Eye-Smoker: A Machine Vision-Based Nose Inference System of Cigarette Smoking Detection using Convolutional Neural Network
2020 IEEE 7th International Conference on Engineering Technologies and Applied Sciences (ICETAS)
(2020)
Jonel R. Macalisang
a
,
Nino E. Merencilla
b
,
Michael Angelo D. Ligayo
c
,
Mark P. Melegrito
d
,
Ryan R. Tejada
e
a Technology Licensing Office - ITSO, Technological University of the Philippines, Manila, Philippines
b Department of Computer Engineering, FEU Institute of Technology, Manila, Philippines
c Department of Electronics Engineering, Quezon City University, Quezon City, Philippines
d Department of Electronics and Communications Engineering, Technological University of the Philippines, Manila, Philippines
e College of Computing Sciences, Ifugao State University, Ifugao, Philippines
Abstract: In the Philippines, at least 16 million Filipinos reported smoking cigarettes amid the campaign against tobacco products due to various concerns about their adverse health effects. Due to health, environmental, and safety concerns, the President of the Philippines issued Executive Order 26 s. 2017, imposing a nationwide ban on smoking (use of tobacco including e-cigarettes) in all public places in the Philippines. Despite the implementation of this order, many are still seen smoking in prohibited smoking areas. A smoke detector can be helpful in this situation. This study proposed a smoker detection system that uses a deep learning algorithm that can detect people that are smoking cigarettes. The study used the Pascal VOC format and LabelImg tool for annotating the datasets. Training, validation, and evaluation of the system is done by presenting images, videos, and live detection using the webcam of a camera. Overall, the system produced 90% testing accuracy.