190119CodeforIkoma_Mapillary

瀬戸寿一 (2019年1月)「オープンなストリート画像共有サービスMapillaryとは？」、Code for Ikoma定例会、市民活動推進センターららポート、2019年1月19日

Transcript

1. / 0 (2
   !
   
   
   "
   G C I G 9
   / S O M / S 1 / )/ /
   

   View Slide

2. C
   -- 0 : 1
   A B 4 C 1 :2
   4 3460 2
   /362 2 1 90
   630 km 2
   

   View Slide

3. 9
   https://twitter.com/jesolem
   03/12/
   

   View Slide

4. 2 0 / 1
   #$! !
   !
   
   )(!'& %"
   

   View Slide

5. 2 0 / 1
   -
   

   View Slide

6. 2 0 / 1
   https://www.mapillary.com/app/
   

   View Slide

7. 2 0 / 1
   
   
   

   View Slide

8. 2 0 / 1
   
   
   

   View Slide

9. 12/ 0
   
   
   https://blog.mapillary.com/update/2018/12/06/mapillary-in-japan.html
   

   View Slide

10. 12/ 0
    
    
    

    View Slide

11. 12/ 0
    
    
    

    View Slide

12. 12/ 0
    #
    https://mapillary.github.io/mapillary_greenhouse/
    

    View Slide

13. 230 / 1
    8
    https://www.johnscreekga.gov/NewsAndEvents/News-Archive/2017-
    News/Complete-the-Map-residents-can-participate-in-str
    

    View Slide

14. 2 0 / 1
    https://mapillary.github.io/mapillary_greenhouse/global-challenge/q2-2018/
    Score = (total unique KM contributed + (images/1000)) * (1 + (count of local participants/10))
    

    View Slide

15. 2 0 / 1
    Mapillary images are available under the Creative Commons Attribution-ShareAlike 4.0 (CC BY-SA 4.0)
    

    View Slide

16. 2 0 / 1
    
    
    https://www.mapillary.com/pricing
    

    View Slide

17. 2 0 / 1
    
    
    http://openaccess.thecvf.com/content_ICCV_2017/papers/Neuhold_The_Mapillary_Vi
    stas_ICCV_2017_paper.pdf
    

    View Slide

18. 2 0 / 1
    1 0 2 5
    0 1 0
    The Mapillary Vistas Dataset for Semantic Understanding of Street Scenes
    doi: 10.1109/ICCV.2017.534
    

    View Slide

19. 12/ 0
    - - - - ( -
    ( - - )
    https://research.mapillary.com/lsun.html
    

    View Slide

20. 12/ 0
    Mapillary Vistas Dataset
    25,000 & 100
    
    https://research.mapillary.com/
    

    View Slide

21. 12/ 0
    0 2 1 ,
    •Road markings: increased from 6 classes to 21 classes including speed limits and direction
    arrows, which are essential classes for building navigation and HD maps.
    •Driveway: a new class that we’ve introduced to help enhance data extraction for last-mile
    •Barriers: increased from 6 to 11 classes. Detailed refinement of the barrier classes is the first
    step to enhance the perception capacity for autonomous driving under different scenarios.
    •Signages: the billboard class is now refined to advertisement, storefront, and general
    information signages. This facilitates automatic and in-depth location-aware data extraction
    for marketing, business, and point-of-interest analysis.
    •Traffic light states: dynamic properties are associated with static objects. In this case, we
    have refined 100K traffic light annotations with different state
    

    View Slide

22. 12/ 0
    http://www.robustvision.net/
    

    View Slide

23. 230 / 1
    https://blog.mapillary.com/update/2018/06/14/robust-cv.html
    

    View Slide

24. 2 0 / 1
    :
    https://arxiv.org/abs/1712.02616
    

    View Slide

25. 2 0 / 1
    3/ ) ( 0 . 3 .
    https://github.com/mapillary/mapillary-js
    

    View Slide

26. 2 0 / 1
    https://blog.mapillary.com/update/2018/06/12/mapillary-for-organizations.html
    

    View Slide

27. 67 wl k m
    3 . (
    https://github.com/enricofer/go2mapillary
    • 0 1 ( o
    Qi tpI
    JS
    • x M e V
    i I
    • 0 x 0 1 G
    J 0375297 G
    • / 9 C.1QW d
    fV f M y
    • b a r
    Thanks! Christopher Beddow
    

    View Slide

28. 4 : 2 kyh g
    . . / .
    • f f O S
    o 0 2
    Oe d
    • 2 9 2 F 0 2
    F19:9 492
    P 9 :
    • r Mvs R tMO
    ipl mnwI
    • f c Waf
    8 2 2 9 .9
    2 .9 / C9
    Thanks! Christopher Beddow
    

    View Slide

29. /0 1 4
    • S P A
    • 9 90
    0 99 ad
    • 2. 9
    • M O Ce
    k f c
    • ek m o C
    j i C I
    . . / .
    Thanks! Christopher Beddow
    

    View Slide

30. 230 / 1
    
    
    

    View Slide

31. 230 / 1
    Thank you!
    tosseto@csis.u-tokyo.ac.jp
    http://researchmap.jp/tosseto
    https://speakerdeck.com/tosseto
    

    View Slide