Experimental frameworks

• AVletters database:

  • 3 repetitions/talker of the letters A-Z; 10 talkers: 5 males (2 with moustache) and 5 females
  • mouth images only; image size: 80x60; lighting: studio ceiling

• Tulips1 audiovisual database:

  • 12 speakers saying the first 4 digits in English; 2 repetitions/subject/digit
  • gray level mouth images only; image size: 75x100

• M2VTS audiovisual database:

  • 37 speakers saying the digits 0-9 in French; 5 repetitions/subject/digit, at 1 week interval; both frontal & rotated view; with/without glasses
  • color videosequence; each frame=entire face (size 286x350)

• XM2VTS audiovisual database: the “English version” of M2VTS

Department of Informatics

Aristotle University of Thessaloniki

• Audio-visual databases (not very large size):

Previous slide

Next slide

Back to first slide

View graphic version