PhD student at Frankfurt Institute for Advanced Studies, Germany
Mingjun Xiang 📡 is a passionate researcher in terahertz imaging and deep learning, currently based at the Frankfurt Institute for Advanced Studies 🇩🇪. With a strong academic background in Information and Communication Engineering and hands-on experience in optical systems, Mingjun bridges theoretical innovation with practical applications. He’s contributed to multiple international conferences 🎤 and has worked on advanced topics like phase retrieval, depth reconstruction, and semi-supervised learning. Fluent in English 🇬🇧 and Mandarin 🇨🇳, he brings a global perspective to his research. In his spare time, he enjoys swimming 🏊, playing traditional Chinese instruments 🎶, and photography 📷.
Professional Profile:
Orcid
Scopus
🔹 Education & Experience
🎓 Education:
-
📘 Ph.D. in AI for Terahertz Imaging – Frankfurt Institute for Advanced Studies (2021–Present), Germany
-
📗 M.Sc. in Information and Communication Engineering – Technische Universität Darmstadt (2017–2020), Germany
-
📙 B.Sc. in Electronics Information Science and Technology – Harbin Institute of Technology (2013–2017), China
💼 Work Experience:
-
🔧 Electronic Engineer Intern – Inspur Group (2016): Worked on fiber coupling modules using VirtualLab
-
📡 Communication Engineer Intern – China Mobile (2018): Developed optical fault location system with LabView & Matlab
-
🧪 Research Assistant – TU Darmstadt (2019): Developed GUI for Zurich Instruments’ devices in Matlab
-
🧫 Microfabrication Projects – Cleanroom experience in antenna design & THz waveguide fabrication
🔹 Professional Development
Mingjun Xiang is deeply committed to continuous learning and professional growth 📘. During his academic journey, he has mastered a diverse set of tools including Python 🐍, Matlab 📊, LabView ⚙️, and CST Microwave Studio 📡, along with cleanroom fabrication skills 🧼. He has presented in renowned global conferences such as IRMMW-THz 🌍 and Digital Holography & 3D Imaging 📸, often as a keynote or session speaker 🎙️. His hands-on engineering internships in China gave him practical insights into fiber optics and telecom systems 🔌. By combining advanced theory with real-world skills, he actively advances in the terahertz research community 🚀.
🔹 Research Focus Category
Mingjun Xiang’s research focuses on AI-enhanced Terahertz Imaging 📶, where he applies deep learning techniques to overcome challenges in phase retrieval, image denoising, and occluded object reconstruction 🧠. His work integrates physics-informed neural networks 🧬 with holographic data for superior image accuracy and depth perception 🧊. His current interest lies in using both supervised and unsupervised models to extract meaningful features from complex terahertz datasets 📂. This intersection of optics 🔍, machine learning 🤖, and microwave technology 📡 positions him at the forefront of the next-generation imaging systems for biomedical, industrial, and security applications 🔬🛡️.
🔹 Awards & Honors
-
🏅 Keynote Speaker – IRMMW-THz 2023, Canada
-
🗣️ Invited Speaker – IRMMW-THz 2022, Netherlands
-
🧠 Speaker – Digital Holography & 3D Imaging 2022, UK
-
📢 Speaker – Face2Phase 2022, Netherlands
-
🖼️ Poster Presenter – 10th Intl. Workshop on THz Tech & Applications, Germany
-
🥇 Master Thesis Achievement – Achieved -5 dB coupling loss in 0.67–1.37 THz range
Publication Top Notes
1. Amplitude/Phase Retrieval for Terahertz Holography With Supervised and Unsupervised Physics-Informed Deep Learning
-
Journal: IEEE Transactions on Terahertz Science and Technology
-
Publication Date: March 2024
-
DOI: 10.1109/TTHZ.2024.3349482
-
Type: Journal Article
-
Summary:
Combines supervised and unsupervised physics-informed deep learning (PIDL) methods for amplitude and phase retrieval in THz holography, enhancing accuracy and generalization without requiring a reference beam.
2. Depth Reconstruction for Reference-Free THz Holography Based on Physics-Informed Deep Learning
-
Conference: 48th International Conference on Infrared, Millimeter, and Terahertz Waves (IRMMW-THz)
-
Date: September 17, 2023
-
DOI: 10.1109/irmmw-thz57677.2023.10298937
-
Type: Conference Paper
-
Summary:
Proposes a PIDL-based method for depth reconstruction in reference-free THz holography. Demonstrates the capability of neural networks to learn physical laws and produce 3D profiles from 2D holograms.
3. Phase Retrieval for Fourier THz Imaging with Physics-Informed Deep Learning
-
Conference: 47th IRMMW-THz
-
Date: August 28, 2022
-
DOI: 10.1109/irmmw-thz50927.2022.9895691
-
Type: Conference Paper
-
Summary:
Applies PIDL to Fourier domain imaging in the THz range, aiming at robust phase retrieval by embedding physical constraints in the training loss.
4. Phase Retrieval for Terahertz Holography with Physics-Informed Deep Learning
-
Conference: Digital Holography and 3-D Imaging 2022 (OSA)
-
Date: 2022
-
DOI: 10.1364/dh.2022.tu4a.4
-
Type: Conference Paper
-
Summary:
Earlier work demonstrating PIDL for phase retrieval in THz holography setups, potentially eliminating the need for phase-shifting or interferometric measurements.
5. Broadband Terahertz Photonic Integrated Circuit with Integrated Active Photonic Devices
-
Journal: Photonics
-
Publication Date: November 3, 2021
-
DOI: 10.3390/photonics8110492
-
Type: Journal Article
-
Summary:
Focuses on the development of a broadband THz photonic integrated circuit (PIC) with embedded active photonic components, marking progress toward compact, tunable THz systems.
Conclusion:
Mingjun xiang is a highly suitable and deserving nominee for the Best Researcher Award, particularly in areas that value cross-disciplinary innovation, high-frequency imaging, and AI integration in physical sciences.
His work advances both theoretical frameworks and practical implementations of next-generation THz systems, showcasing real-world applications and scientific impact. His combination of technical rigor, innovation, international presence, and AI-driven breakthroughs clearly sets him apart as a next-generation leader in his field.