Christoph Feichtenhofer
Christoph Feichtenhofer studies how AI systems learn from video. His work includes efficient recognition networks and learning representations from partly hidden video inputs. Feichtenhofer earned his TU Graz doctorate from 2014 to 2017 with a thesis titled Deep Learning for Video Recognition. He also visited York University and Oxford for research. He worked as a research scientist at Facebook AI Research from 2018 to 2021. He then worked as a research scientist and director at Meta. Feichtenhofer authored X3D, published at CVPR 2020. The work develops video-recognition networks by progressively expanding a small image-classification architecture across spatial, temporal and network dimensions.
- Works
- 81
- Citations
- 27,860
- h-index
- 42
Publication numbers from OpenAlex, as of 14 October 2026.
Timeline
-
2014–2017
Doctorate on video recognition
Feichtenhofer earned his TU Graz doctorate from 2014 to 2017 with a thesis titled Deep Learning for Video Recognition. He also visited York University and Oxford for research.
-
2018–2021
Research at FAIR
He worked as a research scientist at Facebook AI Research from 2018 to 2021. He then worked as a research scientist and director at Meta.
-
2020
Efficient video networks
Feichtenhofer authored X3D, published at CVPR 2020. The work develops video-recognition networks by progressively expanding a small image-classification architecture across spatial, temporal and network dimensions.
-
2022
Learning from masked videos
He shared technical lead contribution with Haoqi Fan on Masked Autoencoders As Spatiotemporal Learners, with Yanghao Li and Kaiming He. The model learns by reconstructing masked space-time patches.
Sources and credits
Photo: feichtenhofer.github.io; creator not stated. License: Not stated by source. Source: feichtenhofer.github.io. Cropped to a square and resized. Ask us to remove or replace this photo.