Nick Weir

@NickWeir09

Senior Data Scientist , SpaceNet 4 challenge manager. Former Health Data , PhD . Skeeball enthusiast. Opinions my own.

Washington, DC
Vrijeme pridruživanja: studeni 2015.

Tweetovi

Blokirali ste korisnika/cu @NickWeir09

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @NickWeir09

  1. proslijedio/la je Tweet
    prije 11 sati

    The podcast series is back with Season 2! To kick off the new season, our own & talk about the value of SAR data and the upcoming challenge with , 's CEO and Founder.

    Prikaži ovu nit
    Poništi
  2. proslijedio/la je Tweet
    prije 17 sati

    It is everyone! As you may know, the upcoming challenge will feature our first multimodal dataset: high resolution imagery from as well as high resolution SAR data from (find out more here: ).

    Poništi
  3. proslijedio/la je Tweet
    prije 17 sati

    Sharks can be found in two locations: the Northern and Southern Hemisphere. But thanks to SAR is about to be found everywhere! Excited to be launching , look out for lots of great content from the partners!

    Poništi
  4. proslijedio/la je Tweet
    prije 17 sati

    Big thanks to for making this dataset available as well as of the SpaceNet Partners: , , , , , and !

    Prikaži ovu nit
    Poništi
  5. proslijedio/la je Tweet
    prije 17 sati

    In preparation for the challenge, we are excited to launch ! We will be releasing a variety of a content throughout the week related to the upcoming challenge and Synthetic aperture radar (SAR) data.

    Prikaži ovu nit
    Poništi
  6. proslijedio/la je Tweet
    28. sij
    Prikaži ovu nit
    Poništi
  7. proslijedio/la je Tweet
    29. sij

    I’ve told this story in private a few times but it’s time to tell it publicly. How I Got a Lifetime Ban from AWS: A Cautionary Tale for Idiots on Computers (Illustrated with Larry David GIFs)

    Prikaži ovu nit
    Poništi
  8. proslijedio/la je Tweet
    29. sij

    In part two of the blog post series, builds and analyzes several baseline models using the real satellite imagery data for detecting aircraft, aircraft role, and aircraft type. Stayed for upcoming posts on our blog, The DownLinQ.

    Poništi
  9. proslijedio/la je Tweet
    23. sij

    We've updated our "Resources" page with 3 new research papers and reports: 1. "Road Network and Travel Time Extraction From Multiple Look Angles with SpaceNet Data" by , , , , & .

    Prikaži ovu nit
    Poništi
  10. proslijedio/la je Tweet
    23. sij

    Really proud of the team! We've released 3 new research papers and reports on our website focused on: (1) for road network extraction & routing; and (2) robustness of limited training data. Let us know if you have any questions.

    Poništi
  11. proslijedio/la je Tweet
    23. sij

    Another great post from on submissions. An important take away from this post, and the wider blog post series, is how fragile model performance was depending on the geography being tested.

    Prikaži ovu nit
    Poništi
  12. proslijedio/la je Tweet
    23. sij

    In the 3rd post of the series, IQT ' , who served as the Challenge Manager for SpaceNet 5, explores neighborhood-level score variances in competitor's Average Path Length Similarity (APLS) scores across the featured 4 cities!

    Prikaži ovu nit
    Poništi
  13. 22. sij

    Well I guess that answers that! However, I *strongly* recommend that any or explore the complementary domain - there’s a lot to learn there which you’ll never get from “standard” computer vision.

    Prikaži ovu nit
    Poništi
  14. proslijedio/la je Tweet
    21. sij

    We partnered with to compare building coverage in OpenStreetMap with data they extracted from imagery in cities around Cambodia. Pretty nifty tool you can play with here: . More at Orbital's blog here:

    Poništi
  15. proslijedio/la je Tweet
    21. sij

    We are hiring a Program Manager with geospatial/remote sensing experience in Berlin!

    Prikaži ovu nit
    Poništi
  16. proslijedio/la je Tweet
    21. sij

    's blog post series analyzing the results from the challenge is a must-read for those focused extracting route networks and routes from data such as satellite imagery.

    Poništi
  17. 20. sij

    Headed by , the team did an analysis on off-nadir road network extraction from imagery using : . Interestingly, road extraxn off-nadir is better than building extraction at those angles.

    Poništi
  18. 17. sij

    “...for [] scoring performance, only a subset of data will be available to participants to ultimately map buildings. We hope that...will incentivize new data fusion methods and other approaches such as domain adaptation. ” ,

    Poništi
  19. 17. sij

    "...many of the challenges with imagery (i.e., relative size, number and variability of objects, size of images, lack of well labelled data) are true for imaging, specifically data."

    Poništi
  20. 17. sij

    This post was a bit more technical than the last one, but we're considering really getting into the weeds on techniques developed in each field that might benefit the other (e.g. COGs and how they might help bio) - is there interest in this?

    Prikaži ovu nit
    Poništi

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:

    ·