OGHENEVOVWERO Z I O N APENE
DELSU ABRAKA, COMPUTER SCIENCE, Graduate Student
Advancements in computer vision and deep learning have led to significant progress in automated crime detection systems. This study focuses on the development of a novel mathematical model for crime detection based on the You Only Look... more
Advancements in computer vision and deep learning have led to significant progress in automated crime detection systems. This study focuses on the development of a novel mathematical model for crime detection based on the You Only Look Once (YOLOv5) network architecture. The proposed model utilizes state-of-the-art object detection techniques to identify, classify, and detect criminal activities in surveillance footage, including images and videos, focusing on critical crime categories such as weapons and violent behaviour. The model's performance is evaluated on seven classes of weapon objects and violent scenes, achieving a precision (P) of 0.842, recall (R) of 0.77, and mAP of 0.811. These results demonstrate the model's efficiency in accurately identifying and categorizing criminal activities, thereby contributing to enhancing public safety and security through the utilization of cutting-edge deep learning technologies in crime prevention and detection.
Research Interests:
Crime prevention and detection are critical components of public safety in any nation. Traditionally, crime prevention and detection approaches relied on human intuition and limited data, resulting in reactive and resource-intensive... more
Crime prevention and detection are critical components of public safety in any nation. Traditionally, crime prevention and detection approaches relied on human intuition and limited data, resulting in reactive and resource-intensive methods. However, recent advancements in artificial intelligence (AI) offer a paradigm shift, enabling proactive, data-driven approaches. This study explores the evolution from conventional crime prevention and detection methods to cutting-edge AI solutions. It employs a literature survey, local observation, and global news approach to examine the current state of the art in AI-driven approaches. Traditional crime prevention methods, such as neighbourhood watch programs, random stop-and-search initiatives, and foot patrols, are examined alongside technological approaches, such as surveillance systems, crime mapping, and geographical profiling. These conventional techniques are tedious and time-consuming leading to inefficiency. Findings from the study revealed that AI has the potential to revolutionize crime prevention and detection through its subfields, such as machine learning and computer vision. Machine learning algorithms can process large amounts of data to forecast potential criminal activity, thus transforming law enforcement operations. Also, computer vision models can utilise visual data from surveillance cameras and other sources to analyse, identify, and respond to crimes. The study recommends the integration of AI into law enforcement agencies for crime prevention and detection to transform societal security. In addition, it emphasizes the need for further research in this domain. The study also recommends the development of an efficient framework and model for crime detection based on deep learning to enhance public safety.