Sunglass detection method for automation of video surveillance system

• Main Authors: Tasriva, Sikandar, Wan Nur Azhani, W. Samsudin, Kamarul Hawari, Ghazali, Izzeldin, I. Mohd, Mohammad Fazle, Rabbi
• Format: Conference or Workshop Item
• Language: English
• Published: IOP Publishing Ltd 2018
• Subjects: TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://umpir.ump.edu.my/id/eprint/21451/1/Sunglass%20detection%20method%20for%20automation%20of%20video%20surveillance%20system.pdf; http://umpir.ump.edu.my/id/eprint/21451/; http://iopscience.iop.org/article/10.1088/1757-899X/342/1/012040/pdf

Wearing sunglass to hide face from surveillance camera is a common activity in criminal incidences. Therefore, sunglass detection from surveillance video has become a demanding issue in automation of security systems. In this paper we propose an image processing method to detect sunglass from surveillance images. Specifically, a unique feature using facial height and width has been employed to identify the covered region of the face. The presence of covered area by sunglass is evaluated using facial height-width ratio. Threshold value of covered area percentage is used to classify the glass wearing face. Two different types of glasses have been considered i.e. eye glass and sunglass. The results of this study demonstrate that the proposed method is able to detect sunglasses in two different illumination conditions such as, room illumination as well as in the presence of sunlight. In addition, due to the multi-level checking in facial region, this method has 100% accuracy of detecting sunglass. However, in an exceptional case where fabric surrounding the face has similar color as skin, the correct detection rate was found 93.33% for eye glass.