Avaliação da capacidade de retenção de água em amostras de carne suína

• Ano de defesa: 2019
• Autor(a) principal: Andrade, Renan Gomes de
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Uberlândia; Brasil; Programa de Pós-graduação em Ciência da Computação
• Palavras-chave em Português: Capacidade de Retenção de Água (CRA); Computação; Qualidade da carne suína; Processamento de imagens - Técnicas digitais; Processamento Digital de Imagens (PDI); Carne de porco - Qualidade; Segmentação; Segmentação de imagem; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO::METODOLOGIA E TECNICAS DA COMPUTACAO::PROCESSAMENTO GRAFICO (GRAPHICS)
• Link de acesso: https://repositorio.ufu.br/handle/123456789/27552; http://doi.org/10.14393/ufu.di.2019.76

Pork is one of the protein source most consumed in Brazil and in the world. Therefore, it is extremely important that the meat offered to consumers pass for a strict process to control of quality. Usually, pork’s quality is associated with some features desired and valued by consumers. Among the many attributes analyzed to obtain a high quality product, Water Holding Capacity (WHC) is one of the most used. The levels of water concentration in pork are directly related to its visual appearance, softness, juiciness, firmness and yield during preparation. WHC is usually quantified by the compression method, which consists of estimating the amount of water in a sample by measuring the volume of juice released after its compression. However, this process is done manually by experts, what makes this a slow, imprecise, variable and expensive operation, vulnerable to failure and human fatigue. In view of the need to automation this process and the scarcity of research using digital image processing techniques to segment and quantify WHC, we developed an automatic segmentation method capable of quantifying the WHCs of pork samples. The efficiency of the proposed method is evaluated by performing experiments where the segmentation of the samples and the accuracy of the extracted WHCs are analyzed. In both experiments, the proposed method obtained satisfactory results. About 93% of all images were successfully segmented and the values of the WHCs extracted by experts and the proposed method were very close. Therefore, this work may help professionals in the area of food control to perform a more efficient, practical and cheaper WHC quantification.