Zhengyuan Zhang

Zhengyuan Zhang

Computational imaging, Medical imaging, Inverse problem, Hardware acceleration

A short introduction

I am a Postdoctoral Research Fellow jointly at the College of Computing and Data Science and the Lee Kong Chian School of Medicine, Nanyang Technological University (NTU), Singapore. From 2023 to 2026, I was a Postdoctoral Research Fellow at NTU’s School of Electrical and Electronic Engineering, where I also received my Ph.D. in 2024, supported by a full NTU Ph.D. scholarship. Before that, I received my M.Eng. from the State Key Laboratory of Robotics and System at Harbin Institute of Technology (HIT), supervised by Academician Hong Liu and Prof. Minghe Jin, and my B.Eng. from the School of Mechatronics Engineering, HIT.

My research is in computational imaging and medical imaging, where I combine physical imaging models with deep learning to solve inverse problems. My work covers photoacoustic microscopy and medical ultrasound: I develop model-based and learning-based methods for photoacoustic imaging enhancement, and diffusion model-based methods that bring portable ultrasound images closer to high-end clinical quality. I also work on hardware acceleration, making these algorithms run in real time on handheld imaging systems through in-memory computing and few-step diffusion inference.

I teach IN6242 Deep Learning Foundations, a postgraduate course at NTU’s Wee Kim Wee School of Communication and Information, and co-supervise Master’s and undergraduate students at NTU.

News

  • Started teaching IN6242 Deep Learning Foundations, a postgraduate course at NTU’s Wee Kim Wee School of Communication and Information.
  • Our paper on physics-guided segmentation for lensless imaging was published in IEEE Transactions on Emerging Topics in Computational Intelligence.
  • Joined the College of Computing and Data Science and the Lee Kong Chian School of Medicine, NTU, as a Postdoctoral Research Fellow.
  • Dual Branch Graph Induced Fusion Network for high temporal-lateral resolution PAM published in IEEE Transactions on Computational Imaging.
  • Two papers presented orally at IEEE ISCAS 2025.
  • FeRAM neuromorphic design for real-time handheld photoacoustic imaging published in IEEE Transactions on Biomedical Circuits and Systems.
  • Group sparsity with deep denoiser prior for AR-PAM published in IEEE Transactions on Image Processing.
  • Completed my Ph.D. at Nanyang Technological University, Singapore.
  • Joined the School of Electrical and Electronic Engineering, NTU, as a Postdoctoral Research Fellow.

Research Interests:

  • Computational Imaging
  • Medical Imaging
  • Inverse Problem
  • Hardware Acceleration

Research Highlights:

Selected Journal Publications

+ co-first author · * corresponding author · Full publication list

For more information

More information about me can be found in my CV or the downloadable CV.

Publications

+ co-first author · * corresponding author · Full and current list on Google Scholar

Journal Articles

2026

2025

2024

2023

2022

Under Review

  • “Physics-Guided Proximal Diffusion Sampling with Learned Consistency Weighting for Photoacoustic Microscopy Enhancement.” IEEE Transactions on Image Processing, under review.
  • “Real-Time Imaging Enhancement for Portable Medical Ultrasound via Unsupervised Latent Diffusion Model.” IEEE Transactions on Image Processing, under review.
  • “Unsupervised Physics-Driven Generative Model for Adaptive Photoacoustic Microscopy Enhancement.” IEEE Transactions on Image Processing, under review.

Conference Papers

Thesis

Research

Projects I have led or worked on, most recent first.

Diffusion Model-Based Ultrasound Imaging Enhancement

Project Leader · CICS Laboratory, NTU · Oct 2023 – present

  • Proposed UNIT-LDM, a two-stage unsupervised framework that aligns portable and high-end ultrasound images in a shared latent space, then refines them with latent diffusion guided by a self-supervised edge condition that preserves anatomy.
  • Distilled the 50-step diffusion sampler into a few-step student network, raising throughput to 32.3 FPS, above the 24 FPS native scan rate of the portable device, while giving the strongest structural preservation among compared methods on carotid ultrasound.
  • Proposed SonoSDE, an unsupervised score-based diffusion framework that enhances handheld ultrasound images toward high-end clinical quality without paired training data.
  • Designed physics-guided experts for SonoSDE: a Realistic Expert that matches frequency spectrum, speckle statistics and texture, and a training-free Faithful Expert based on Log-Gabor filters and phase congruency that preserves anatomy.

Related: IEEE TIP (under review)

Real-Time Handheld Photoacoustic Imaging Enhancement with Hardware Acceleration

Project Leader · CICS Laboratory, NTU · Oct 2023 – Jan 2025

  • Developed a real-time enhancement platform for a handheld photoacoustic imaging system, running a ternary-quantized MultiResU-Net on a FeRAM crossbar array for in-memory computing.
  • Proposed a fast crossbar charging model that accounts for line and parasitic conductance, enabling hardware-aware training.
  • Validated the platform in vivo: lateral resolution improved from 850 µm to 78 µm, network inference finished within microseconds, and 30 Hz real-time imaging was kept.

