Van-Thanh Hoang
Research Subjects
Van-Thanh Hoang has a total of 15 co-authors. Here are the top co-authors by number of shared publications:
- Kang-Hyun Jo
University of Ulsan
17 shared publications - Tu Minh Phuong
Posts and Telecommunications Institute of Technology, Hanoi, Vietnam
3 shared publications Nguyen Duy Quang
2 shared publicationsTu Minh Phuong
2 shared publicationsVu Huu Dao
Posts and Telecommunications Institute of Technology,Faculty of Information Technology,Hanoi,Vietnam
2 shared publications- Ajmal Shahbaz
Centre for Machine Vision, University of the West of England (UWE), Bristol BS16 1QY, UK
1 shared publication
Latest Publications
- Open Access
- 22 September 2026
- Published by Universitas Mercu Buana in SINERGI
- Vol. 30 (3) , 943-960
WS-Det: an efficient whale shark detection via integrated feature modulation module
Whale sharks are vital plankton feeders that help maintain marine ecosystem stability and serve as key ecotourism assets supporting Gorontalo's blue economy. Reliable detection and identification systems are essential for monitoring their habitats and behaviors. Vision-based methods offer non-invasive observation without physical contact or electromagnetic interference. YOLOv12, a state-of-the-art object detector, achieves high localization accuracy with its lightweight "nano" variant. However, lightweight models often struggle with complex feature discrimination, which can be mitigated by integrating enhancement blocks to boost detection performance while maintaining computational efficiency. In this paper, we introduce a novel vision-based whale shark detection framework (WS-Det), which enhances YOLOv12-nano with MOGA attention modules. The proposed model incorporates lightweight convolutional operations to ensure efficient feature extraction while keeping computational cost low. The Integrated Feature Modulation (IFM) module improves backbone performance by selectively emphasizing spatially salient regions while interacting with multi-spatial-frequency information, thereby improving the overall robustness of feature learning. To support this study, we also present a dedicated whale shark dataset comprising underwater and aerial (drone-based) images for direct and practical monitoring. Experimental evaluations demonstrate that the WS-Det outperforms the baseline YOLOv12-nano and other efficient detectors, achieving higher detection precision. On the proposed dataset, the model achieves an mAP@0.5 of 0.979 and an mAP@0.5:0.95 of 0.868. It contains 2,191,459 parameters and requires 6.3 GFLOPs, indicating a low computational cost. Inference speed evaluations show the model runs at 17.19 FPS on a desktop PC and 6.64 FPS on a Raspberry Pi 5, confirming its suitability for deployment on resource-constrained devices.
- 22 September 2026
- Published by Universitas Mercu Buana in SINERGI
- Vol. 30 (3) , 943-960
Open AccessWS-Det: an efficient whale shark detection via integrated feature modulation module
Whale sharks are vital plankton feeders that help maintain marine ecosystem stability and serve as key ecotourism assets supporting Gorontalo's blue economy. Reliable detection and identification systems are essential for monitoring their habitats and behaviors. Vision-based methods offer non-invasive observation without physical contact or electromagnetic interference. YOLOv12, a state-of-the-art object detector, achieves high localization accuracy with its lightweight "nano" variant. However, lightweight models often struggle with complex feature discrimination, which can be mitigated by integrating enhancement blocks to boost detection performance while maintaining computational efficiency. In this paper, we introduce a novel vision-based whale shark detection framework (WS-Det), which enhances YOLOv12-nano with MOGA attention modules. The proposed model incorporates lightweight convolutional operations to ensure efficient feature extraction while keeping computational cost low. The Integrated Feature Modulation (IFM) module improves backbone performance by selectively emphasizing spatially salient regions while interacting with multi-spatial-frequency information, thereby improving the overall robustness of feature learning. To support this study, we also present a dedicated whale shark dataset comprising underwater and aerial (drone-based) images for direct and practical monitoring. Experimental evaluations demonstrate that the WS-Det outperforms the baseline YOLOv12-nano and other efficient detectors, achieving higher detection precision. On the proposed dataset, the model achieves an mAP@0.5 of 0.979 and an mAP@0.5:0.95 of 0.868. It contains 2,191,459 parameters and requires 6.3 GFLOPs, indicating a low computational cost. Inference speed evaluations show the model runs at 17.19 FPS on a desktop PC and 6.64 FPS on a Raspberry Pi 5, confirming its suitability for deployment on resource-constrained devices.
