Image feature extraction via local binary patterns for marbling score classification in beef cattle using tree-based algorithms

Página do item simplificado
dc.contributor.authorPinto, Diógenes Lodi
dc.contributor.authorSelli, Alana
dc.contributor.authorTulpan, Dan
dc.contributor.authorAndrietta, Lucas Tassoni
dc.contributor.authorGarbossa, Pollyana Leite Matioli
dc.contributor.authorVoort, Gordon Vander
dc.contributor.authorMunro, Jasper
dc.contributor.authorMcMorris, Mike
dc.contributor.authorAlves, Anderson Antonio Carvalho
dc.contributor.authorCarvalheiro, Roberto [UNESP]
dc.contributor.authorPoleti, Mirele Daiana
dc.contributor.authorBalieiro, Júlio Cesar de Carvalho
dc.contributor.authorVentura, Ricardo Vieira
dc.contributor.institutionUniversidade de São Paulo (USP)
dc.contributor.institutionUniversity of Guelph
dc.contributor.institutionAgSights
dc.contributor.institutionLivestock Research Innovation Corporation
dc.contributor.institutionUniversity of Wisconsin-Madison
dc.contributor.institutionUniversidade Estadual Paulista (UNESP)
dc.date.accessioned2023-07-29T14:00:23Z
dc.date.available2023-07-29T14:00:23Z
dc.date.issued2023-01-01
dc.description.abstractThe objective of this study was to investigate the potential of creating a pipeline to classify the marbling score obtained from ribeye area (REA) images using computer vision and machine learning methods. Our database consisted of images and measurements (N = 2,446) from the transversal cut between the 12th and 13th ribs of the Longissimus dorsi muscle from carcasses of a beef cattle population (Bos taurus). Each sample was previously labeled by the industry using a low, medium or high marbling score. The prediction accuracies of two tree-based Machine Learning (ML) algorithms (Decision Tree - DT and Random Forest - RF) were compared, as well as different measures for extracting features from the REA images, which were used as input for the ML algorithms. In order to extract features based on detectable color patterns and textures contained in smaller parts of the grayscale image, we proposed the application of the local binary pattern (LBP) method prior to the adoption of ML methods. Mean classification accuracies for the test set ranged from 45.78% to 91.25% for different test scenarios. The results were mostly impacted by the feature extraction metrics, ML methods, potential subjectivity during the classification process by the industry, and the number of classes evaluated together. The best prediction accuracy results were achieved after performing the cross-validation (20% in each balanced group, 5 folds, and 10 repetitions), considering solely the extreme groups (low and high marbling scores) and pre-selecting from each group the 400 most visually representative samples. The RF algorithm outperformed the DT for most scenarios. After increasing the number of images to 580 samples for the same two groups, the highest testing accuracies were reduced to 83.05% for RF and 75.58% for DT. Such a decrease in the classification accuracies may be associated with the addition of erroneously classified images, due to the subjective nature of the industry evaluation. In conclusion, our preliminary studies showed the LBP method as a powerful feature extraction strategy considering a scenario where the labels were well defined. Our results revealed high accuracies for the classification of marbling extremes, but there is an evident need to improve the understanding of the biological and visual aspects that led to a sharp drop in classification accuracy after the insertion of the intermediate groups of marbling. In addition, the authors highlight the importance of an accurate labeling process for achieving better classification accuracy when applying supervised classification methods.en
dc.description.affiliationDepartment of Nutrition and Animal Production (VNP) School of Veterinary Medicine and Animal Science University of São Paulo, SP
dc.description.affiliationCentre for Genetic Improvement of Livestock (CGIL) Departament of Animal Biosciences University of Guelph
dc.description.affiliationAgSights
dc.description.affiliationLivestock Research Innovation Corporation
dc.description.affiliationDepartment of Animal & Dairy Sciences University of Wisconsin-Madison
dc.description.affiliationAnimal Science Department School of Agricultural and Veterinarian Sciences São Paulo State University, SP
dc.description.affiliationDepartment of Basic Science The Faculty of Animal Science and Food Engineering University of São Paulo, SP
dc.description.affiliationUnespAnimal Science Department School of Agricultural and Veterinarian Sciences São Paulo State University, SP
dc.identifierhttp://dx.doi.org/10.1016/j.livsci.2022.105152
dc.identifier.citationLivestock Science, v. 267.
dc.identifier.doi10.1016/j.livsci.2022.105152
dc.identifier.issn1871-1413
dc.identifier.scopus2-s2.0-85145974639
dc.identifier.urihttp://hdl.handle.net/11449/249028
dc.language.isoeng
dc.relation.ispartofLivestock Science
dc.sourceScopus
dc.subjectImage features
dc.subjectLocal binary pattern
dc.subjectMachine learning
dc.subjectMarbling score
dc.subjectRibeye area
dc.titleImage feature extraction via local binary patterns for marbling score classification in beef cattle using tree-based algorithmsen
dc.typeArtigo
unesp.author.orcid0000-0001-5218-1593[1]
unesp.author.orcid0000-0002-2242-6315[2]
unesp.author.orcid0000-0003-1100-646X[3]
unesp.author.orcid0000-0002-3504-165X[4]
unesp.author.orcid0000-0002-5151-7005[5]
unesp.author.orcid0000-0001-5604-9207 0000-0001-5604-9207[6]
unesp.author.orcid0000-0003-0540-1850[12]
unesp.author.orcid0000-0002-8758-3020[13]

Arquivos

Coleções

Artigos - Produção Animal - FMVZ