INVEDRIFAC - A Video and Image Database of Faces of In-vehicle Automotive Drivers
Contributors: Aurobinda Routray, Supratim Gupta, Anirban Dasgupta, SL Happy
Drowsiness detection in automotive drivers using image-based methods has gained popularity over the past decade. Algorithms to detect the level of drowsiness in drivers using features including eye closure rates, eye gaze, head pose, blink frequency, etc. are quite popular. However, due to the unavailability of a standard database to evaluate such algorithms, proper benchmarking of the same has not been carried out. This paper describes such a database created under on-board in-vehicle conditions during the day as well as night. We have named it the INVEDRIFAC (IN VEhicle DRIver FACe) Database. The database consists of images as well as videos, with appropriate ground truths which will aid in the testing of speed as well as accuracies of the algorithms. We have used passive Near Infra-red (NIR) illumination for illuminating the face during night driving. INVEDRIFAC contains variations in head orientations and with different facial expressions, facial occlusions and illumination variation. This new database can be a treasured resource for development and evaluation of algorithms for the video based detection of driver fatigue.
The database consists of videos of 30 drivers in the age group of 20 to 40 years age under various conditions of day as well as night in NIR illumination. Prior consent was obtained from the subjects before the experiments.
ACKNOWLEDGEMENTS
The authors would like to acknowledge Department of Information Technology, Government of India for financially supporting the research. The authors would also like to acknowledge the subjects for voluntary participation in the experiment for database creation.