XM2VTSDB: Available DataSets


CDS001 + CDS006: Frontal Image Set (sample)+ Frontal Image Set II(sample)
Medium: 1 x DVD
Cost: £100 Academic; £200 Industry
CDS001
This dataset contains 1 frontal view for each of the 295 subjects and each of the four sessions. This image was taken at the beginning of the head rotation shot. There are a total of 1,180 colour images. The images are at resolution 720x576 pixels. They have been individually compressed using zip. They can be uncompressed using pkzip on a DOS/WIN95/WINNT machine and by using 'unzip' on a UNIX based machine. The images are stored in PPM (portable pixmap format).
 
CDS006
This dataset contains 1 frontal view for each of the 295 subjects and each of the four sessions. This image was taken from the middle of the head rotation shot when the subject had returned their head to the middle. They are different from those contained in CDS001. There are a total of 1,180 colour images. The images are at resolution 720x576 pixels. They have been individually compressed using zip. They can be uncompressed using pkzip on a DOS/WIN95/WINNT machine and by using 'unzip' on a UNIX based machine. The images are stored in PPM (portable pixmap format).

CDS003: Audio Data (sample)
Medium: 1 x DVD
Cost: £100 Academic; £200 Industry
This dataset contain all the audio data from all 4 sessions. The data has been split into seperate sentences. There are a total of 7,080 files. The quality is 16 bit, mono, 32KHz. Files have been compressed using zip and are stored in WAV format.

CDS005: 3D VRML Models (sample)
Medium: 1 x CD
Cost: £100 Academic; £200 Industry
This set contains 1xVRML Model of the head for 293 of the subjects present at the recordings of the XM2VTSDB.

CDS002 + CDS008: Side Profile Image Set(sample) + Darkened Frontal View Images (sample)
Medium: 1 x DVD
Cost: £100 Academic; £200 Industry
CDS002
This dataset contains 1 left and 1 right profile image per person, per session. A total of 2,360 images.

The images are at resolution 720x576 pixels and have been individually compressed using zip. They can be uncompressed using 'pkzip' on a DOS/WIN95/WINNT machine and by using 'unzip' on a UNIX based machine. The images are stored in PPM (portable pixmap format).

 
CDS008
This dataset contains 4 frontal views for each of the 295 subjects taken from the final session. In two of the images the studio light illuminating the left side of the face was turned off. In the other two images the light illuminating the right side of the face was turned off.

There are a total of 1,180 colour images. The images are at resolution 720x576 pixels. They have been individually compressed using zip. They can be uncompressed using pkzip on a DOS/WIN95/WINNT machine and by using 'unzip' on a UNIX based machine. The images are stored in PPM (portable pixmap format).


MPEG7: MPEG7 Test Set (sample)
Medium: 1 x DVD
Cost: £100 Academic; £200 Industry
The data required for the MPEG 7 experiments. So far the orignal 2950 xm2vts images and the 520 Banca images. Plus the normalised images.

DVD001 + DVD002 + DVD003a + DVD003b:
Sentence 3 (sample)
Head Rotation Shots (sample)
Sentences 1,2, 4 and 5 - Client Set (sample)
Sentences 1,2, 4 and 5 - Imposter and Test Set (sample)
Medium: 1 x External USB2 Hard drive
Cost: £1000 Academic; £2000 Industry
Information: The sequences are stored in DV encoded AVI
DVD001
This set contains the full motion video (with audio) of sentence 3 for each of the 295 subjects and each of the four sessions. There are a total of 1,180 sequences. Each sequence is approximately 5 seconds long and contains the subject speaking the sentence

............ "joe tooks fathers green shoe bench out".

The sequences are stored in DV encoded AVI file format.

 
DVD002
This set contains the full motion video of the first head rotation shot. Each sequence contains a subject rotating his/her head from the centre, to the left, to the right, then up to down, before returning to the centre. This set contains all 295 subjects for all 4 sessions, a total of 1,180 video clips.
 
DVD003a
This set contains two video sequences (with audio) for 200 subjects (the clients from the XM2VTDB protocol) across all four sessions. The first sequence consists of the subject speaking sentences 1 and 2. The second sequence contains sentence 3 and 4. There are a total of 1,600 files.

sentences 1 and 4 ...... "zero one two three four five size seven eight nine".

sentences 2 and 5 ...... "five zero six nine two eight one three seven four".

 
DVD003b
This set contains two video sequences (with audio) for 95 subjects (the imposters and test set from the XM2VTSDB protocol) across all four sessions. The first sequence consists of the subject speaking sentences 1 and 2. The second sequence contains sentence 3 and 4. There are a total of 760 files.

sentences 1 and 4 ...... "zero one two three four five size seven eight nine".

sentences 2 and 5 ...... "five zero six nine two eight one three seven four".