Enhancing Crime Prevention via Human Scream Detection with Deep Learning and Machine Learning

Raheem Basha, Ashraf Hussain (2024) Enhancing Crime Prevention via Human Scream Detection with Deep Learning and Machine Learning. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration Manual) Download (1MB) | Preview

Abstract

Scream recognition is one of the very important components of audio analysis with prospects to be used in cases such as crime prevention, emergency situations, and security. The current study examines the efficacy of machine learning classifiers: Support Vector Machines (SVM), Multilayer Perceptron (MLP), and ResNet-34 deep learning for scream detection. Features including the Mel-Frequency Cepstral Coefficients (MFCC) and spectrogram visualisations were applied when feeding the inputs for classification models.

Advanced data augmentation approaches were used to fix drawbacks such as noise and class skewness so that both clean and noisy data sets were created. Specific performance indicators like precision, recall, F1 score and noise robustness of each model were measures used in the evaluation process. However, when it came to clean well-controlled data, SVM and MLP gave a much better result than ResNet-34, but the latter did have overall higher accuracy, adaptability, and worked better in noisy inputs.

While this paper only provides evidence of the screams’ detection, it demonstrates the usefulness of each model separately to help further evaluate their applicability in practice. The objective of the study is to identify the aspects that are needed in the improvement of audio-based safety systems while recommending them in the application of emergency response and public safety.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Tomer, Vikas UNSPECIFIED
Uncontrolled Keywords:	Scream Detection; Machine Learning; Deep Learning; Resnet-34; Support Vector Machines; Multilayer Perceptron; Data Augmentation; Spectrograms; Public Safety
Subjects:	H Social Sciences > HV Social pathology. Social and public welfare > Criminology > Crimes and Offences Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	04 Sep 2025 10:52
Last Modified:	04 Sep 2025 10:52
URI:	https://norma.ncirl.ie/id/eprint/8779

Actions (login required)

View Item