Application of computer vision techniques in the development of classifiers for feeding and dominance behaviors in Zebu cattle

• Grant number: 24/22540-1
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: April 01, 2025
• End date: March 31, 2026
• Field of knowledge: Agronomical Sciences - Agricultural Engineering - Rural Buildings and Ambience
• Principal Investigator: Rafael Vieira de Sousa
• Grantee: Marcos Jun Orikassa
• Host Institution: Faculdade de Zootecnia e Engenharia de Alimentos (FZEA). Universidade de São Paulo (USP). Pirassununga , SP, Brazil

Abstract

In intensive systems, climate directly influences cattle feeding behavior, affecting water and dry matter intake as well as the timing and frequency of consumption. Dominance hierarchy also plays an important role in confined environments, impacting social interactions, access to resources, and the animals' overall behavior. The Performance and Feed Efficiency Test (TDEA) of the National Young Bulls Evaluation Program (PNAT), promoted by the Brazilian Association of Zebu Breeders (ABCZ), seeks to identify young bulls with relevant traits for zebu cattle production. However, the assessment of performance, behavior, and hierarchy traditionally requires in-person data collection and manual observation, making the process labor-intensive and inefficient. This study, therefore, proposes the development of a computer vision model to classify feeding and dominance behaviors. During the PNAT, video footage from the feeding area will be collected to serve as input for the model. Classification will be performed using a Long Short-Term Memory (LSTM) Recurrent Neural Network, designed to capture temporal patterns such as animal movements and interactions. The scientific initiation project is linked to the CNPq project "Non-invasive automatic systems for integrated monitoring of performance and welfare in beef cattle farming," which investigates the hypothesis that animal performance is related to feeding behavior, response to thermal environment, and dominance hierarchy. Based on these correlations, it is believed that it will be possible to develop automatic systems capable of monitoring animal performance and welfare in a non-invasive and real-time manner.