Zhao Mingfan

Mingfan Zhao

PhD Candidate in Electronic Science & Technology
zhaomf7@mail2.sysu.edu.cn
1594313483
Sun Yat-sen University
Guangzhou, China

Professional Profile

Personal Introduction

Hello, I'm Mingfan Zhao(赵明帆), a Ph.D. candidate at Sun Yat-sen University, advised by Dr. Wen-Dong Mao and Prof. Zhongfeng Wang (IEEE Fellow). I received my M.S. from Xi'an Institute of Optics and Precision Mechanics, CAS in 2023 under Prof. Xiaoqiang Lu, and my B.S. from Northwest University (China) in 2018. My research interests include Deep Learning, AI Hardware Acceleration, and Robotics.

Research Experience

Robot Path Planning

Exploring innovative algorithms and techniques for efficient and reliable robot path planning in dynamic environments.

  • My Role: Algorithm design and implementation
Transformer Hardware Acceleration for AI

Hardware acceleration for neural network models with focus on optimizing power efficiency.

  • My Role: Algorithm optimization and hardware design
Cycle and Saliency Map-Guided GANs for Infrared Image Colorization

Developed GAN-based methods for both paired and unpaired infrared image colorization, addressing artifacts and improving colorization quality using saliency maps and cycle consistency.

  • My Role: Redesigned network architectures and optimized training strategies
Deep Learning-based Polarization Image Demodulation

Proposed a deep learning approach for demodulating polarization images, effectively removing stripe noise while preserving critical information.

  • My Role: Redesigned generator and loss functions to utilize hierarchical features

Technical Skills

Core Skills

C++
Python
Verilog
MATLAB

Publications

DATA figure
DATA: A Memory-Efficient Deformable Transformer Accelerator via Neural Architecture Search
M. Zhao, W. Mao, Z. Wang
IEEE Transactions on Very Large Scale Integration (VLSI) Systems, vol. 33, no. 11, pp. 3097-3108, Nov. 2025
First Author
SCI Journal Paper

This paper presents DATA, a memory-efficient deformable transformer accelerator designed through neural architecture search. The work focuses on optimizing hardware implementation of transformer models with deformable attention mechanisms, addressing key challenges in memory management and computational efficiency.

DOI: 10.1109/TVLSI.2025.3589359

View Paper
Optics Express figure
A novel demodulation method of the channeled modulated polarization imaging pictures by hybrid feature modulated autoencoders
Ning Zhang, Mingfan Zhao, Zhinan Zhang, Jie Liu, Yunyao Zhang, and Siyuan Li
Optics Express, vol. 32, no. 18, pp. 31473-31485, Aug. 2024
Co-First Author
SCI Journal Paper

This paper presents a novel demodulation method for channeled modulated polarization imaging using hybrid feature modulated autoencoders. The approach addresses key challenges in polarization image processing and demonstrates improved performance in automated image analysis.

View Paper
Applied Optics figure
Guide to the design of broadband full polarization imager based on dispersion-compensation Savart plates
Zhang N, Zhao M, Zhang Y, et al.
Applied optics, 2023(19):62
Second Author
SCI Journal Paper

This paper provides a comprehensive guide for designing broadband full polarization imagers using dispersion-compensation Savart plates, with focus on optical system optimization and performance enhancement.

View Paper
CSTGAN figure
CSTGAN: Cycle Swin Transformer GAN for Unpaired Infrared Image Colorization
Zhao M, Feng G, Tan J, et al.
2022 3rd International Conference on Control, Robotics and Intelligent System (CCRIS), 2022
First Author
EI Conference Paper

This paper presents a novel GAN-based approach for colorizing unpaired infrared images, addressing common artifacts and distortion issues through a cycle-consistent Swin Transformer architecture.

View Paper

Research in Progress

Currently working on additional papers in the areas of transformer hardware acceleration and robotics.

Education Background

Sun Yat-sen University
Sun Yat-sen University

PhD Candidate in Electronic Science & Technology

Advisor: Wendong Mao, Zhongfeng Wang

  • Doctoral First-Class Scholarship (2023-2024)
  • Research focus: Hardware acceleration for neural networks, computer vision algorithms
University of Chinese Academy of Sciences
University of Chinese Academy of Sciences

Master of Engineering in Electronic Information

Cultivation Unit: Xi'an Institute of Optics and Precision Mechanics, CAS

Advisor: Xiaoqiang Lu

  • Outstanding Student Award (2020-2021)
  • Outstanding Student Award (2021-2022)
  • Master's thesis focused on infrared image colorization using GANs
Northwest University
Northwest University

Bachelor of Engineering in Optoelectronic Information Science & Engineering

  • University Scholarship (2014-2015)
  • University Scholarship (2015-2016)
  • University Scholarship (2016-2017)

Work Experience

Southwest Computer Co., Ltd. (Chongqing)

C++ Software Developer

  • One of the main developers for C++ application software
  • Responsibilities: Software framework design and sub-module development
  • Achievements: Completed framework construction and module interaction; delivered initial version before departure

Hobbies & Interests

Outside of my academic and research work, I enjoy various activities that help me maintain a balanced lifestyle and foster creativity. Here are some of my interests:

Volleyball

Passionate about playing volleyball, enjoying the teamwork and physical engagement that the sport provides.

Frisbee

Enjoying frisbee as a recreational sport that combines skill, strategy, and outdoor fun with friends.

Tennis

Dedicated to tennis as a sport that challenges both mind and body, providing excellent cardiovascular exercise.

Life Photos

Extracurricular Activities

  • Northwest University Science & Entrepreneurship Association - Vice Minister of Science Popularization Department (2015.6 - 2016.6)
  • UCAS School of Optoelectronics Student Union - Minister of Arts Department (2020.9 - 2021.6)

Additional Notes

About Me

I am an optimistic and positive person, always open to communication and collaboration. Throughout my journey, I feel incredibly lucky to have received support and help from so many people. I aspire to continue pursuing scientific research and hope to make meaningful contributions in the future!