INSTRUCTIONS FOR DOWNLOADING TunnelDBSoftBio

1. Download license agreement, send by email one signed and scanned copy to atvsuam.es according to the instructions given in point 2.
    

2. Send an email to atvsuam.es, as follows:
   Subject: [DATABASE download: TunnelDBSoftBio]
   
   Body: Your name, e-mail, telephone number, organization, postal mail, purpose for which you will use the database, time and date at which you sent the email with the signed license agreement.
    

3. Once the email copy of the license agreement has been received at ATVS, you will receive an email with a username, a password, and a time slot to download the database.
    

4. Download the database, for which you will need to provide the authentication information given in step 4. After you finish the download, please notify by email to atvsuam.es that you have successfully completed the transaction.
    

5. For more information, please contact: atvsuam.es

DESCRIPTION OF TunnelDBSoftBio

The database is comprised of 13,340 soft biometric signals (from 23 physical trait labels) from 58 users annotated by 10 different annotators, as described in [TIFS2013].

FILES FORMAT

The physical trait labels are stored in .mat (MATLAB) file. After loading each file, soft biometric signals corresponding to the 23 physical trait labels are stored respectively in one matrix (TunnelDBSoftBio) of dimensions 26 by 580, i.e., 3 rows of identificators and 23 rows of physical trait labels by 10 annotators from 58 subjects. Each column in the matrix represents a soft biometric signature from the 58 subjects.

As described in [TIFS2013] and [TPAMI2013], the database was annotated against recordings taken of the individuals in laboratory conditions. The annotation process was as follows: an annotator visualized the full video of a subject walking toward the camera and then generated one set of soft labels per each video. It is important to note that the process followed here is independent of the distance. A range of discrete values is given to each trait label, e.g. Arm length marked as 1 (very short), 2 (short), 3 (average), 4 (long), and 5 (very long).

INDEX FILES

The nomenclature followed to identify the physical soft biometric label is as follows:

• ROW 1 = sss: subject identifier

• ROW 2 = xxxx: sample identifier

• ROW 3 = aaa: annotator identifier

• ROW 4...26 = soft labels based on the right table

• ROW 4 = arm length [1...5]

• ROW 5 = arm thickness [1...5]

• ROW 6 = chest [1...5]

• ROW 7 = figure [1...5]

• ROW 8 = height [1...5]

• ROW 9 = hips [1...5]

• ROW 10 = leg length [1...5]

• ROW 11 = leg direction [1...5]

• ROW 12 = leg thickness [1...5]

• ROW 13 = muscle build [1...5]

• ROW 14 = proportions [1...2]

• ROW 15 = shoulder shape [1...5]

• ROW 16 = weight [1...5]

• ROW 17 = age [1...7]

• ROW 18 = ethnicity [1...7]

• ROW 19 = sex [1...2]

• ROW 20 = skin colour [1...4]

• ROW 21 = facial hair colour [1...6]

• ROW 22 = facial hair length [1...5]

• ROW 23 = hair colour [1...6]

• ROW 24 = hair length [1...5]

• ROW 25 = neck length [1...5]

• ROW 26 = neck thickness [1...5]

REFERENCES

For further information regarding the database, the different features approaches extracted from the data and the experimental work we refer the reader to:

• [TIFS2013] P. Tome, J. Fierrez, R. Vera-Rodriguez, M. S. Nixon, "Soft Biometrics and their Application in Person Recognition at a Distance", in IEEE Transactions on Information, Forensic and Security (TIFS), 2013. (DOI: 10.1109/TIFS.2014.2299975)

• [TPAMI2013] D. Reid, M. S. Nixon and S. V. Stevenage, "Soft Biometrics; Human Identification using Comparative Descriptions", IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013. (DOI: 10.1109/TPAMI.2013.219)

Please remember to reference articles [TIFS2013] and [TPAMI2013] on any work made public, whatever the form, based directly or indirectly on any part of the TunnelDBSoftBio.

Data collected at CSPC - University of Southampton and processed at ATVS - Universidad Autonoma de Madrid.