Håkon Hukkelås

I’m a computer science Ph.D. candidate at the Norwegian University of Science and Technology (NTNU) working with application of computer vision and generative models for realistic anonymization of faces. My work on anonymization received the best AI-related master thesis from the Norwegian Open AI Lab. I am one of the main instructors of NTNU’s courses on visual computing enrolling roughly 200 students each year, an undergraduate course on Visual Computing, and a graduate course on Computer Vision and Deep Learning. I have a passion for deep learning, open-source, and education.

Publications

  • [PDF] [Github] DeepPrivacy: A Generative Adversarial Network for Face Anonymization
    Håkon Hukkelås, Rudolf Mester, Frank Lindseth. ISVC 2019. Won best paper award.
  • [PDF] [GitHub] Image Inpainting with Learnable Feature Imputation.
    Håkon Hukkelås, Rudolf Mester, Frank Lindseth. DAGM GCPR 2020.
  • [PDF] [GitHub] Deep Active Learning for Autonomous Perception.
    Navjot Singh, Håkon Hukkelås, Frank Lindseth. NIKT 2020.

Open Source Projects

  • DeepPrivacy: A Generative Adversarial Network for Face Anonymization
  • Face detection: A High-Performance Pytorch Implementation of face detection models, including RetinaFace and DSFD

Timeline

  • 2019-now: Ph.D. Candidate at the Department of Computer Science, Norwegian University of Science and Technology.
  • 2014-2019: Norwegian University of Science and Technology. Master of Science in Computer Science. Best AI-related master thesis 2019, awarded by the Norwegian Open AI Lab.
  • Summer 2019: Internship at Cisco; Deep learning for low-resource face detection.
  • 2017-2018: Exchange student at University of California, San Diego.
  • Summer 2018: Internship at Cisco; Deep learning for key-word spotting in audio.
  • Summer 2017: Internship at Itera; .NET core and React Native development.

Talks

  • [Youtube] Introduction to GANs. Given at the 2021 Northern Light Deep Learning Workshop (January 2021)
  • [Youtube] DeepPrivacy. Given to BRAIN NTNU (student AI organization) (October 2020).
  • [Slides] DeepPrivacy: Realistic Face Anonymization. Given to the “personvernskomiteen” appointed by the Norwegian government (October 2020).
  • [Youtube] Image Inpainting With Learnable Feature Imputation. Talk at GCPR 2020 (September 2020)
  • January 2021: Improving upon DeepPrivacy for Realistic Image Anonymization. Given to the AI Master Thesis Award, NTNU. (January 2021)
  • Article: NRKBeta article (Norway’s broadcasting company) on DeepPrivacy.
  • Podcast: Guest on the podcast “Teknologi: Av og for mennesker” about DeepFakes.
  • Article: Vice Motherboard article on DeepPrivacy.