Year: 2024
Type: article
Abstract: In response to the increasing demand for enhanced industrialsafety and efficiency, this research delves into the domain of sound anomalydetection within industrial environments, leveraging the power of deep learning.Focused on addressing the limitations of traditional methods, the studyinvestigates various deep learning architectures, including convolutional neuralnetworks (CNNs), recurrent neural networks (RNNs), and hybrid models, todiscern their efficacy in detecting abnormal sounds. The survey rigorouslyevaluates datasets, preprocessing techniques, and benchmarks, providing acomprehensive overview of the state-of-the-art models and their applicationsacross diverse industrial sectors.The paper scrutinizes performance evaluation metrics, drawingcomparisons between deep learning and traditional methods in sound anomalydetection. Real-world applications and case studies underscore the practicalsignificance of these advancements. While acknowledging achievements, theresearch identifies challenges and proposes future directions, emphasizing theneed for innovative solutions to enhance the robustness and real-worldapplicability of deep learning-based sound anomaly detection in industrialsettings.This research not only contributes valuable insights into the intersectionof deep learning and industrial sound analysis but also serves as a pivotal guidefor researchers and practitioners seeking to navigate the complexities ofdeploying effective sound anomaly detection systems.
Author: Magzhan Zhailau
Language: English
Topic: Music and Audio Processing
Subfield: Signal Processing
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4407224410