Remembering Tang Xiaoyu: The Visionary Behind Modern Facial Recognition

AI pioneer Tang Xiaoyu, co‑founder of SenseTime and former director of leading computer‑vision labs, passed away in December 2023, leaving a legacy of groundbreaking facial‑recognition algorithms, influential mentorship, and a profound impact on the global artificial‑intelligence community.

21CTO
21CTO
21CTO
Remembering Tang Xiaoyu: The Visionary Behind Modern Facial Recognition

On December 16, 2023, SenseTime (stock code 0020.HK) announced the passing of its beloved founder, AI scientist, and professor Tang Xiaoyu, who died at 23:45 on December 15, 2023.

Born in 1968 in Anshan, Liaoning, Tang entered the University of Science and Technology of China at 17, majoring in precision mechanics and instrumentation. After graduating in 1990, he earned a master’s degree at the University of Rochester, pursued further studies at MIT, and later joined the Chinese University of Hong Kong.

In 2001, Tang founded the Multimedia Laboratory (MMLab) at CUHK, mentoring a generation of prominent AI researchers, including the co‑CEO of Horizon Robotics, the director of Huawei’s Noah’s Ark Lab, JD’s former chief scientist, and DJI’s former AI vision lead. The lab earned the nickname “the Huangpu Military Academy of computer vision.”

From 2005 to 2007, he led the Visual Computing Group at Microsoft Research Asia. In 2008, he became director and researcher at the Multimedia Integrated Technology Research Office of the Shenzhen Institute of Advanced Technology.

His team’s 2014 GaussianFace algorithm achieved a 98.52% accuracy on the LFW benchmark, surpassing human performance for the first time. The DeepID series of face‑recognition algorithms also emerged from his lab, propelling his and his team’s fame.

In 2014, Tang co‑founded SenseTime, one of China’s “AI Four Little Dragons,” holding a 20.63% stake and serving as the controlling shareholder.

In 2016, his AI team was selected as one of the world’s top ten AI pioneer labs, the only one from Asia. In 2020, Tang was listed among the “Top 2000 Most Influential AI Scholars Worldwide.”

Global Pioneer of Face‑Recognition Technology

Beyond research, Tang applied computer‑vision techniques to personal photo management. He launched the “Photo Tagging” project, using his son’s extensive photo collection to develop tagging software, multi‑image segmentation, real‑time image retrieval, and image embedding technologies.

In 2009, IEEE elected him a Fellow. That same year, he, his PhD student He Kaiming, and Microsoft Research Asia’s Dr. Sun Jian won the Best Paper Award at CVPR for their work on single‑image dehazing based on dark channel prior—the first such honor for an Asian scholar in the conference’s 25‑year history.

At the 2023 World Artificial Intelligence Conference (WAIC), Tang highlighted the achievements of his three students: Wang Xiaogang’s breakthrough in surpassing human face‑recognition, He Kaiming’s development of the ResNet architecture that solved deep‑network gradient issues and underpins large‑model families like GPT, and Lin Dahua’s leadership of the influential OpenMMLab open‑source project.

He concluded his speech with a quote from the film “Teacher·Good,” saying, “It’s not that I met you in the best of times; it’s that meeting you gave me the best of times.”

R.I.P.

Original Source

Signed-in readers can open the original source through BestHub's protected redirect.

Sign in to view source
Republication Notice

This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactadmin@besthub.devand we will review it promptly.

artificial intelligenceComputer Visionfacial recognitionAI Pioneer
21CTO
Written by

21CTO

21CTO (21CTO.com) offers developers community, training, and services, making it your go‑to learning and service platform.

0 followers
Reader feedback

How this landed with the community

Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.