High-value Target Detection
Name
Anton Prokopov
Abstract
This work describes an automatic high-value target detection for the purpose of reducing the human workload in analyzing video feed from the given source. The aim of this thesis is to investigate the mechanism of neural networks in general and to fine-tune an existing pre-trained neural network with suitable data. Further adjustment of the parameters was required to achieve better results in performing robust target detection in real-time. This kind of a system can recognize different classes of targets depending on the data it was trained with. The research was focused on detecting particular car marks and models, but target could be defined as any object. The potential application of such system could be found in surveillance systems, border control, monitoring the animals.
Graduation Thesis language
English
Graduation Thesis type
Master - Computer Science
Supervisor(s)
Gholamreza Anbarjafari
Defence year
2018