- 21 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
An Efficient Object Detection Model Based on NanoDet
Object detection is a fundamental task in computer vision, involving the prediction of bounding boxes and class labels for Regions of Interest (ROI) within images. Traditionally, anchor-based detectors have dominated this field, leveraging predefined anchor boxes for object localization and classification. However, anchor-free detectors have emerged as a compelling alternative due to their reduced computational complexity and efficient detection capabilities. Among these innovations, Nan-oDet has garnered significant attention in the AI community for its speed and lightweight design as an anchor-free object detection model. This study introduces an enhanced variant of NanoDet, leveraging the EfficientNetB0-compact model as its backbone. By replacing the original ShuffleNetV2 backbone, this integration aims to improve detection performance without compromising speed, thereby enhancing suitability for mobile devices. Experimental results demonstrate the effectiveness of this modification in achieving superior performance metrics, highlighting the potential of advanced network architectures in optimizing object detection tasks for mobile and edge computing environments.
- 21 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
An Efficient Object Detection Model Based on NanoDet
Object detection is a fundamental task in computer vision, involving the prediction of bounding boxes and class labels for Regions of Interest (ROI) within images. Traditionally, anchor-based detectors have dominated this field, leveraging predefined anchor boxes for object localization and classification. However, anchor-free detectors have emerged as a compelling alternative due to their reduced computational complexity and efficient detection capabilities. Among these innovations, Nan-oDet has garnered significant attention in the AI community for its speed and lightweight design as an anchor-free object detection model. This study introduces an enhanced variant of NanoDet, leveraging the EfficientNetB0-compact model as its backbone. By replacing the original ShuffleNetV2 backbone, this integration aims to improve detection performance without compromising speed, thereby enhancing suitability for mobile devices. Experimental results demonstrate the effectiveness of this modification in achieving superior performance metrics, highlighting the potential of advanced network architectures in optimizing object detection tasks for mobile and edge computing environments.
- 16 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Enhanced Breast Cancer Detection: A Transfer Learning Framework with Auxiliary Classification for Mammographic Image Analysis
Breast cancer remains one of the most prevalent malignancies affecting women worldwide, with over 2.3 million new cases diagnosed annually. While conventional diagnostic methods like mammography have reduced mortality rates, artificial intelligence offers opportunities to further enhance detection accuracy. In this paper, we present a comprehensive framework for breast cancer identification using deep learning techniques applied to mammographic images. Our approach leverages transfer learning from ImageNet pretrained models and incorporates an auxiliary classification mechanism to improve diagnostic performance. We evaluated multiple state-of-the-art convolutional neural network architectures, including MobileNetV3, ResNet50, ConvNeXtV2-nano, ResNeXt50, MobileViT-small, and EfficientNet-B3, trained on four publicly available datasets (BMCD, CDD-CESM, CMMD, and MiniDDSM). Performance was assessed on separate external test datasets (VinDr and RSNA) to simulate real-world clinical deployment scenarios. Our experimental results demonstrate that EfficientNet-B3 with auxiliary classification achieved superior performance across most metrics, with an accuracy of 86.14%, F1-score of 85.80%, and AUC of 90.39% on validation data. We further introduce the Probabilistic F1-score as a clinically relevant evaluation metric that accounts for prediction confidence rather than binary decisions alone. The proposed framework, available as an open-source implementation, offers a promising approach for enhancing breast cancer detection while providing insights into the challenges of deploying AI systems in diverse clinical environments. We released our codebase at: https://github.com/thanhhnvnqb/mmbreast
