您好,欢迎来到叨叨游戏网。
搜索
您的当前位置:首页数据集(casia,msu,replay,oulu)的详细信息

数据集(casia,msu,replay,oulu)的详细信息

来源:叨叨游戏网

1.casia

总共有50个subject
其中30个作为test,20个作为train
每个subject总共有12个视频(故总共600个视频),其中3个为真,剩余9个为假
具体来说:
前缀为1,2,HR_1的视频为real faces
剩余的视频则为attack faces
总结:

2.msu-mfsd

总共有35个subject
人物标号从1到55(其中有些标号没有,即标号不连续,所以是35个)
其中01,13, 14, 23, 24 ,26, 28 ,29 ,30, 32, 33,35,36,37,39,42,48,49,50,51共20个对象作为test,剩余的15个作为train

对于每个对象,总共有2个真视频和6个假视频
所以总共就是35x8=280个视频
总结:

  1. train:15x8=120个视频,test:20x8=160个视频
  2. real:35x2=70个视频,attack:35x6=210个视频
  3. total:280个视频

3.replay

总共有50个subjects
每个subject有20个attack和4个real
enroll里面有100个视频(enroll中好像是复用real视频)
train,devel,test里面有1200个视频(我们只需要考虑这1200个)
分别对应的人数是15,15,20(也是人物标号不连续的那种)
总共就是1300个视频(enroll视频用不上,所以可以看成1200个视频)
这么多视频分成了四个子数据集
train,test,devel, enroll
The 1300 real-accesses and attacks videos were then divided in the following
way:

  • Training set: contains 60 real-accesses and 300 attacks under
    different lighting conditions;

  • Development set: contains 60 real-accesses and 300 attacks under
    different lighting conditions;

  • Test set: contains 80 real-accesses and 400 attacks under
    different lighting conditions;

  • Enrollment set: contains 100 real-accesses under different
    lighting conditions, to be used exclusively for studying the baseline
    performance of face recognition systems.

总结:

  1. train:15x24=360个视频,devel:15x24=360个视频,test:20x24=480个视频
  2. real:50x4=200个视频,attack:50x20=1000个视频
  3. total:视频

4.oulu

总共有55个subject
20个train,15个devel,20个test
每个subject有18个真视频,72个假视频
总共4950个视频
总结:

  1. train:20x90=1800个视频,devel:15x90=1350个视频,test:20x90=1800个视频
  2. real:55x18=990个视频,attack:55x72=3960个视频
  3. total:4950个视频

因篇幅问题不能全部显示,请点此查看更多更全内容

Copyright © 2019- gamedaodao.net 版权所有 湘ICP备2024080961号-6

违法及侵权请联系:TEL:199 18 7713 E-MAIL:2724546146@qq.com

本站由北京市万商天勤律师事务所王兴未律师提供法律服务