Domestic Waste Segregation using Deep Neural Networks

Yogeashwaran, Iswarya (2022) Domestic Waste Segregation using Deep Neural Networks. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration manual) Download (1MB) | Preview

Abstract

Everyone’s life has been dramatically affected by significant environmental challenges. Environment and living species suffer from a variety of difficulties because of improper waste disposal. People’s negligence, insufficient information about waste types, and bad habits are factors that influence appropriate waste disposal. Generally, rubbish is physically classified. The garbage truck, that separates garbage into biodegradable and non-disposable categories, has innovated this process. The image recognition methodology is used in this system. In this study, the techniques of Custom CNN, VGG16, Mobile-net, and Xception are assessed using analytics to sort garbage into corresponding classes. We may minimize this wrong waste disposal issue at source by developing smart home bins that mishandle fewer items. We expect higher waste segregation efficiency by using smart bins than solutions used at final stages. It transfers far more responsibilities to users and waste collection, and therefore increase the final waste disposal efficiency. A range of techniques, such as VGG16, Custom CNN, Mobile-net, and Xception techniques were tested in this work. In this research, a custom CNN model achieved 90.21% of validation accuracy, 90.27% of validation precision and 90.05% of validation recall. The enhanced model training efficiency allows it to run over 22,564 images, while other models can classify 6443 images in the same allocated time. As Custom CNN has scored really good accuracy even by compiling more data, it is being chosen as Optimum model and implemented into the prototype. This model is saved in a path to integrate with Flask front end to create a web application which has predicted the image into proper category when it has uploaded.

Item Type:	Thesis (Masters)
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science T Technology > TD Environmental technology. Sanitary engineering
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Tamara Malone
Date Deposited:	14 Mar 2023 16:07
Last Modified:	14 Mar 2023 16:07
URI:	https://norma.ncirl.ie/id/eprint/6347

Actions (login required)

View Item