PeopleofSuperintelligence

Researcher

Kai Zhang

Also known as 张凯, cszn

Kai Zhang is a computer vision researcher who builds deep networks for image denoising, super-resolution and other image restoration problems. He wrote the DnCNN denoiser with Wangmeng Zuo, Yunjin Chen, Deyu Meng and Lei Zhang, published in 2017, and followed it with FFDNet in 2018. After a Ph.D. at Harbin Institute of Technology in 2019, he spent 2020 to 2024 as a postdoctoral researcher at the Computer Vision Lab of ETH Zurich and, in 2021, introduced the BSRGAN degradation model and coauthored SwinIR. FFDNet earned the IEEE Signal Processing Society's 2024 Best Paper Award, presented in 2025, and DnCNN received the 2026 Frontiers of Science Award of the International Congress of Basic Science. Since March 2024, he is an associate professor at the School of Intelligence Science and Technology at Nanjing University.

Timeline

  1. 2016–2017

    The DnCNN denoiser

    In 2016, he posts DnCNN on arXiv with Wangmeng Zuo, Yunjin Chen, Deyu Meng and Lei Zhang. The feed-forward denoising network uses residual learning and batch normalization and handles blind Gaussian denoising with unknown noise level. IEEE Transactions on Image Processing publishes the paper Beyond a Gaussian Denoiser in 2017.

  2. 2018

    FFDNet for flexible denoising

    In 2018, IEEE Transactions on Image Processing publishes FFDNet, which he writes with Wangmeng Zuo and Lei Zhang. The convolutional denoiser takes a tunable noise level map as input, so a single network handles noise levels from 0 to 75.

  3. 2019

    PhD from Harbin Institute of Technology

    He earns a Ph.D. in Computer Science and Technology from Harbin Institute of Technology in 2019. During his doctoral studies, he spends three years as a research assistant at The Hong Kong Polytechnic University.

  4. 2020–2024

    Postdoc at ETH Zurich

    From 2020 to 2024, he is a postdoctoral researcher at the Computer Vision Lab of ETH Zurich. In 2020, he posts Plug-and-Play Image Restoration with Deep Denoiser Prior (DPIR) with Yawei Li, Wangmeng Zuo, Lei Zhang, Luc Van Gool and Radu Timofte, which plugs a CNN denoiser into a half quadratic splitting algorithm to solve deblurring, super-resolution and demosaicing. He also releases KAIR, a PyTorch toolbox with training and testing code for DPIR, DnCNN, FFDNet, BSRGAN and SwinIR.

  5. 2021

    BSRGAN and SwinIR

    In 2021, he writes the ICCV paper on a practical degradation model for blind image super-resolution, released as the BSRGAN degradation model, which combines randomly shuffled blur, downsampling and noise. He also coauthors SwinIR with Jingyun Liang, Jiezhang Cao, Guolei Sun, Luc Van Gool and Radu Timofte, an image restoration model built on the Swin Transformer and tested on super-resolution, denoising and JPEG artifact reduction.

  6. 2025

    IEEE best paper award for FFDNet

    In 2025, he receives the IEEE Signal Processing Society's 2024 Best Paper Award with Wangmeng Zuo and Lei Zhang for FFDNet, presented during ICASSP 2025 in India.

  7. 2026

    A Frontiers of Science Award

    The DnCNN paper from 2017 receives the 2026 Frontiers of Science Award of the International Congress of Basic Science, in the field of information science and engineering.

  8. Since March 2024

    Associate professor at Nanjing University

    He joins Nanjing University in March 2024 as an associate professor at the School of Intelligence Science and Technology on the Suzhou campus. His research there centers on low-level vision, with a focus on image and video restoration and enhancement.

Career

  1. Since 2024Nanjing UniversityAssociate Professor
  2. 2020–2024ETH ZurichPostdoctoral Researcher
  3. Hong Kong Polytechnic UniversityResearch Assistant

Education

  1. 2019Harbin Institute of TechnologyPh.D.

From the biography.

  1. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Kai Zhang. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.
  2. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes Nanjing University. We estimate where the people shaping AI studied and where their careers led. Massachusetts Institute of Technology leads with 467 former students, including 291 recorded doctorates.
  3. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes Nanjing University. We estimate where the people shaping AI worked when their most-cited papers appeared: the company share was 31% in 2020–2024 and 8% in 1970–1974, and Google led all organizations in 2020–2024.
  4. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Nanjing University. The median h-index is 48 at universities and 14 at frontier AI labs. We compare publishing careers, recent citations, dual appointments and job moves across 2,420 university researchers and 308 frontier-lab researchers.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Kai Zhang yet. Send us a free-licensed photo.

  1. Kai Zhang Homepage cszn.github.io. Checked 21 October 2026.
  2. [1608.03981] Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising arxiv.org. Checked 21 October 2026.
  3. [1710.04026] FFDNet: Toward a Fast and Flexible Solution for CNN based Image Denoising arxiv.org. Checked 21 October 2026.
  4. [2008.13751] Plug-and-Play Image Restoration with Deep Denoiser Prior arxiv.org. Checked 21 October 2026.
  5. GitHub - cszn/KAIR: Image Restoration Toolbox (PyTorch). Training and testing codes for DPIR, USRNet, DnCNN, FFDNet, SRMD, DPSR, BSRGAN, SwinIR github.com. Checked 21 October 2026.
  6. [2103.14006] Designing a Practical Degradation Model for Deep Blind Image Super-Resolution arxiv.org. Checked 21 October 2026.
  7. [2108.10257] SwinIR: Image Restoration Using Swin Transformer arxiv.org. Checked 21 October 2026.
  8. 2024 IEEE Signal Processing Society Awards Presented in India! | IEEE Signal Processing Society signalprocessingsociety.org. Checked 21 October 2026.