Hedvig Kjellström

Professor

Kungliga Tekniska Högskolanhttps://www.kth.se/profile/hedvig/

Works for

Robotics, Perception & Learning

Telephone

+46 8 790 69 06

Unit address

Lindstedtsvägen 24

E-mail

hedvig@kth.se

Room

416c

Researcher ID

ORCID IDWeb of Science ResearcherIDScopus IDGoogle Scholar IDResearchGate

About me

I am a Professor in the Department of Robotics, Perception and Learning, KTH, and also affiliated with Swedish e-Science Research Centre and ELLIS Unit Sweden. A bio and cv is found here.

My main area of research is Computer Vision, which is a sub-field of AI. My group develop computer methods to observe and make inferences and predictions about humans and other animals. As outlined in the Overview page, three partly overlapping themes in my research are AI for Performing Arts, AI for Healthcare, and AI for Animal Science.

I am an Editor-in-Chief for CVIU and was a Program Chair for CVPR 2025.

I teach in the Bachelor program Engineering Mathematics and in the Master programs Machine Learning, Systems, Control and Robotics and Biostatistics and Data Science at KTH. My current courses are found below.

In my free time I play the double bass in different settings, more info here.

News

	Since 2023 I am part of the AI exhibition at Tekniska museet in Stockholm (in Swedish with English subtitles). I can really recommend a visit to this incredibly exciting museum, crossing science, technology, and art.
	On December 19, 2024, I took part in the radio show Förmiddag med Louise Epstein and talked about how AI and humans can live together. Here is the espisode (in Swedish).
	NEW PAPER: Yifan Lu, Taras Kucherenko, Hedvig Kjellström, and Judith Bütepage. RA-ClipScore: Making generative model evaluation more interpretable. arXiv preprint arXiv:2608.12088, 2026.
	NEW PAPER: Mert Mermerci, Emile Pascoe, Fredrik Edström, and Hedvig Kjellström. Real-time control of a virtual orchestra by recognition of conducting gestures. arXiv preprint arXiv:2604.27957, 2026. Data and videos
	NEW PAPER: Chen Ling, Henglin Shi, and Hedvig Kjellström. FIELDS: Face reconstruction with accurate inference of expression using learning with direct supervision. arXiv preprint arXiv:2511.21245v3, 2026.