NORMA eResearch @NCI Library

Trash Image Classification System using Machine Learning and Deep Learning Algorithms

Gupta, Himanshu (2020) Trash Image Classification System using Machine Learning and Deep Learning Algorithms. Masters thesis, Dublin, National College of Ireland.

[thumbnail of Master of Science]
PDF (Master of Science)
Download (777kB) | Preview
[thumbnail of Configuration manual]
PDF (Configuration manual)
Download (1MB) | Preview


In today’s fast pacing world of the internet age with all the amenities and latest gadgets, the major urban cities in the world are still struggling with trash management. Only a few countries using the recycling of wastes but most of them dumping all the trash to the landfills. The quantity of generated trash in day to day life is affecting land, water, and air which causes a serious threat to the aquatic species and their surroundings and ultimately to humans if not managed properly. The objective of this study is to develop a system that can classify these trash images into their correct categories with the help of machine learning and deep learning methodologies. The dataset used for achieving this objective is released by the TACO dataset consist of 1500 images of litter and trash with annotations and labeled. There are five categories of trash with the name ’Plastic straw’, ’Drink can ’, 'Cigarette’, and ’Clear plastic bottle’ considered for this work. This dataset is recently released and very rarely used in any research work. To make the dataset more scalable and balanced various data augmentation techniques will be adopted. Four classification algorithms such as the Sequential Keras model,transfer learning ResNet-50, and VGG-19 models and XGBoost classifier model will be developed for features extraction and classification of images. The performance of the model will be evaluated based on accuracy and comprehensive comparison will occur among all the four models on several parameters. The outcomes showed that pre-trained transfer learning models can be used for high classification accuracy and assured that data augmentation techniques assists in improving the overall results.
Keywords: Data Augmentation, Trash Classification, Image Classification, ResNet-50, VGG-19, Sequential Keras Model, Extreme Gradient Boosting, Transfer Learning.

Item Type: Thesis (Masters)
Subjects: Q Science > QA Mathematics > Electronic computers. Computer science
T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science
Q Science > QA Mathematics > Computer software
T Technology > T Technology (General) > Information Technology > Computer software
Divisions: School of Computing > Master of Science in Data Analytics
Depositing User: Dan English
Date Deposited: 20 Jan 2021 14:25
Last Modified: 20 Jan 2021 14:25

Actions (login required)

View Item View Item