Andrew Davison

@AjdDavison

From SLAM to Spatial AI; Professor of Robot Vision, Imperial College London; Director of the Dyson Robotics Lab; Co-Founder of SLAMcore.

London, England
doc.ic.ac.uk/~ajd/
Vrijeme pridruživanja: svibanj 2013.
96 fotografija ili videozapisa Fotografije i videozapisi

Tweetovi

Blokirali ste korisnika/cu @AjdDavison

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @AjdDavison

  1. Prikvačeni tweet
    31. lis 2019.

    New on arXiv: FutureMapping 2: Gaussian Belief Propagation for Spatial AI, with GBP is ready for the new generation of super-parallel AI chips and edge networks, for general, graph-based . Try my Python demos and see what you think!

    7 replies 42 proslijeđena tweeta 114 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  2. 4. velj

    Useful thread and I will try some of these... I love pygame but it definitely doesn't work properly on my Mac.

    Does anyone know a simple, *fast* way to display a numpy array as pixels, e.g. for watching cellular automata in real time? (I usually use pygame but it has some weird issues on retina displays, so I'm wondering if there's a better option.)
    Prikaži ovu nit
    1 reply 3 korisnika označavaju da im se sviđa
    Poništi
  3. 3. velj

    In a pure rotation SLAM approach, jointly estimated camera motion and a gradient map which could be integrated to HDR/superresolution intensity. Paper link: With Sio-How Ieng, Ryad Benosman. Still my favourite event camera paper 😀

    2 proslijeđena tweeta 7 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  4. 3. velj

    started his PhD just when I got excited about event cameras as custom hardware for , and his brilliant first paper (Best Industry Paper at BMVC 2014) showed the first high quality image reconstruction results from an event camera.

    1 reply 4 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  5. 29. sij

    Looks like impressive learned visual grasping technology from CovariantAI.

    After 2+ years in stealth, we’re excited to launch today! Thank you to our team, customers, partners and investors, we couldn’t have done it w/o your support and trust. Exciting milestone, even more exciting journey ahead!
    1 reply 2 proslijeđena tweeta 12 korisnika označava da im se sviđa
    Poništi
  6. proslijedio/la je Tweet
    27. sij

    Real-time path tracing, the holy grail of graphic techniques, has finally reached our voxel world. Let's all take that walk now.

    68 replies 835 proslijeđenih tweetova 3.963 korisnika označavaju da im se sviđa
    Poništi
  7. 27. sij

    Robert then showed how to robustly fit parametric room models to omnidirectional depth maps, even when there is little texture, which is useful for room identification and planning (this is our old kitchen!) , , ICRA 2017. 33/n

    1 reply 4 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  8. 27. sij

    Very inspiring talk from and a reminder to do research on crazy/fun things.

    Ramesh Raskar (associate professor at MIT) eventually managed how to deal with failures by becoming a pathfinder: always tense towards your goal but also critical and ready to adjust his strategy to reach it. An incredible life lesson in his talk at FAIL!
    8 korisnika označava da im se sviđa
    Poništi
  9. 27. sij

    's PhD work, funded by Dyson, used omnidirectional images to rapidly obtain global geometry when a robot enters a room. These are omni depth maps and confident free space from a short circular manouevre, which could then feed into planning. 32/n

    1 reply 1 proslijeđeni tweet 10 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  10. 27. sij

    Enjoyed this thread, and a good reminder of how technology builds on itself in layers with innovation/fragmentation alternating with consolidation/stability.

    Excellent thread on the history of operating systems - and what might be coming next
    2 korisnika označavaju da im se sviđa
    Poništi
  11. 24. sij

    Nice clip of Owen from SLAMcore on the positive impact of robotics.

    Our CEO made his TV debut this week! See him here discussing rights on :
    7 korisnika označava da im se sviđa
    Poništi
  12. proslijedio/la je Tweet
    24. sij

    Our Visionaries Programme is now open! We're looking for commercial development partners to be the first to integrate our software into their platforms. Apply now at

    5 proslijeđenih tweetova 6 korisnika označava da im se sviđa
    Poništi
  13. 23. sij

    Yes, love these visualisations, and it seems like just the right type of scene representation to be heading towards to me .

    Look at 's cool 3d scene graph work here. She also makes amazing visualizations of her work!
    6 korisnika označava da im se sviđa
    Poništi
  14. 22. sij

    Great visualisations of depth from single photos.

    0:06
    3D Ken Burns Effect, based on Python ML library delivers astonishing effects! These are just some of my photos becoming parallaxed. (Thread)
    Prikaži ovu nit
    10 proslijeđenih tweetova 53 korisnika označavaju da im se sviđa
    Poništi
  15. 21. sij

    That's nice!

    Tiling a 16x16 board, with one unit removed, using L shaped tiles. Recursively divide into four smaller boards, place the tile on those not containing the omitted square. This reduces to four smaller instances of the problem. Works with any square board of size 2^n.
    Prikaži ovu nit
    1 korisnik označava da mu se sviđa
    Poništi
  16. 20. sij

    Looks like interesting work on computational patterns for efficient training from sequential video data. Efficient continual learning of persistent models from incremental (real-time) data is exactly what I think I've always been most interested in.

    Given the smoothness of videos, can we learn models more efficiently than with ? We present Sideways - a step towards a high-throughput, approximate backprop that considers the one-way direction of time and pipelines forward and backward passes.
    1 reply 3 proslijeđena tweeta 25 korisnika označava da im se sviđa
    Poništi
  17. 16. sij

    Paper link: To appear in RAL, 2020.

    5 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  18. 16. sij

    DeepFactors with , Tristan Laidlow, from the Dyson Robotics Lab. Unified real-time monocular SLAM with a general factor graph formulation (GTSAM), pushing what's possible combining deep networks with probabilistic optimisation.

    30 proslijeđenih tweetova 107 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  19. 15. sij

    Very nice work. We did something similar with Jan Jachnik a few years ago in (much simpler) case of a shiny planar object .

    1:12
    Excited to share some work with colleagues at the University of Washington which is now up on arxiv! "Seeing the World in a Bag of Chips" Here's a video snippet showing environment estimation and novel view synthesis results.
    Prikaži ovu nit
    7 proslijeđenih tweetova 23 korisnika označavaju da im se sviđa
    Poništi
  20. proslijedio/la je Tweet
    14. sij

    SLAMcore makes mapping systems for and drones. Hear where this is headed from co-founder Owen Nicholson 📈🤖

    1 reply 9 proslijeđenih tweetova 7 korisnika označava da im se sviđa
    Poništi
  21. 12. sij

    Really happy to welcome Mark to our department.

    I'm excited to share that I have joined Imperial College London as a lecturer (asst prof)! I'm convinced it will be a great environment to continue working on GPs, Bayesian Deep Learning, and model-based RL. Do get in touch if you're interested joining to do a PhD!
    Prikaži ovu nit
    2 proslijeđena tweeta 19 korisnika označava da im se sviđa
    Poništi

@AjdDavison još nije objavio nijedan Tweet.

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·