- 16 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Enhanced Breast Cancer Detection: A Transfer Learning Framework with Auxiliary Classification for Mammographic Image Analysis
Breast cancer remains one of the most prevalent malignancies affecting women worldwide, with over 2.3 million new cases diagnosed annually. While conventional diagnostic methods like mammography have reduced mortality rates, artificial intelligence offers opportunities to further enhance detection accuracy. In this paper, we present a comprehensive framework for breast cancer identification using deep learning techniques applied to mammographic images. Our approach leverages transfer learning from ImageNet pretrained models and incorporates an auxiliary classification mechanism to improve diagnostic performance. We evaluated multiple state-of-the-art convolutional neural network architectures, including MobileNetV3, ResNet50, ConvNeXtV2-nano, ResNeXt50, MobileViT-small, and EfficientNet-B3, trained on four publicly available datasets (BMCD, CDD-CESM, CMMD, and MiniDDSM). Performance was assessed on separate external test datasets (VinDr and RSNA) to simulate real-world clinical deployment scenarios. Our experimental results demonstrate that EfficientNet-B3 with auxiliary classification achieved superior performance across most metrics, with an accuracy of 86.14%, F1-score of 85.80%, and AUC of 90.39% on validation data. We further introduce the Probabilistic F1-score as a clinically relevant evaluation metric that accounts for prediction confidence rather than binary decisions alone. The proposed framework, available as an open-source implementation, offers a promising approach for enhancing breast cancer detection while providing insights into the challenges of deploying AI systems in diverse clinical environments. We released our codebase at: https://github.com/thanhhnvnqb/mmbreast
- 15 December 2024
- Published by Elsevier in Neurocomputing
- Vol. 620
A compact version of EfficientNet for skin disease diagnosis application
- 15 December 2024
- Published by Elsevier in Neurocomputing
- Vol. 620
A compact version of EfficientNet for skin disease diagnosis application
- Top Cited PapersOpen Access
- 28 February 2024
- Published by Ho Chi Minh City University of Technology and Engineering in Journal of Technical Education Science
- Vol. 19 (1) , 64-73
Impact of Integrated Artificial Intelligence and Internet of Things Technologies on Smart City Transformation
Rapid urbanization is placing tremendous pressure on limited resources and aging infrastructure in cities worldwide. Meanwhile, new technologies are emerging to help address urban challenges through data-driven solutions. This paper explores how the strategic integration of artificial intelligence (AI) and Internet of Things (IoT) can transform urban management and services delivery for smart and sustainable cities. The Internet of Things enables the ubiquitous collection of real-time data across urban systems through embedded sensors. However, extracting actionable insights requires advanced analytics. Concurrently, artificial intelligence provides techniques to autonomously analyze huge volumes of IoT-sensed urban data. When combined effectively, AI and IoT can automatically monitor infrastructure, optimize operations, and enhance citizen experiences. This paper first defines key concepts and outlines applications of AI and IoT independently in areas like transportation, energy, environment, and public safety. It then examines how both technologies can be integrated for mutual benefit. Examples of integrated solutions such as predictive maintenance, intelligent transportation, and emergency response optimization are discussed. Challenges to adoption like data privacy, infrastructure costs, skills gaps, and technical standardization are also covered. The conclusion underscores the tremendous potential of AI and IoT to create efficient, resilient and livable urban environments through ubiquitous sensing and autonomous management. With proper policy support and collaborations, cities worldwide can leverage these smart technologies to sustainably combat problems facing urbanization.