Related: IEEE TBCAS 2025 · ISCAS 2025

Model-Based Photoacoustic Imaging Enhancement

Project Leader · CICS Laboratory, NTU · Sep 2021 – May 2025

  • Developed the degradation model of acoustic-resolution photoacoustic microscopy.
  • Designed a deep CNN denoiser prior and its optimization scheme to enhance imaging results.
  • Designed a joint group sparsity and deep prior model with an adaptive optimization scheme.

Related: Photoacoustics 2023 · IEEE TIP 2025 · BioCAS 2023

Learning-Based Photoacoustic Imaging Enhancement

Project Leader · CICS Laboratory, NTU · Jul 2019 – Sep 2023

  • Developed the degradation model of photoacoustic imaging for training data synthesis.
  • Used GAN-based learning to achieve realistic, high-resolution imaging enhancement.
  • Employed a Dual Branch Gated Fusion Network to improve imaging resolution and speed together.

Related: IEEE TMI 2022 · IEEE TCI 2025 · ISCAS 2021 · ISCAS 2022 · BioCAS 2022

Target Capturing Based on Computer Vision

M.Eng. Student · State Key Laboratory of Robotics and System, HIT · Apr 2018 – Jul 2019

  • Used a binocular stereo camera to extract the position and orientation of the target.
  • Reconstructed the 3D model of the target with structure from motion.
  • Detected the handle and nozzle of the object for manipulator capturing.

Target Manipulation Based on Impedance Control

B.Eng. Student · State Key Laboratory of Robotics and System, HIT · Oct 2016 – Dec 2017

  • Developed the communication program between the force/torque sensor and the controller.
  • Modeled the kinematics and dynamics of the manipulator and developed an impedance control strategy.
  • Simulated the controller in Simulink and verified the algorithm on a real manipulator.

Teaching & Service

Course Teaching

Aug – Nov 2026
IN6242 Deep Learning Foundations

Lecturer of this postgraduate course at the Wee Kim Wee School of Communication and Information, NTU. AY2026/27 Semester 1, 13 weeks, one 3-hour class per week.

Student Co-supervision

Master’s dissertations, NTU

  • Yang Sijin (2025)
  • Zhengke Sun (2024)
  • Ze Feng (2023), Yudi Zhang (2023)
  • Chenyang Liu (2022), Zeliang Hao (2022)
  • Yibing Duan (2021), Xiaoshi Sun (2021)

Bachelor’s final year projects, NTU

  • Joey Tee Hui Xing (2026)
  • Leyi Li (2025)
  • Jing Lin Ong (2024)
  • Yu Myat Noe Oo (2023)
  • Potipireddi Sai Pratyusha (2022), Muralidharan Vaishnavi (2022)
  • Su Wai Wai Zin (2021)

Reviewer Service

Journal reviewer since 2021 for:

  • IEEE Transactions on Multimedia
  • IEEE Sensors Journal
  • Neurocomputing
  • Photoacoustics

CV

Download the full CV (PDF)

Education

2019 – 2024
Ph.D., School of Electrical and Electronic Engineering

Nanyang Technological University, Singapore. Supervisor: Prof. Yuanjin Zheng. GPA 4.75/5.0.

2017 – 2019
M.Eng., State Key Laboratory of Robotics and System

Harbin Institute of Technology, Harbin, China. Supervisors: Academician Hong Liu (Member of the Chinese Academy of Engineering) and Prof. Minghe Jin. GPA 88.5/100, top 10%.

2013 – 2017
B.Eng., School of Mechatronics Engineering

Harbin Institute of Technology, Harbin, China. Supervisors: Academician Hong Liu (Member of the Chinese Academy of Engineering) and Prof. Minghe Jin. GPA 90.47/100, rank 22/237.

Work Experience

Feb 2026 – present
Postdoctoral Research Fellow

College of Computing and Data Science and Lee Kong Chian School of Medicine, Nanyang Technological University, Singapore. Working with Prof. Fan Xiuyi.

Oct 2023 – Jan 2026
Postdoctoral Research Fellow

School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore. Worked with Prof. Yuanjin Zheng.

Awards

2019 – 2023

Full Scholarship for NTU Ph.D. (4 years)

2017

First-Class Scholarship, Harbin Institute of Technology

2016

Engineering Leadership Program, Harbin Institute of Technology (10/237)

2016

First-Grade Scholarship, Harbin Institute of Technology (4/237)

2015

First-Grade Scholarship, Harbin Institute of Technology (11/237)

Skills

Programming

Python, PyTorch, TensorFlow, OpenCV, MATLAB, C/C++, Vivado HLS, Verilog

Areas

Computational imaging, image restoration, inverse problems, model-based image processing, SLAM, machine learning, deep learning, matrix analysis, convex optimization, probability and statistics, hardware acceleration

Coursework

Image Analysis & Pattern Recognition, Machine Vision, Computer Networks, Data Structures & Algorithms, Computer Architecture & Acceleration for Machine Learning, Operating Systems

Languages

Mandarin (native), English (fluent), German (fluent)