PeopleofSuperintelligence

Researcher

Helmut Grabner

Helmut Grabner is a computer vision researcher at the Zurich University of Applied Sciences (ZHAW) whose on-line boosting methods let visual trackers keep training their classifiers as new frames arrive. After a PhD in computer science at Graz University of Technology in 2008, he coauthored “Semi-supervised On-Line Boosting for Robust Tracking” with Christian Leistner and Horst Bischof, a paper that won the Koenderink Prize in 2018. As a postdoc at ETH Zurich's Computer Vision Lab from 2008 to 2014, he studied what makes images and videos interesting. From 2012 to 2015 he was CTO of upicto, a start-up he co-founded, and he then worked at Logitech until 2019. He now heads the Visual Intelligence and Applications Group at ZHAW, where he and Fitim Abdullahu tested in 2025 how closely GPT-4o's judgements of visual interestingness match human ones.

Timeline

  1. 1999–2008

    Telematics and a Graz PhD

    He studies at Graz University of Technology from 1999 and earns a master's degree (Dipl.-Ing.) in telematics in 2004. He completes a PhD (Dr.techn.) in computer science there in 2008.

  2. 2006–2008

    On-line boosting for tracking

    In 2006, he and Horst Bischof present “On-Line Boosting and Vision” at CVPR, a real-time on-line AdaBoost feature-selection method that trains classifiers incrementally and serves background modelling, visual tracking and object detection. At ECCV 2008, he, Christian Leistner and Bischof introduce semi-supervised on-line boosting, whose SemiBoost tracker greatly reduces drifting in real-time tracking.

  3. 2008–2014

    Postdoc at ETH Zurich

    From October 2008 to December 2014, he is a postdoctoral researcher at ETH Zurich's Computer Vision Lab, working on on-line and incremental machine learning for object detection, tracking, recognition and event detection in visual surveillance. His CVPR papers include “Tracking the Invisible: Learning Where the Object Might Be” in 2010, which estimates an object's position even when it is fully occluded, and “What Makes a Chair a Chair?” in 2011, which learns an affordance detector for sitting.

  4. 2012–2019

    Co-founding upicto

    From May 2012 to January 2015, he is CTO and co-founder of upicto. He works as lead software engineer at Logitech Upicto from February 2015 and, from June 2017 to May 2019, as lead computer vision engineer in Logitech's CTO office.

  5. 2013–2014

    The interestingness of images

    At ICCV 2013, he coauthors “The Interestingness of Images” with Michael Gygli, Hayko Riemenschneider, Fabian Nater and Luc Van Gool, which names aesthetics, unusualness and general preferences as cues and builds an interestingness predictor from them. At ECCV 2014, he and colleagues present “Creating Summaries from User Videos,” which scores video segments by visual interestingness and adds a benchmark of human-made summaries.

  6. 2018

    The Koenderink Prize

    In September 2018, ECCV gives the Koenderink Prize to his ECCV 2008 paper “Semi-supervised On-Line Boosting for Robust Tracking,” coauthored with Christian Leistner and Horst Bischof.

  7. 2023–2026

    Interestingness and GPT-4o

    From January 2023 to June 2026, he leads “Visual Interestingness,” a CHF 600,000 project at ZHAW's Institute for Data Science funded through the SNF Practice-to-Science programme. With Fitim Abdullahu, he publishes “Commonly Interesting Images” at ECCV 2024, an analysis of 2.5k Flickr image collections, and “Visual Interestingness Decoded” at ICCV 2025, which finds partial alignment between human judgements and those of GPT-4o.

  8. As of October 2026

    Visual intelligence group at ZHAW

    He is a professor at ZHAW's School of Engineering in Winterthur, where he heads the Visual Intelligence and Applications Group at the Institute for Data Science and serves as the school's head of entrepreneurship. His group develops an augmented-reality simulator for orthopaedic surgical training that combines adaptive AR guidance with AI-based analysis of human activity, and its art-and-science project Watching the World analyses more than 10,000 public webcams in real time.

  1. Google had the most people shaping AI every year from 2014 to 2025Our estimates of which organizations employed the most people shaping AI from 1961 to 2025, with a year-by-year chart.
  2. The United States employs an estimated 53% of people in our data with a known employment countryOur estimates of where AI careers are concentrated now and how countries’ shares of dated positions changed over time.
  3. 46% of living people in our data with known birth and employment countries work outside their birth countryOur estimate of cross-border careers in AI, from country of birth and the country of each current organization.
  4. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionOur estimates of where the people shaping AI studied and where their careers led, from education records in cited biographies.

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 Helmut Grabner yet. Send us a free-licensed photo.

  1. Helmut Grabner new.videolectures.net. Checked 19 October 2026.
  2. Prof. Dr. Helmut Grabner | ZHAW Zurich University of Applied Sciences zhaw.ch. Checked 19 October 2026.
  3. On-Line Boosting and Vision | ML Anthology mlanthology.org. Checked 19 October 2026.
  4. Semi-Supervised On-Line Boosting for Robust Tracking | ML Anthology mlanthology.org. Checked 19 October 2026.
  5. Tracking the Invisible: Learning Where the Object Might Be | ML Anthology mlanthology.org. Checked 19 October 2026.
  6. What Makes a Chair a Chair? | ML Anthology mlanthology.org. Checked 19 October 2026.
  7. ICCV 2013 Open Access Repository: The Interestingness of Images openaccess.thecvf.com. Checked 19 October 2026.
  8. Creating Summaries from User Videos | ML Anthology mlanthology.org. Checked 19 October 2026.
  9. ECVA | European Computer Vision Association ecva.net. Checked 19 October 2026.
  10. Visual Interestingness - All images are equal but some images are more equal than others | ZHAW zhaw.ch. Checked 19 October 2026.
  11. [2409.16736] Commonly Interesting Images arxiv.org. Checked 19 October 2026.
  12. [2510.13316] Visual Interestingness Decoded: How GPT-4o Mirrors Human Interests arxiv.org. Checked 19 October 2026.
  13. Visual Intelligence and Applications | ZHAW Institut für Data Science IDS zhaw.ch. Checked 19 October 2026.