- 28 February 2024
- Published by Ho Chi Minh City University of Technology and Engineering in Journal of Technical Education Science
- Vol. 19 (1) , 64-73
Top Cited PapersOpen AccessImpact of Integrated Artificial Intelligence and Internet of Things Technologies on Smart City Transformation
Rapid urbanization is placing tremendous pressure on limited resources and aging infrastructure in cities worldwide. Meanwhile, new technologies are emerging to help address urban challenges through data-driven solutions. This paper explores how the strategic integration of artificial intelligence (AI) and Internet of Things (IoT) can transform urban management and services delivery for smart and sustainable cities. The Internet of Things enables the ubiquitous collection of real-time data across urban systems through embedded sensors. However, extracting actionable insights requires advanced analytics. Concurrently, artificial intelligence provides techniques to autonomously analyze huge volumes of IoT-sensed urban data. When combined effectively, AI and IoT can automatically monitor infrastructure, optimize operations, and enhance citizen experiences. This paper first defines key concepts and outlines applications of AI and IoT independently in areas like transportation, energy, environment, and public safety. It then examines how both technologies can be integrated for mutual benefit. Examples of integrated solutions such as predictive maintenance, intelligent transportation, and emergency response optimization are discussed. Challenges to adoption like data privacy, infrastructure costs, skills gaps, and technical standardization are also covered. The conclusion underscores the tremendous potential of AI and IoT to create efficient, resilient and livable urban environments through ubiquitous sensing and autonomous management. With proper policy support and collaborations, cities worldwide can leverage these smart technologies to sustainably combat problems facing urbanization.
- 9 August 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Modern Deep Learning Models for Skin Disease Diagnosis
Skin cancer is a pressing global health concern, representing a significant challenge in terms of achieving accurate and timely diagnoses. The introduction of deep learning algorithms has heralded a new era of progress across various domains, particularly in the domain of skin disease diagnosis. In our research, we embark on a journey to meticulously evaluate the effectiveness of state-of-the-art deep learning models, utilizing the comprehensive HAM10000 dataset. Our overarching objective is to identify the foremost model for the diagnosis of skin cancer, leveraging the immense potential of modern machine learning techniques. By conducting a comprehensive and in-depth analysis of these models, our aim transcends the mere pursuit of technical excellence. We seek to make a substantial and meaningful contribution to the field of healthcare, particularly in the critical area of skin disease diagnosis. Our vision is to enhance the precision, reliability, and efficiency of diagnostic methodologies, ultimately leading to earlier detection, more effective treatments, and improved outcomes for patients grappling with skin cancer.
- 9 August 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Modern Deep Learning Models for Skin Disease Diagnosis
Skin cancer is a pressing global health concern, representing a significant challenge in terms of achieving accurate and timely diagnoses. The introduction of deep learning algorithms has heralded a new era of progress across various domains, particularly in the domain of skin disease diagnosis. In our research, we embark on a journey to meticulously evaluate the effectiveness of state-of-the-art deep learning models, utilizing the comprehensive HAM10000 dataset. Our overarching objective is to identify the foremost model for the diagnosis of skin cancer, leveraging the immense potential of modern machine learning techniques. By conducting a comprehensive and in-depth analysis of these models, our aim transcends the mere pursuit of technical excellence. We seek to make a substantial and meaningful contribution to the field of healthcare, particularly in the critical area of skin disease diagnosis. Our vision is to enhance the precision, reliability, and efficiency of diagnostic methodologies, ultimately leading to earlier detection, more effective treatments, and improved outcomes for patients grappling with skin cancer.
- 15 December 2022
- Published by Springer Nature
- p. 435-445
Rethinking Mobile Inverted Bottleneck Convolution for EfficientNet
- 15 December 2022
- Published by Springer Nature
- p. 435-445
Rethinking Mobile Inverted Bottleneck Convolution for EfficientNet
- 19 November 2020
- Published by Springer Nature
- p. 166-174
Simple Pose Network with Skip-Connections for Single Human Pose Estimation
- 19 November 2020
- Published by Springer Nature
- p. 166-174
Simple Pose Network with Skip-Connections for Single Human Pose Estimation
- 4 March 2020
- Published by Springer Nature
- p. 163-173
Slice Operator for Efficient Convolutional Neural Network Architecture
- 4 March 2020
- Published by Springer Nature
- p. 163-173
Slice Operator for Efficient Convolutional Neural Network Architecture