Indexing of acoustic features for ecological audio retrieval

• Grant number: 20/09397-4
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: January 01, 2021
• End date: December 31, 2021
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Maria Cristina Ferreira de Oliveira
• Grantee: Leonardo Rossi Luiz
• Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil

Abstract

The analysis of Acoustic Landscapes has aroused great interest in the scientific community as a resource to assist decision making related to monitoring and understanding environmental issues in diverse environments and landscapes. Analysis relies on capturing the environmental sounds, producing audio recordings that may be inspected to identify relevant acoustic events. This strategy reduces costs, increases the analysis capabilities and minimizes the impact of external influence on the ecological environments. However, it also generates an ever increasing volume of recordings, which makes it difficult to perform analysis and discovery tasks, and poses many practical challenges for the identification of relevant information. Techniques for extracting audio features and indexing acoustic data can be applied to summarize and facilitate access to the data, enabling searches for acoustic patterns of interest. In this project we propose to define and implement a framework to support the retrieval of acoustic events of interest in a set of recordings. This requires extracting the relevant audio features and generating an index to support retrieval.