Pork is the most consumed meat in the world and, in Brazil, out of the 2.4 million existing pigs, 66.67% have been produced in a highly technological way. The main goal of the market is to achieve maximum productivity using the lowest possible financial resources, which directly affects the life quality of the animal, as the stocking rate per pen increases and climatic conditions are not controlled. Due to the growing concern about the quality of food and ethics involved in production by consumers, investors, and importers; food companies began to prioritize welfare measurements, influencing final meat quality. In order to measure animal welfare, presential observations are commonly used, which generate inaccurate results due to the subjectivity and time demanded for analysis, so the management to be employed is affected. This study proposes to collect animals' image and sound data, aiming to develop a non-invasive behavioral measurement methodology (using camera and microphone) by applying machine learning techniques. It is expected to perform behavioral analysis instantly and accurately, generating more accurate decision-making, which can positively impact on productivity results.
News published in Agência FAPESP Newsletter about the scholarship: