CHIBUKO, SAMUEL (2026) “DEVELOPMENT OF AN AI-BASED CROWD DENSITY ESTIMATION SYSTEM”. Other thesis, Godfrey Okoye University, Enugu Nigeria.
![[img]](https://eprints.gouni.edu.ng/style/images/fileicons/text.png) |
Text
CHIBUKO SAMUIEL.pdf Download (5MB) |
Abstract
This project work of research focused mainly on developing a very simple AI-based crowd density estimation system that is very effective and efficient. During the process of the research, the researcher employed the use of the ShanghaiTech dataset as the underlying data source for training the AI model. The dataset was made up of diverse crowd images which were group into dense and sparse groups. The dataset was therefore cleaned and enhanced in other to optimize the training of the machine learning model. Using the cleaned dataset, a deep learning model was trained using the Convolutional Neural Networks algorithm along side Adam and Stochastic Gradient Descent optimizers in other to generate density maps under conditions that have heavy occlusion. However, at the end of the development phase, the system was tested using different evaluation metrices, where the system performed effectively, achieving a Mean Absolute Error (MAE) of 12.4 and a Root Mean Squared Error (RMSE) of 18.2, while being able to successfully classify crowd density into low, medium, high , and critical levels. Moreover, at the end of the research, all the system requirements were fully met successfully, thereby achieving the project’s core objectives for public safety monitoring.
| Item Type: | Thesis (Other) |
|---|---|
| Subjects: | Q Science > Q Science (General) |
| Divisions: | Faculty of Computing And Information Technology (FACIT) |
| Depositing User: | Cynthia Ugwuoti |
| Date Deposited: | 04 Aug 2026 14:32 |
| Last Modified: | 04 Aug 2026 14:32 |
| URI: | http://eprints.gouni.edu.ng/id/eprint/6182 |
Actions (login required)
 |
View Item |
