Scholarworks @ Morehead State

Title

Visual Pollution Classification using Convolutional Neural Networks

Authors

Jacob Vogelpohl
Heba Elgazzar, Faculty Mentor

Files

Description

Visual pollution is an impairment on an individual's ability to enjoy their surroundings. It usually takes the form of a messy and chaotic environment that can cause overstimulation of the visual senses. This includes trash, advertisements, construction, electric cables, and similar objects. Convolutional Neural Networks (CNNs) are a form of artificial intelligence that use supervised learning to process and classify images. In this research, a CNN processed images of city streets and classified them as polluted or not polluted based on the visual characteristics that it learned from during its training period. The CNN achieved a training accuracy of 98% and a validation accuracy of 80%.

Publication Date

4-2023

Disciplines

Arts and Humanities | Business | Education | Engineering | Higher Education | Life Sciences | Medicine and Health Sciences | Physical Sciences and Mathematics | Social and Behavioral Sciences

Recommended Citation

Vogelpohl, Jacob and Elgazzar, Heba, "Visual Pollution Classification using Convolutional Neural Networks" (2023). 2023 Celebration of Student Scholarship - Poster Presentations. 21.
https://scholarworks.moreheadstate.edu/celebration_posters_2023/21