Loading…

Flash Player 9 (or above) is needed to view presentations.
We have detected that you do not have it on your computer. To install it, go here.

Like this presentation? Why not share!

Oppai-Detect 3

on

  • 21,425 views

Oppai-Detect 3 on GXEB#01

Oppai-Detect 3 on GXEB#01
by yusukebe

Statistics

Views

Total Views
21,425
Views on SlideShare
15,406
Embed Views
6,019

Actions

Likes
14
Downloads
39
Comments
1

29 Embeds 6,019

http://yusukebe.com 4665
http://kamawada.com 478
http://gxeb.org 445
http://d.hatena.ne.jp 260
http://webcache.googleusercontent.com 33
https://twitter.com 32
https://si0.twimg.com 27
http://futureinsight.info 20
http://yusukebe.typepad.jp 18
http://a0.twimg.com 4
http://asakusa-satellite.org 4
http://asakusa-satellite.herokuapp.com 4
https://twimg0-a.akamaihd.net 3
http://cache.yahoofs.jp 3
http://www.iweb34.com 3
http://translate.googleusercontent.com 2
http://ancient-temple-7751.herokuapp.com 2
http://us-w1.rockmelt.com 2
http://web.archive.org 2
http://harupong.tumblr.com 2
http://b.hatena.ne.jp 2
http://www.linkedin.com 1
http://jp.hanrss.com 1
http://www.mefeedia.com 1
http://paper.li 1
http://twitter.com 1
http://static.slidesharecdn.com 1
http://www.onlydoo.com 1
http://ipoir.com 1
More...

Accessibility

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
  • Yes very much
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

Oppai-Detect 3 Oppai-Detect 3 Presentation Transcript

  • Oppai-Detect 3 yusukebe 2010/11/26GXEB #01 @ NAVER JAPAN
  • •  Yusuke Wada•  1981/12/23•  Kamakura, Japan.•  Wadit Inc. h"p://youravhost.net/•  Perl Engineer•  Founder of YourAVHost etc.
  • Oppai-Detect
  • • • • • 
  • •  – • • • 
  • •  OpenCV –  Intel –  Object- Detection•  Training•  Detection
  • OpenCV•  Using “haarcascade_frontalface_alt2.xml” use Image::ObjectDetect;  use Imager;  my $file = $ARGV[0] or die Filename is needed!;  my $img = Imager‐>new;  $img‐>read( file => $file ) or die $img‐>errstr;  my $cascade  = ./haarcascade_frontalface_alt2.xml;  my $detector = Image::ObjectDetect‐>new($cascade);  my @faces    = $detector‐>detect($file);  for my $face (@faces) {      $img = $img‐>box(          xmin   => $face‐>{x},          xmax   => $face‐>{x} + $face‐>{width},          ymin   => $face‐>{y},          ymax   => $face‐>{y} + $face‐>{height},          color  => red,          filled => 0,      ) or die $img‐>errstr;  }  $img‐>write( file => out.jpg ) or die $img‐>errstr; 
  • Result
  • 1. 2. 3. 
  • Term of Images•  Positive Image –  –  7000-8000•  Negative Image –  –  3000
  • 1.• •  Web API –  Google Ajax Search •  64 –  Yahoo API •  1,000•  10,000 –  AV •  AV A •  AV B •  …
  • use Acme::Porn::JP;•  http://github.com/yusukebe/Acme-Porn-JPuse Acme::Porn::JP; …; my $porn = Acme::Porn::JP‐>new; my $actress_list = $porn‐>actress(); for my $actress ( @$actress_list ){     search("$query $actress”); } …; $ perl ‐MAcme::Porn::JP ‐e $p = Ace::Porn::JP‐>new; print scalar @{$p‐>actress};’ 4502 
  • 29,065
  • Haartraining1.  positive.dat/negative.dat   2.  $ opencv_createsamples ‐info positive.dat ‐ vec a.vec 3.  $ opencv_haartraining ‐data haarcascade ‐ vec a.vec ‐bg negative.dat ‐nstages 20  positive.dat [filename] [# of objects] [[x y width height] [... 2nd object] ...] images/01.jpg 1 140 100 45 45 nega2ve.dat images/neg‐01.jpg images/neg‐02.jpg 
  • •  positive.dat negative.dat•  –  positive.dat –  negative.dat•  Web•  HTTP/Web –  use Plack;•  Javascript –  strokeRect/HTML5
  • Haartraining-App•  http://github.com/yusukebe/Haartraining-App yusuke@macmini:~/work/oppai‐detect/temp/Haartraining‐App$ tree ./  ./  |‐‐ data  |   |‐‐ images  |   |   `‐‐ 000001.jpg  |   |‐‐ negative.dat  |   `‐‐ positive.dat  |‐‐ haartraining.psgi  `‐‐ html      |‐‐ haartraining.js      `‐‐ index.html  3 directories, 6 files  $ plackup haartraining.psgi   HTTP::Server::PSGI: Accepting connections at http://0:5000/ 
  • Haartraining-App
  • opencv_haartraining•  Positive Image 2,000•  Negative Image 2,000•  nstage 10 stages
  • Oppai-Detect
  • • •  –  – 
  • •  –  •  / / / / / … – •  – 
  • • • •  Oppai-Detect
  • Oppai-Detect End.