Applied Intelligence and Computing
CNN-Mobilized Health: Visual Classification of Smokers using MobileNetV2 CNN Model
Authors: Mohit Beri, Kanwarpartap Singh Gill, Deepak Upadhyay and Swati Devliyal
Publishing Date: 05-11-2024
ISBN: 978-81-955020-9-7
Abstract
This research investigates the application of Convolutional Neural Networks (CNN) with the MobileNetV2 architecture for the critical task of smoker classification based on visual data. Tobacco use remains a pervasive global health concern, with profound physical and mental health implications. The study employs a diverse dataset encompassing various tobacco consumption forms, addressing the need for an automated system to identify smoking individuals. The motivation lies in developing a robust tool for early detection, facilitating timely public health interventions. The research focuses on the immediate health benefits of quitting tobacco and underscores the urgency of encouraging smoking cessation. The methodology involves meticulous data preprocessing, utilizing a three-tiered dataset division, and leveraging MobileNetV2 for model training. The evaluation includes accuracy metrics, visualizations of accuracy and loss curves, and detailed analyses using classification reports and confusion matrices. The promising results highlight the potential of CNNs in automating smoking detection, contributing to public health campaigns and interventions. Further exploration involves refining the model with larger datasets and considering real-world deployment scenarios.
Keywords
Tobacco Classification, Convolutional Neural Networks, MobileNetV2, Smoker Detection, Deep Learning, Health Intervention, Public Health, Image-based Classification.
Cite as
Mohit Beri, Kanwarpartap Singh Gill, Deepak Upadhyay and Swati Devliyal, "CNN-Mobilized Health: Visual Classification of Smokers using MobileNetV2 CNN Model", In: Mukesh Saraswat and Rajani Kumari (eds), Applied Intelligence and Computing, SCRS, India, 2024, pp. 41-48. https://doi.org/10.56155/978-81-955020-9-7-5