We are publishing a research paper titled “Comparative Analysis of CNN and TNN Models in Environmental Noise Sources,” which our team presented at the Internoise conference in Nantes.
This paper presents a focused examination of the challenges associated with environmental noise, particularly in urban settings where the amalgamation of various sound sources complicates the task of noise classification.
Traditional methodologies, while effective in capturing level exceedance, fall short in addressing the nuanced requirements of noise source identification.
With the emergence of IoT Noise Monitoring Terminals we have seen a pivotal shift from conventional manual auditory analysis methods to more sophisticated, data-driven approaches.
The objective of this paper aims to compare the efficacy of TNN’s, specifically the adaptation of the AST model, against the previously favored CNN architecture.
The UrbanSound8K dataset was used for both the CNN and the TNN.
Read the full research paper at the link below.
