Introduction:
The ULFMT was gathered at MiViM research chair of Canada at Université Laval during the PhD study of Dr. Reza Shoja Ghiass in the laboratory of Prof. Hakim Bendada and Prof. Xavier Maldague. The database has been gathered by an Indigo Phoenix Thermal Camera between 2010 and 2014. It has some unique properties which distinguish it from the other databases:
(i) It has been acquired with an expensive and high resolution MWIR camera. So, it is possible to visualize the fine under-skin details in thermal images.
(ii) The database consists of 238 subjects, which makes it the largest thermal database in the world. It is divided into Genuine (134 Subjects) and Impostor (104 Subjects).
(iii) This is the only thermal video database in the world. So, you can use it for tracking research projects.
(iv) Most of the subject faces were captured in multiple sessions ranging from two years to four years. So, it is possible investigate the robustness of thermal features after timelapse.
(v) Pose, facial expression, ethnicity, aging, time-Lapse, and opaqueness to Eyeglasses are all considered in this database.
For further details about this database, please see the following paper:
Ghiass, Reza & Bendada, Hakim & Maldague, X. (2018). Université Laval Face Motion and Time-Lapse Video Database (UL-FMTV). 10.21611/qirt.2018.051. (download PDF)
To download the database, you will need a username and password. Please write an email to Prof. Xavier Maldague (Xavier.Maldague@gel.ulaval.ca) and mention that you accept the following terms of use.
Terms of Use:
The database is made available for non-commercial use only. If you use this dataset, please cite the following thesis:
Ghiass, Reza Shoja. "Face Recognition Using Infrared Vision." Ph.D. diss., Université Laval, 2014. (download PDF)
Database Contents:
main.m is the main application to read the frame sequences either for a single image (*.sfimg) or a video sequence (*.sfmov).
sfmov_read.m is the video reader function called by main.m .
The database is stored in ThermalFaces.zip (you will need a username and password. Please write an email to Prof. Xavier Maldague (Xavier.Maldague@gel.ulaval.ca))
Acknowledgement:
We acknowledge the students of Université Laval, LVSN members, NSERC, Matthieu Klein, Dr. Ognjen Arandjelovic and Denis Ouellet for their support.