1. X
  2. Martin Danelljan
Log inSign up
Martin Danelljan
113 posts
user avatar
Martin Danelljan
@MDanelljan
3D Computer Vision at Apple. Previously Research Group Leader at ETH Zurich.
Zürich, Schweiz
martin-danelljan.github.io
Joined October 2020
339
Following
1,234
Followers
RepliesRepliesMediaMedia

New to X?

Sign up now to get your own personalized timeline!

Create account

By signing up, you agree to the Terms of Service and Privacy Policy, including Cookie Use.

Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
Don't miss what's happening
People on X are the first to know.
Log inSign up
  • Pinned
    user avatar
    Martin Danelljan
    @MDanelljan
    Nov 7, 2023
    I’m happy to share that I’m joining Apple in SF Bay Area as Senior Research Engineer in Computer Vision! I will be working on all things 3D Vision and more. I’m looking for extremely talented people to work with me and the team on amazing things. Reach out if you are interested!
    Image
    48K048K
  • user avatar
    Martin Danelljan
    @MDanelljan
    Jul 24, 2021
    6 papers accepted to #ICCV2021 Congratulations to the dedicated and hard-working students! 🎉🏆
  • user avatar
    Martin Danelljan
    @MDanelljan
    Sep 15, 2021
    2.5 years ago we released the PyTracking library. The large interest has been somewhat overwhelming. Excited to continue making it even better! We have now released our #iccv2021 work KeepTrack w/ @chris_mayer_ch! paper arxiv.org/abs/2103.16556 PyTracking github.com/visionml/pytra…
    arXiv logo
    arxiv.org
    Learning Target Candidate Association to Keep Track of What Not to Track
    The presence of objects that are confusingly similar to the tracked target, poses a fundamental challenge in appearance-based visual tracking. Such distractor objects are easily misclassified as...
  • user avatar
    Martin Danelljan
    @MDanelljan
    Sep 13, 2021
    Our paper "Deep Reparametrization of Multi-Frame Super-Resolution and Denoising" was accepted at #iccv2021 as Oral! Instead of using Attention, we fuse multiple frames through deep unrolled optimization, inspired by classical image restoration. arXiv: arxiv.org/abs/2108.08286
  • user avatar
    Martin Danelljan
    @MDanelljan
    Jun 24, 2021
    Code released for our #CVPR2021 oral paper "DeFlow": A method for unpaired learning of image-to-image translation with normalizing flows. Excellent master's thesis by @va_wolf👏 Also with @AndreasLugmayr et al. arXiv: arxiv.org/abs/2101.05796 code: github.com/volflow/DeFlow
    Image
    00:00
  • user avatar
    Martin Danelljan
    @MDanelljan
    Oct 5, 2020
    So happy for @alxandrecarlier that our work DeepSVG was accepted at #neurips2020. Generation of vector graphics is a very underexplored subject. In the paper we introduce a new dataset and method. Check out the amazing repository by @alxandrecarlier! github.com/alexandre01/de…
    This post is from an account you blocked.
  • user avatar
    Martin Danelljan
    @MDanelljan
    Jul 26, 2021
    Now you can do unsupervised learning of dense matching and optical flow networks without relying on photometric consistency assumptions! Check out our #iccv2021 oral paper on Warp Consistency. arXiv: arxiv.org/pdf/2104.03308
    user avatar
    Prune Truong
    @prunetruong
    Jul 24, 2021
    It got accepted as an oral at #ICCV2021!! 🎉🎉
  • user avatar
    Martin Danelljan
    @MDanelljan
    Oct 16, 2020
    Check out @prunetruong and my #NeurIPS2020 paper proposing the GOCor network module for dense matching. Highly related to Meta-learning and Deep Declarative Networks. arXiv: arxiv.org/abs/2009.07823 Code (coming soon): github.com/PruneTruong/GO…
    user avatar
    Prune Truong
    @prunetruong
    Oct 14, 2020
    Very excited that our paper 'GOCor: Bringing Globally Optimized Correspondence Volumes into Your Neural Network' got accepted to NeurIPS 2020 ! This is a joint work with @MDanelljan, Luc Van Gool and Radu Timofte. Check out the paper at arxiv.org/pdf/2009.07823… !
    Image
  • user avatar
    Martin Danelljan
    @MDanelljan
    Dec 7, 2020
    You are welcome to @prunetruong and my NeurIPS poster session for "GOCor: Bringing Globally Optimized Correspondence Volumes into Your Neural Network". A meta learning inspired module for SOTA dense correspondences and flow. Time: Tue 9 am PT / 18.00 CET
    youtube.com
    [NeurIPS 2020] GOCor: Bringing Globally Optimized Correspondence Volumes into Your Neural Network
    This is a 3 minutes video for our NeurIPS paper: GOCor: Bringing ...
  • user avatar
    Martin Danelljan
    @MDanelljan
    Dec 7, 2020
    Check out the poster session of our NeurIPS 2020 paper "DeepSVG: A Hierarchical Generative Network for Vector Graphics Animation". Time: Mon 9 pm PT / Tue 06:00 CET @alxandrecarlier made an awesome video as well!
    Image
    DeepSVG (NeurIPS 2020 3min video)
    From youtube.com
  • user avatar
    Martin Danelljan
    @MDanelljan
    Jun 22, 2022
    See you soon at one of our three posters in the afternoon session at #CVPR22! P75: Probabilistic Warp Consistency for Weakly-Supervised Semantic Correspondences P77: Transforming Model Prediction for Tracking P85: Adiabatic Quantum Computing for Multi Object Tracking
  • user avatar
    Martin Danelljan
    @MDanelljan
    Mar 17, 2022
    I'm very proud that @prunetruong received the @Apple AI/ML PhD Fellowship for 2022!
    user avatar
    Prune Truong
    @prunetruong
    Mar 17, 2022
    I'm very grateful to receive the @Apple AI/ML PhD Fellowship for 2022! This would not have been possible without my incredible mentors and collaborators! Special thanks to my supervisor @MDanelljan. machinelearning.apple.com/updates/apple-…
  • user avatar
    Martin Danelljan
    @MDanelljan
    Jun 20, 2021
    Using probabilistic regression to estimate dense correspondences. Turns out that the most difficult challenge is to learn robust confidences in the unsupervised setting. Check out our paper and oral presentation at #CVPR2021 w/ @prunetruong to see how!
    user avatar
    Prune Truong
    @prunetruong
    Jun 19, 2021
    Looking for dense and trustworthy correspondences? Check out our #CVPR2021 oral paper w/ @MDanelljan: Learning Accurate Dense Correspondences and When to Trust Them Arxiv: arxiv.org/abs/2101.01710 Code: github.com/PruneTruong/De… Project page: prunetruong.com/research/pdcnet
    Image
    00:00
  • user avatar
    Martin Danelljan
    @MDanelljan
    Jan 29, 2021
    Our upcoming NTIRE 2021 workshop at CVPR hosts multiple challenges that are currently ongoing. I am co-organizing two completely new ones. Burst Super-Resolution: github.com/goutamgmb/NTIR… Learning the Super-Resolution Space: github.com/andreas128/NTI… You're welcome to participate!
Advertisement
Advertisement