Title: Deep learning based tobacco products classification

Authors: Murat Taşkıran; Sibel Çimen Yetiş

Addresses: Department of Electronics and Communication Engineering, Yildiz Technical University, Istanbul, 34220, Turkey ' Department of Electronics and Communication Engineering, Yildiz Technical University, Istanbul, 34220, Turkey

Abstract: Various images and videos are uploaded every day on Instagram. Shared images include tobacco products and can be encouraging for young people when they are accessible. In this study, it is aimed to classify tobacco products with various convolutional neural networks (CNNs) and to limit the access of young users to these classified tobacco products over the internet. 2008 public images were collected from Instagram, and feature vectors were extracted with various CNNs and CNN was determined to be proper for classification tobacco products. The classification of 5 different tobacco products was realized by using the networks and the classification performance rate was obtained as 99.50% for 402 test images via MobileNet, which gave the highest results 99.11% as average. In this way, the content including tobacco products, can be filtered with a high accuracy rate and a secure Internet environment can be provided for young people.

Keywords: tobacco products; CNN; convolutional neural network; classification; social Media; health; Instagram.

DOI: 10.1504/IJCSM.2021.114193

International Journal of Computing Science and Mathematics, 2021 Vol.13 No.2, pp.167 - 176

Received: 02 Dec 2019
Accepted: 10 Jun 2020
Published online: 13 Apr 2021 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article