả năng xử lý lượng lớn ảnh màu với chi phí thấp để trích chọn đặc điểm kiểu hình của hạt lúa. Kết quả này có tiềm năng trong việc hỗ trợ các chương trình lai tạo giống lúa và xác định các gene chức năng của các giống lúa.
- Open Access
- 16 May 2024
- Published by Publishing House for Science and Technology, Vietnam Academy of Science and Technology (Publications) in Journal of Computer Science and Cybernetics
- Vol. 40 (2) , 135-146
A NEW METHOD FOR MEDICAL IMAGE FUSION BASED ON GAUSSIAN BLUR FILTER AND ROBINSON COMPASS OPERATOR
Medical image fusion is a process of extracting features from multi-modal medical images and combining them into a composite image. It brings huge support in medical imaging and clinical diagnosis. However, the extraction of both structural and functional information from input MRI and PET images using multi-scale transform fusion methods poses a challenge of providing high-quality decomposition layers since during the decomposition process, images can still lose information such as blur or noise at the edges of the image. To address this limitation, we present a new method to improve the visual information fidelity of medical image fusion. Firstly, the YCbCr color space is utilized to prevent distortion when merging color and grey images. The second algorithm uses the CLAHE model, which allows the input images to have good contrast. Then, a Gaussian blur filter is employed to decompose the images into base and detail layers. The use of Gaussian blur ensures a smoothing filter of the edges. After that, the Robinson compass operator is applied to create the fusion rule of detail components. Finally, the fused base and detail layers are concatenated together to form the final composite image. The experimental results show that the proposed approach outperforms the latest methods in bringing visual information fidelity of the input images to the fused image, which is helpful in supporting doctors and radiologists in visual analysis of the medical images.
- 16 May 2024
- Published by Publishing House for Science and Technology, Vietnam Academy of Science and Technology (Publications) in Journal of Computer Science and Cybernetics
- Vol. 40 (2) , 135-146
Open AccessA NEW METHOD FOR MEDICAL IMAGE FUSION BASED ON GAUSSIAN BLUR FILTER AND ROBINSON COMPASS OPERATOR
Medical image fusion is a process of extracting features from multi-modal medical images and combining them into a composite image. It brings huge support in medical imaging and clinical diagnosis. However, the extraction of both structural and functional information from input MRI and PET images using multi-scale transform fusion methods poses a challenge of providing high-quality decomposition layers since during the decomposition process, images can still lose information such as blur or noise at the edges of the image. To address this limitation, we present a new method to improve the visual information fidelity of medical image fusion. Firstly, the YCbCr color space is utilized to prevent distortion when merging color and grey images. The second algorithm uses the CLAHE model, which allows the input images to have good contrast. Then, a Gaussian blur filter is employed to decompose the images into base and detail layers. The use of Gaussian blur ensures a smoothing filter of the edges. After that, the Robinson compass operator is applied to create the fusion rule of detail components. Finally, the fused base and detail layers are concatenated together to form the final composite image. The experimental results show that the proposed approach outperforms the latest methods in bringing visual information fidelity of the input images to the fused image, which is helpful in supporting doctors and radiologists in visual analysis of the medical images.
- Open Access
- 10 May 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Asian and Low-Resource Language Information Processing
- Vol. 23 (5) , 1-16
A Novel Pretrained General-purpose Vision Language Model for the Vietnamese Language
Lying in the cross-section of computer vision and natural language processing, vision language models are capable of processing images and text at once. These models are helpful in various tasks: text generation from image and vice versa, image-text retrieval, or visual navigation. Besides building a model trained on a dataset for a task, people also study general-purpose models to utilize many datasets for multitasks. Their two primary applications are image captioning and visual question answering. For English, large datasets and foundation models are already abundant. However, for Vietnamese, they are still limited. To expand the language range, this work proposes a pretrained general-purpose image-text model named VisualRoBERTa. A dataset of 600k images with captions (translated MS COCO 2017 from English to Vietnamese) is introduced to pretrain VisualRoBERTa. The model’s architecture is built using Convolutional Neural Network and Transformer blocks. Fine-tuning VisualRoBERTa shows promising results on the ViVQA dataset with 34.49% accuracy, 0.4173 BLEU 4, and 0.4390 RougeL (in visual question answering task), and best outcomes on the sViIC dataset with 0.6685 BLEU 4, 0.6320 RougeL (in image captioning task).
