PeopleofSuperintelligence

Researcher

Florian Schroff

Florian Schroff is a computer vision researcher at Google whose work centers on describing people in images and video. He took a DPhil in Oxford's Visual Geometry Group in 2009, completed a two-year postdoc in the UCSD vision group, and joined Google in 2011. His 2015 paper FaceNet mapped face images into a compact Euclidean space and reached 99.63% accuracy on Labeled Faces in the Wild, and he coauthored DeepLabv3+ in 2018, FEELVOS in 2019 and VideoPrism in 2024. In 2020, he was among the authors of a Nature study that applied deep neural networks to 6 million videos from 144 countries to study facial expressions. He is a Senior Research Engineering Manager at Google, where he develops computer vision approaches for people-centric understanding and biometrics.

Now at
Google
Badges

Timeline

  1. 2007–2009

    A DPhil at Oxford

    At the University of Oxford's Visual Geometry Group, he publishes Harvesting Image Databases from the Web at ICCV 2007 with A. Criminisi and A. Zisserman. He graduates with a DPhil in 2009, with a thesis titled Semantic Image Segmentation and Web-Supervised Visual Learning.

  2. 2011

    Postdoc at UCSD, then Google

    He completes a two-year postdoc in the UCSD vision group, researching unconstrained face and object recognition, and joins Google in 2011. His projects there focus on describing people in images and videos.

  3. 2015

    FaceNet face embeddings

    He is first author of FaceNet, published in the CVPR 2015 proceedings with Dmitry Kalenichenko and James Philbin. The system learns a mapping from face images to a compact Euclidean space where distances correspond to face similarity, and it reaches 99.63% accuracy on Labeled Faces in the Wild and 95.12% on YouTube Faces DB.

  4. 2017–2018

    Semantic segmentation with DeepLab

    He coauthors Rethinking Atrous Convolution for Semantic Image Segmentation in 2017 and DeepLabv3+ in 2018 with Liang-Chieh Chen, George Papandreou and Hartwig Adam. DeepLabv3+ extends DeepLabv3 with a decoder module that refines segmentation results along object boundaries.

  5. 2019

    FEELVOS video segmentation

    He is a coauthor of FEELVOS, a CVPR 2019 method for video object segmentation developed with Paul Voigtlaender, Yuning Chai, Hartwig Adam, Bastian Leibe and Liang-Chieh Chen. It segments video without fine-tuning and reaches a J&F measure of 71.5% on the DAVIS 2017 validation set.

  6. 2020

    Facial expressions across cultures

    In December 2020, Nature publishes Sixteen facial expressions occur in similar contexts worldwide, which he coauthors with Alan Cowen, Dacher Keltner, Brendan Jou, Hartwig Adam and Gautam Prasad. The study uses deep neural networks to examine 16 types of facial expression in 6 million videos from 144 countries.

  7. 2024

    The VideoPrism video encoder

    He is a coauthor of VideoPrism, a general-purpose video encoder accepted to ICML 2024. It reaches state-of-the-art performance on 31 of 33 video understanding benchmarks.

  8. As of October 2026

    Research engineering manager at Google

    He is a Senior Research Engineering Manager at Google, where he develops computer vision approaches for people-centric understanding and biometrics. He leads development from research in metric learning and large vision-language models through to product launches.

Career

  1. Since 2011GoogleSenior Research Engineering Manager
  2. Until 2011University of California, San DiegoPostdoctoral Researcher

Education

  1. 2009University of OxfordDPhil

From the biography.

  1. Google had the most people shaping AI every year from 2014 to 2025Includes Google. Our estimates of which organizations employed the most people shaping AI each year from 1955 to 2025, tracing the shift from university labs to companies.
  2. Google has the highest combined h-index among AI organizationsIncludes Google. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.
  3. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes Google. We estimate that 539 of the 6,766 people with a known current organization (8%) hold both a university role and a company role, including 20% of people at companies and 13% of people at universities.
  4. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes Google. 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.

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 Florian Schroff yet. Send us a free-licensed photo.

  1. Florian Schroff research.google. Checked 21 October 2026.
  2. Florian Schroff - Publications florian-schroff.de. Checked 21 October 2026.
  3. FaceNet: A Unified Embedding for Face Recognition and Clustering arxiv.org. Checked 21 October 2026.
  4. CVPR 2015 Open Access Repository openaccess.thecvf.com. Checked 21 October 2026.
  5. Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation arxiv.org. Checked 21 October 2026.
  6. FEELVOS: Fast End-to-End Embedding Learning for Video Object Segmentation arxiv.org. Checked 21 October 2026.
  7. Sixteen facial expressions occur in similar contexts worldwide | Nature nature.com. Checked 21 October 2026.
  8. VideoPrism: A Foundational Visual Encoder for Video Understanding arxiv.org. Checked 21 October 2026.
  9. Florian Schroff - Home florian-schroff.de. Checked 21 October 2026.