PeopleofSuperintelligence

Researcher

Chunjing Xu

Chunjing Xu is a computer-vision researcher whose work spans efficient neural networks, image restoration and driving-scene perception. The 2007 clustering paper coauthored by Xu bypassed spectral clustering's eigenproblem with comparable performance at O(n²) complexity; AdderNet traded neural-network multiplications for additions in 2019. In 2020, a portable image-translation GAN coauthored by Xu used knowledge distillation to train a smaller generator, and Xu coauthored SCOP, GhostNet and the Image Processing Transformer. TopoNet, coauthored by Xu in 2023, extended topology reasoning for driving scenes, and Shanghai Jiao Tong University was Xu's affiliation in 2025. As of October 2026, Xu is affiliated with Huawei Technologies Co., Ltd. in China, with artificial intelligence and image processing as Xu's field; a 2026 patent application named Xu as an inventor on an image-enhancement model.

Now at
Huawei
Works
116
Citations
19,081
h-index
36
Badges

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. 2007

    Clustering without eigenproblems

    In 2007, a clustering method coauthored by Xu avoids the eigenproblem used by spectral clustering and offers comparable performance with O(n²) computational complexity.

  2. 2019

    AdderNet with additions

    In 2019, AdderNet, coauthored by Xu, trades multiplication-heavy neural-network operations for cheaper additions to reduce computation costs.

  3. 2020

    A portable image-translation GAN

    In 2020, a portable image-translation GAN coauthored by Xu uses knowledge distillation to train a smaller student generator from a larger teacher model.

  4. 2020

    SCOP, GhostNet and IPT

    In 2020, SCOP, coauthored by Xu, uses a scientific control group in neural-network pruning, while GhostNet, also coauthored by Xu, generates more feature maps through cheap operations. The Image Processing Transformer, also coauthored by Xu in 2020, targets low-level vision tasks such as denoising, super-resolution and deraining.

  5. 2023

    TopoNet for driving scenes

    In 2023, TopoNet, coauthored by Xu, is an end-to-end framework for driving-scene topology that abstracts traffic knowledge beyond conventional perception tasks.

  6. 2025

    At Shanghai Jiao Tong

    In 2025, Xu is affiliated with Shanghai Jiao Tong University in China.

  7. June 2026

    An image-enhancement patent

    In June 2026, a Huawei patent application names Xu as an inventor on an image-enhancement model with selection, segmentation, transformer and recombination modules.

  8. As of October 2026

    Huawei research affiliation

    As of October 2026, Xu is affiliated with Huawei Technologies Co., Ltd. in China, with artificial intelligence and image processing as Xu's field.

Career

  1. Since 2018HuaweiResearcher
  2. 2023–2025Shanghai Jiao Tong University
  3. 2019–2021Peking University
  4. 2019–2022University of Sydney
  5. 2009–2026Chinese Academy of Sciences
  6. 2006–2013Chinese University of Hong Kong

From the biography and OpenAlex affiliations (the years of papers that list the organization), so some years are approximate.

  1. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes Huawei. We estimate that alumni of 17 major AI labs went on to found or lead 350 other organizations, including 226 they founded, and we trace 472 career moves directly between 23 labs since 2010.
  2. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes Huawei. 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.
  3. Universities keep AI researchers for a median of 6 years, compared with 2 years at frontier labsIncludes Chinese Academy of Sciences. Across 15,811 dated work stays (6,529 people), median role tenure runs from 6 years at universities to 2 years at frontier labs. We compare all five organization types, starting decades from the 1970s to the 2020s, and the organizations that keep people longest.
  4. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Chinese University of Hong Kong. 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.

Colleagues

People with dated positions at the same organization in the same years.

Sources and credits

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

  1. Clustering with Transitive Distance and K-Means Duality arxiv.org. Checked 22 October 2026.
  2. AdderNet: Do We Really Need Multiplications in Deep Learning? arxiv.org. Checked 22 October 2026.
  3. Distilling Portable Generative Adversarial Networks for Image Translation ojs.aaai.org. Checked 22 October 2026.
  4. SCOP: Scientific Control for Reliable Neural Network Pruning arxiv.org. Checked 22 October 2026.
  5. GhostNet: More Features From Cheap Operations openaccess.thecvf.com. Checked 22 October 2026.
  6. Pre-Trained Image Processing Transformer arxiv.org. Checked 22 October 2026.
  7. Graph-based Topology Reasoning for Driving Scenes arxiv.org. Checked 22 October 2026.
  8. Xu, Chunjing, Shanghai Jiao Tong University | TOPSCINET topscinet.com. Checked 22 October 2026.
  9. US20260179360A1 - Model structure, method for training model, image enhancement method, and device - Google Patents patents.google.com. Checked 22 October 2026.