- 10 May 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Asian and Low-Resource Language Information Processing
- Vol. 23 (5) , 1-16
Open AccessA Novel Pretrained General-purpose Vision Language Model for the Vietnamese Language
Lying in the cross-section of computer vision and natural language processing, vision language models are capable of processing images and text at once. These models are helpful in various tasks: text generation from image and vice versa, image-text retrieval, or visual navigation. Besides building a model trained on a dataset for a task, people also study general-purpose models to utilize many datasets for multitasks. Their two primary applications are image captioning and visual question answering. For English, large datasets and foundation models are already abundant. However, for Vietnamese, they are still limited. To expand the language range, this work proposes a pretrained general-purpose image-text model named VisualRoBERTa. A dataset of 600k images with captions (translated MS COCO 2017 from English to Vietnamese) is introduced to pretrain VisualRoBERTa. The model’s architecture is built using Convolutional Neural Network and Transformer blocks. Fine-tuning VisualRoBERTa shows promising results on the ViVQA dataset with 34.49% accuracy, 0.4173 BLEU 4, and 0.4390 RougeL (in visual question answering task), and best outcomes on the sViIC dataset with 0.6685 BLEU 4, 0.6320 RougeL (in image captioning task).
- Open Access
- 23 February 2024
- Published by Association for Computing Machinery (ACM)
- p. 186-196
Overview of fine-tuning CNN-Based Models for X-ray Image Classification
A lung infection is usually the cause of pneumonia, a common medical condition. It irritates the lungs’ tissues and reduces their functionality. The severity of pneumonia can vary from a minor illness to a serious one. Identifying the exact infection causing the problem can be difficult. Diagnosis is often based on symptoms and physical examination, sometimes with chest X-rays. On the other hand, reviewing chest X-rays is a challenging and subjective task. In this work, we improve the CNN architecture to improve the X-ray image classification score performance. The objective of this study is to evaluate the fine-tuning of ResNet50V2. The ensemble technique that has been recommended yields very strong classification results, outperforming other models with an improvement of almost 97% in accuracy.
- 23 February 2024
- Published by Association for Computing Machinery (ACM)
- p. 186-196
Open AccessOverview of fine-tuning CNN-Based Models for X-ray Image Classification
A lung infection is usually the cause of pneumonia, a common medical condition. It irritates the lungs’ tissues and reduces their functionality. The severity of pneumonia can vary from a minor illness to a serious one. Identifying the exact infection causing the problem can be difficult. Diagnosis is often based on symptoms and physical examination, sometimes with chest X-rays. On the other hand, reviewing chest X-rays is a challenging and subjective task. In this work, we improve the CNN architecture to improve the X-ray image classification score performance. The objective of this study is to evaluate the fine-tuning of ResNet50V2. The ensemble technique that has been recommended yields very strong classification results, outperforming other models with an improvement of almost 97% in accuracy.
- Open Access
- 1 January 2024
- Published by Engineering and Technology Publishing in Journal of Advances in Information Technology
- Vol. 15 (11) , 1205-1214
Apply a CNN-Based Ensemble Model to Chest-X Ray Image-Based Pneumonia Classification
JAIT is a scientific open access journal which focuses on empirical research results and critical analysis of technology development, use, management and impacts in information technology.
- 1 January 2024
- Published by Engineering and Technology Publishing in Journal of Advances in Information Technology
- Vol. 15 (11) , 1205-1214
Open AccessApply a CNN-Based Ensemble Model to Chest-X Ray Image-Based Pneumonia Classification
JAIT is a scientific open access journal which focuses on empirical research results and critical analysis of technology development, use, management and impacts in information technology.
- 14 December 2023
- Published by Springer Nature in Earth Science Informatics
- Vol. 17 (1) , 527-538
Hyperspectral image classification using an encoder-decoder model with depthwise separable convolution, squeeze and excitation blocks
- 14 December 2023
- Published by Springer Nature in Earth Science Informatics
- Vol. 17 (1) , 527-538
Hyperspectral image classification using an encoder-decoder model with depthwise separable convolution, squeeze and excitation blocks
- Open Access
- 20 September 2023
- Published by Association for Computing Machinery (ACM)
- p. 199-202
False Positive Reduction of Pulmonary Nodule on CT image using Attention-based Multiple Instance Learning
- 20 September 2023
- Published by Association for Computing Machinery (ACM)
- p. 199-202
Open AccessFalse Positive Reduction of Pulmonary Nodule on CT image using Attention-based Multiple Instance Learning