Shmuel Peleg

Shmuel Peleg is a computer vision researcher at the Hebrew University of Jerusalem whose work spans panoramic imaging, surveillance video, and audio-visual speech. He earned a PhD at the University of Maryland in 1979 under Azriel Rosenfeld, spent a year at Ben-Gurion University, and joined the Hebrew University faculty in 1981; he chaired its Institute of Computer Science from 1990 to 1993, co-chaired ICPR in 1994, and became an IAPR Fellow. He coauthored OmniStereo in 2001, won a 2002 Kaye Innovation Award for the technology, and developed patented computer-vision work that supports HumanEyes. His Video Synopsis won a 2010 Kaye Award and was licensed to BriefCam, where he was a co-founder and chief scientist by 2018; in 2017 he coauthored papers on speech from silent video and face-guided audio filtering. He coauthored papers on crowdsourced audio enhancement and clothes-changing person re-identification in 2024; since April 2025 he is a Visiting Faculty Researcher at Google and a full professor at Hebrew University, and he coauthored LiveSVG in May 2026.

Now at: Hebrew University of Jerusalem and Google
Works: 35
Citations: 0
h-index: 0
Badges: Founder
Bridge
Loyal

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

1979–1981

A Maryland PhD and faculty post

He earns a PhD at the University of Maryland in 1979 under Azriel Rosenfeld and spends a year at Ben-Gurion University after graduation. He joins the faculty of the Hebrew University of Jerusalem in 1981.
1990–1994

Institute chair and IAPR Fellow

He chairs the Institute of Computer Science at the Hebrew University from 1990 to 1993. In 1994, he co-chairs the 10th ICPR in Jerusalem and becomes an IAPR Fellow for contributions to image analysis and computer vision and service to the IAPR.
2001–2002

OmniStereo and HumanEyes

In 2001, he coauthors “Omnistereo: Panoramic stereo imaging” with Moshe Ben-Ezra and Yael Pritch in IEEE Transactions on Pattern Analysis and Machine Intelligence; its panoramic stereo pair creates a stereo sensation up to a full 360 degrees. In 2002, he wins a Kaye Innovation Award for OmniStereo. His patented computer-vision technologies also provide technical foundations for HumanEyes, which creates and prints stereoscopic 3D images.
2010–2018

Video Synopsis and BriefCam

In 2010, he wins a Kaye Innovation Award for Video Synopsis, software that separates the static background from moving events and condenses hours of surveillance video into minutes. The invention is licensed to BriefCam, which receives the WSJ’s 2010 Technology Innovation Award in physical security. In 2018, he is BriefCam’s co-founder and chief scientist.
2017

Speech from silent video

In 2017, he and Ariel Ephrat report Vid2speech, an end-to-end convolutional neural network that generates intelligible speech from silent frames of a speaking person; the paper is accepted at ICASSP.
2017

Seeing Through Noise

In 2017, he coauthors with Aviv Gabbay, Ariel Ephrat, and Tavi Halperin a method that estimates a speaker’s voice from face motion and uses it to filter noisy audio.
2024

Audio enhancement and person ReID

In 2024, he coauthors an INTERSPEECH paper on audio enhancement from multiple crowdsourced recordings and GEFF, a WACVW paper that enriches the gallery with face features to improve clothes-changing person re-identification models.
Since April 2025

Google Visiting Faculty Researcher

Since April 2025, he is a Visiting Faculty Researcher at Google and a full professor at the Rachel and Selim Benin School of Engineering and Computer Science at the Hebrew University of Jerusalem. In May 2026, he coauthors LiveSVG, a zero-shot method that generates SVG animations with video diffusion models.