Dinh Tuan Tran

ORCID: 0000-0001-7443-9102
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Robotics and Sensor-Based Localization
  • Video Surveillance and Tracking Methods
  • Human Pose and Action Recognition
  • Robotics and Automated Systems
  • Modular Robots and Swarm Intelligence
  • Robotic Path Planning Algorithms
  • Context-Aware Activity Recognition Systems
  • Water Quality Monitoring Technologies
  • Virtual Reality Applications and Impacts
  • Underwater Vehicles and Communication Systems
  • Medical Image Segmentation Techniques
  • Face recognition and analysis
  • Advanced Vision and Imaging
  • Advanced Image and Video Retrieval Techniques
  • Stroke Rehabilitation and Recovery
  • Hand Gesture Recognition Systems
  • Control and Dynamics of Mobile Robots
  • Anomaly Detection Techniques and Applications
  • Gait Recognition and Analysis
  • Multimodal Machine Learning Applications
  • Surgical Simulation and Training
  • Indoor and Outdoor Localization Technologies
  • Optimization and Search Problems
  • Advanced Manufacturing and Logistics Optimization
  • Time Series Analysis and Forecasting

Ritsumeikan University
2015-2025

Curtin University
2005-2006

Abstract Gait analysis has been studied for a long time and applied to fields such as security, sport, medicine. In particular, clinical gait played significant role in improving the quality of healthcare. With growth machine learning technology recent years, deep learning-based approaches have become popular. However, large number samples are required training models when using learning, where amount available gait-related data may be limited several reasons. This paper discusses certain...

10.1093/jcde/qwab054 article EN cc-by-nc Journal of Computational Design and Engineering 2021-09-12

Traditional methods for visualizing dynamic human expressions, particularly in medical training, often rely on flat-screen displays or static mannequins, which have proven inefficient realistic simulation. In response, we propose a platform that leverages 3D interactive facial avatar capable of displaying non-verbal feedback, including pain signals. This is projected onto stereoscopic, view-dependent display, offering more immersive and simulated patient experience assessment practice....

10.48550/arxiv.2502.08085 preprint EN arXiv (Cornell University) 2025-02-11

10.1109/iceic64972.2025.10879770 article EN 2020 International Conference on Electronics, Information, and Communication (ICEIC) 2025-01-19

Night security is known for its long hours and heavy tasks. In Japan, a labor shortage of guards has become an issue in recent years. To solve these problems, increasing number robotic methods are being used. However, several problems exist with existing robots. For example, wheeled robots traveling on the ground have difficulty dealing obstacles such as steps, while most drones only monitoring do not function to help people. this study, aerial ubiquitous display (AUD) night drone been...

10.3390/drones7050307 article EN cc-by Drones 2023-05-05

Abstract Recently, measuring users and community influences on social media networks play significant roles in science engineering. To address the problems, many researchers have investigated with these by dealing huge data sets. However, it is hard to enhance performances of studies multiple attributes together networks. This paper has presented a novel model for network. In this model, suggested algorithm combines Knowledge Graph learning techniques based vote rank mechanism reflect user...

10.2478/jaiscr-2023-0013 article EN Journal of Artificial Intelligence and Soft Computing Research 2023-06-01

In this paper, we present robust methods for automatically segmenting phases in a specified surgical workflow by using latent Dirichlet allocation (LDA) and hidden Markov model (HMM) approaches. More specifically, our goal is to output an appropriate phase label each given time point of operating room. The fundamental idea behind work lies constructing HMM based on observed values obtained via LDA topic covering optical flow motion features general working contexts, including medical staff,...

10.1155/2017/1985796 article EN cc-by International Journal of Biomedical Imaging 2017-01-01

In contemporary scenarios, privacy is paramount, especially in applications such as video interviews. This study introduces a privacy-focused real-time face and hair swapping method designed to conceal identity while retaining essential facial attributes of the original subject. Unlike conventional methods that rely on reference image realistic person, our proposed eliminates this need for enhanced privacy. Instead, work use synthetic generated by Generative Adversarial Networks (GANs),...

10.1109/access.2024.3420452 article EN cc-by-nc-nd IEEE Access 2024-01-01

In a wearable robot arm, the minimum joint configuration and link length must be considered to avoid increasing burden on user. This work investigated how configuration, of arm links, mounting position affect cooperative invasive workspaces overall workspace. We configurations lengths passive active joints in our proposed which is called Assist Oriented Arm (AOA). addition, we comprehensively studied As result, three locations around shoulders two waist were chosen as potential sites....

10.3390/robotics11010019 article EN cc-by Robotics 2022-01-29

To enable machines to understand human-centric images and videos, they need the capability detect human–object interactions. This has been studied using various approaches, but previous research mainly focused only on recognition accuracy widely used open datasets. Given for advanced machine-learning systems that provide spatial analysis services, model should be robust changes, have high extensibility, sufficient speed even with minimal computational overhead. Therefore, we propose a novel...

10.1080/18824889.2023.2292353 article EN cc-by SICE Journal of Control Measurement and System Integration 2023-12-23

Traditional models for various machine learning problems such as image classification perform well only under the assumption of a closed set. This implies that inputs must belong to classes which were trained. Data collected in real world may not any finite set classes, and training model with an infinite number would obviously be impossible. Rather than incorrectly classifying outlier unknown members one on was trained, recognize reject data samples, or request human assistance labeling...

