About
Highly motivated and results-driven Deep Learning Researcher specializing in Computer Vision, with a strong academic background from Nanjing Tech University and the University of Burgundy Europe. Proven expertise in developing advanced Transformer-based models, generative AI frameworks (ControlNet, Diffusion), and optimizing deep learning architectures for object detection and pedestrian attribute recognition. Seeking PhD or research/work opportunities to leverage extensive experience in SOTA model development, UAV visual perception, and medical image segmentation to drive impactful advancements in AI.
Work
Laboratoire Connaissance et Intelligence Artificielle Distribuées (CIAD)
|Stage Researcher
Montbeliard, Bourgogne-Franche-Comté, France
→
Summary
Led advanced research in Artificial Intelligence, focusing on innovative generative frameworks and State-of-the-Art (SOTA) model development within a public research laboratory.
Highlights
Developed a comprehensive generative framework leveraging ControlNet and diffusion technologies, constructing a simulation dataset from confidential data to augment attributes and alleviate class imbalance for Pedestrian Attribute Recognition (PAR) tasks.
Designed and implemented a novel State-of-the-Art (SOTA) PAR model based on Transformer architecture, attention mechanisms, and optimized DINO self-distillation, achieving superior performance in attribute recognition.
Co-authored a research paper documenting the generative framework and PAR model, currently under peer review by Prof. Yassine, with anticipated publication.
Nanjing Tech University
|Research-based Master
Nanjing, Jiangsu, China
→
Summary
Conducted advanced research in computer vision and deep learning, contributing to significant projects and a peer-reviewed publication under expert guidance.
Highlights
Co-authored and presented "MRPVT: A novel Multiscale Rain-cutting and Pooling Vision Transformer model for Object Detection on Drone captured Scenarios" at the 2024 ICCVIT, significantly enhancing object detection capabilities in challenging UAV environments.
Participated in extensive laboratory work, completing multiple project tasks focused on cutting-edge computer vision applications and deep learning methodologies.
Education
University of Burgundy Europe
→
Master
Computer Vision
Courses
Optimized RT-DETR-Based Neck Network for UAV Visual Perception Tasks, enhancing drone-based object detection.
Developed ROS robot control, navigation, and SLAM systems for Turtlebot autorace platforms, demonstrating expertise in robotics and autonomous systems.
Implemented Medical Image Segmentation for nailfold capillary using the MaskFormer method, showcasing advanced skills in specialized image analysis.
Nanjing Tech University
→
Master of Science
Control Science and Engineering
Courses
Conducted advanced research in control systems and engineering principles, contributing to a robust foundation in complex system design.
Engaged in experimental group studies, applying theoretical knowledge to practical challenges in control science.
Jiangsu Normal University
→
Bachelor
Rail Transportation and Signal Control
Languages
English
Chinese
French
Russian
Skills
Programming & Development
Python, ROS.
Deep Learning & Computer Vision
Object Detection, Pedestrian Attribute Recognition (PAR), Transformer-based Models, Generative AI (Diffusion), ControlNet, DINO Self-distillation, MaskFormer, SLAM, UAV Visual Perception, Medical Image Segmentation, RT-DETR.
Data Analysis & Tools
Matlab, Markdown, Microsoft Word, PowerPoint.
Research & Methodologies
Paper Writing, Research, Experimental Design, Problem-solving, Result-oriented.
Collaboration & Management
Collaboration, Time Management, Project Management, User Interface Design (UI).