Classifying plastic waste using deep convolutional neural networks for efficient plastic waste management

• Main Authors: Ahmad Radzi, Nur Iwana, Ismail, Nurul Fadhillah, Olowolayemo, Akeem
• Format: Article
• Language: English
• Published: IIUM Press 2022
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://irep.iium.edu.my/95719/1/95719_Classifying%20plastic%20waste%20using%20deep%20convolutional%20neural.pdf; http://irep.iium.edu.my/95719/; https://journals.iium.edu.my/kict/index.php/IJPCC/article/view/282

Plastic waste recycling has not been adopted by a large percentage of plastic manufacturing companies due to the enormous amount of effort required to sort the plastic waste and remove dirt. Consequently, the lack of efficient practice of automated sorting and separation of different types of plastic during the management of plastic waste has caused most of it to end up in landfills instead of being reused and recycled back into society’s consumption. Accumulation of plastic waste eventually causes pollution which will then result in negative effects on ecosystems, underwater and on the ground as well as carbon emission. To leverage machine learning technology in optimizing the process of recycling plastic waste, this study proposes an intelligent plastic classification model developed using a Convolutional Neural Network (CNN) with 50-layer residual net pre-train (ResNet-50) architecture. The proposed model was trained with a dataset consisting of over 2000 images that were compiled and organized into seven plastic categories. The model compared favourably with related previous studies producing a considerably high accuracy classification model of 94.1%.