Carleton University - School of Computer Science Honours Project
Fall 2020
Anomaly Detection through Video Analysis
Tyson Steele
SCS Honours Project Image
ABSTRACT
This project provides an automated tool capable of detecting anomalies and outliers given a video file without relying on machine learning or neural networks. This was accomplished by using a combination of existing tools and technologies such as colour detection, motion detection, facial recognition, and object detection. Once the frames are processed and categorized using those methods, the program is able to identify the frames in which anomalies exist. Further manual analysis is then possible using the JSON output of the program.