10.1109/access.2022.3192621 article EN cc-by IEEE Access 2022-01-01

In this paper, we proposed a pointing interaction system which allows users to control devises in Intelligent Space by using hand pointing. This consists of multiple camera devices and pan-tilt projector. recognizes user's face, head orientation spot where user points with the devices. However, achieve real interaction, an intuitive feedback is required let know what recognizes. We placed projector developed method based on location. The experimental results show performance validity paper.

10.1109/sii.2011.6147451 article EN 2022 IEEE/SICE International Symposium on System Integration (SII) 2011-12-01

We propose a robust method of estimating head orientation based on HOG. The proposed is able to estimate with camera even though when user not facing the camera. With this method, can be estimated precisely in all three axes: roll, yaw, pitch. Furthermore, simple and identification composed by using results Approximate Nearest Neighbor Search. Finally, we combine these estimation methods apply them into real applications provide useful services users any intelligent environment.

10.1109/sii.2011.6147450 article EN 2022 IEEE/SICE International Symposium on System Integration (SII) 2011-12-01

Current probabilistic models for activity recognition do not incorporate much sensory input data due to the problem of state space explosion. In this paper, we propose a model recognition, called Factored State-Abtract Hidden Markov Model (FS-AHMM) allow us integrate many sensors improving performance. The proposed FS-AHMM is an extension Abstract which applies concept factored representations compactly represent transitions. parameters are estimated using EM algorithm from acquired through...

10.1109/issnip.2005.1595601 article EN International Conference on Intelligent Sensors, Sensor Networks and Information Processing 2005-01-01

To tackle the problem of increasing numbers state transition parameters when number sensors increases, we present a probabilistic model together with several parsinomious representations for sensor fusion. These include context specific independence (CSI), mixtures smaller multinomials and softmax function to compactly represent transitions large sensors. The is evaluated on real-world data acquired through ubiquitous in recognizing daily morning activities. results show that combination CSI...

10.1109/icpr.2006.154 article EN 2022 26th International Conference on Pattern Recognition (ICPR) 2006-01-01

In this paper, we propose a method that automatically estimates surgical phases in specified workflow with multi-camera system. More specifically, our goal is to output an appropriate phase label for each one-second of input videos captured by multiple cameras operating room. The fundamental idea behind work lies constructing hidden Markov model based on motion features, which are actually the PCA-compressed histograms optical flow general working contexts, including medical staff,...

10.1109/urai.2017.7992903 article EN 2017-06-01

In this study, we evaluate efficient traveling strategies for a water surface robot to cover multiple aquatic environment observation points. the previous have developed an autonomous sensing called BIWAKO-X and evaluated position-keeping long-term in-situ observation. paper, focus on multi-point strategy that is designed comprehensive environmental across locations. Assuming scenario in which travels several points, travelling was constructed with cycles site distances. simulation practical...

10.1109/sii58957.2024.10417541 article EN 2022 IEEE/SICE International Symposium on System Integration (SII) 2024-01-08

2D image-based face tracking is a core feature for multiple AR/VR applications. The latest advancements in self-supervised 3DMM reconstruction maintained high-accuracy analysis-by-synthesis but were not designed online inference settings with low latency performance. Recently, state-of-the-art models such as MICA [1] has demonstrated significant improvement term of accuracy the offline construction task design ill-suited practical use cases due to their long processing time on and middle-end...

10.1109/sii58957.2024.10417546 article EN 2022 IEEE/SICE International Symposium on System Integration (SII) 2024-01-08

Understanding how human utilize space is important to design optimal layouts and indoor robot navigation considering surrounding conditions. Tasks such as abnormal activity detection human-robot interaction could also leverage the utilization data provide deeper information operators control in an intelligent respectively. However, recognizing actions time-consuming labor difficult define. In this paper, we proposed a novel method infer from state recognition open-set solution using word...

10.1109/sii58957.2024.10417211 article EN 2022 IEEE/SICE International Symposium on System Integration (SII) 2024-01-08

In recent years, deep learning methodologies have been increasingly applied to the intricate challenges of visual-inertial odometry (VIO), especially in scenarios with rapid movements and scenes lacking clear structure. This paper introduces a novel hybrid approach that leverages inherent strengths traditional VIO techniques, while harnessing potential advanced machine technologies. By seamlessly integrating an iterated extended Kalman filter our systematically takes into account...

10.1109/access.2024.3440182 article EN cc-by-nc-nd IEEE Access 2024-01-01

Pain is a more intuitive and user-friendly way of communicating problems, making it especially useful in rehabilitation nurse training robots. While most previous methods have focused on classifying or recognizing pain expressions, these approaches often result unnatural, jiggling robot faces. We introduce PainDiffusion, model that generates facial expressions response to stimuli, with controllable expressiveness emotion status. PainDiffusion leverages diffusion forcing roll out predictions...

10.48550/arxiv.2409.11635 preprint EN arXiv (Cornell University) 2024-09-17
Coming Soon ...