Automated detection of nut-cracking in camera trap audio using convolutional neural network

• Grant number: 25/01069-1
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: April 01, 2025
• End date: March 31, 2026
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Rafael Luiz Testa
• Grantee: Pedro Schneider
• Host Institution: Centro Universitário FEI (UNIFEI). Campus de São Bernardo do Campo. São Bernardo do Campo , SP, Brazil

Abstract

This research project proposes the development of a deep learning system for automated analysis of acoustic data from camera traps, focusing on the detection and classification of nut-cracking behavior by Capuchin monkeys (Sapajus spp.). Although camera traps are widely used in wildlife monitoring, analyses have traditionally focused only on visual data, underutilizing the potential of simultaneously generated sound recordings. The complexity and volume of these data requires a laborious manual analysis, necessitating efficient automated solutions.The proposed system will be developed using audiovisual recordings collected in the Chapada dos Veadeiros and Ubajara National Parks, containing coconut-cracking events annotated by experts. The model will be specifically designed to discriminate between coconut-cracking events of babaçu and macaúba, important species in the diet of capuchin monkeys, using a convolutional neural network architecture to classify acoustic patterns.The methodology is based on three components: development of techniques for detecting relevant acoustic events, implementation of methods for mitigating environmental noise, and design of a neural architecture for classifying acoustic patterns. The final product will be a complete computational framework that will allow not only the efficient analysis of large volumes of data, but will also contribute to the understanding of foraging patterns and tool use by primates, representing a significant contribution to the fields of bioacoustic signal processing and animal behavior studies.