About me
Hello! I am currently pursuing my master’s degree at the School of Engineering, University of Edinburgh, where I’m fortunate to be supervised by Prof. Sotirios A. Tsaftaris, with co-supervision from Dr. Panos Dimitrakopoulos and Mr. Konstantinos Vilouras.
My primary research interests lie in machine learning, image processing, computer vision, and digital signal processing. I am particularly interested in enhancing the efficiency and accuracy of AI algorithms, with applications spanning object detection and human pose estimation. Throughout my academic journey, I have sought to bridge theoretical knowledge and practical applications. My work has included object detection models such as YOLO, DETR, SSD, and Faster R-CNN, as well as human pose estimation methods like YOLOPose, OpenPose, and VideoPose3D.
Before pursuing my master’s studies, I had the privilege of working with Prof. Chengguang Sun, Prof. Jianqiang Mei, Prof. Dan Cheng, and Prof. Xuewen Ding on topics related to the Internet of Things, Embedded Systems, Image Processing, and Machine Learning. Under the supervision of Prof. Sun, I completed my undergraduate thesis “Design of Intelligent Greenhouse Garden Based on AIoT Technology”, which was awarded The Excellent Graduation Project (Thesis) in 2023. Later, under Prof. Mei’s supervision, I further strengthened my research and academic writing skills in deep learning and computer vision.
🔍 Current Research
I am currently conducting research on Visual Anagrams and Optical Illusions at the University of Edinburgh, under the supervision of Prof. Tsaftaris, Dr. Dimitrakopoulos, and Mr. Vilouras. My work investigates how generative diffusion models and Generative Adversarial Networks (GANs) can be used to create multi-view optical illusions that explore the boundaries between visual perception and machine generation. This research focuses on understanding how generative models represent ambiguous visual information, and how content–style and content–content interactions influence the creation of perceptually coherent yet semantically distinct images. The ultimate goal is to provide insights into human perception and AI interpretability through the lens of generative modeling.
🖼️ Visual Anagram Examples
| Experiment 1 | Experiment 2 | Experiment 3 |
|---|---|---|
 |  |  |
| Identity View (Original) | Identity View (Original) | Identity View (Original) |
 |  |  |
| Transformed View (Illusion) | Transformed View (Illusion) | Transformed View (Illusion) |
🎥 Animation (Transformation Process)
| Experiment 1 | Experiment 2 | Experiment 3 |
|---|---|---|
| Transformation Video | Transformation Video | Transformation Video |
🔗 Related Projects
- Visual Anagrams — exploring multi-view optical illusion generation with diffusion models
- GAN — experimenting with adversarial frameworks for visual realism and feature disentanglement
- Hugging Face Diffusers — implementing and customizing diffusion-based generation for perceptual alignment studies
🎨 Personal Interests
I also enjoy anime, which inspires my interest in artistic image generation and creative AI.
🚀 Looking Ahead
I am excited to continue exploring the intersections of machine learning, image processing, computer vision, and digital signal processing, and to contribute to research that is both scientifically impactful and practically meaningful.
If you would like to connect or discuss potential collaborations, please feel free to reach me at:
📧 18805644868@139.com | 📧 s2604458@ed.ac.uk
📚 For More Information
For an up-to-date list of my scholarly works, including publications and citations, please visit my
Google Scholar profile.
Welcome to explore and connect! 🌟