Orientador: Prof. Dr. Marcelo Zanchetta do Nascimento / Dissertação (mestrado) - Universidade Federal do ABC, Programa de Pós-Graduação em Engenharia da Informação, 2014. / In vitro production has been employed in bovine embryos and quantification of lipids
is fundamental to understand the metabolism of these embryos. This paper presents an
unsupervised segmentation method for histological images of bovine embryos. During the
process of mounting slides some droplets of dye and other fragments could be remain
out of the embryo. With this method, the irregularities were removed by applying an
algorithm developed in the pre-processing step, in which a calculation is done to eliminate
regions whose pixels have values of hypotenuse greater than 22% of the image center.
After removal of the irregularities a transformation of the RGB color model to the YIQ
was applied. Then the global thresholding technique was applied with the threshold value
obtained by the maximum entropy in order to separate the lipid droplets of histological
slides at diferent stages: early cleavage, morula and blastocyst. After this step, false
positive regions were removed using the technique of connected components. For this step,
area information, center and amount of pixels were used. It enabled outliers removal
of segmented regions. The proposed segmentation method was applied in 60 histological
images of bovine embryos. The images were evaluated qualitatively and quantitatively with
respect to the gold standard images. The method presented accuracy values of 96 2%,
942% and 952%, respectively in the tissues of early cleavage, morula and blastocyst.
The developed method obtained better results than classical segmentation methods.
| Identifer | oai:union.ndltd.org:IBICT/oai:BDTD:76608 |
| Date | January 2014 |
| Creators | Melo, Douglas Henrique de |
| Contributors | Nascimento, Marcelo Zanchetta do, Marietto, Maria das Graças Bruno, Ramos, Rodrigo Pereira |
| Source Sets | IBICT Brazilian ETDs |
| Language | Portuguese |
| Detected Language | English |
| Type | info:eu-repo/semantics/publishedVersion, info:eu-repo/semantics/masterThesis |
| Format | application/pdf, 78 f. : il. |
| Source | reponame:Repositório Institucional da UFABC, instname:Universidade Federal do ABC, instacron:UFABC |
| Rights | info:eu-repo/semantics/openAccess |
| Relation | http://biblioteca.ufabc.edu.br/index.php?codigo_sophia=76608&midiaext=69777, http://biblioteca.ufabc.edu.br/index.php?codigo_sophia=76608&midiaext=69778, Cover: http://biblioteca.ufabc.edu.brphp/capa.php?obra=76608 |
Page generated in 0.0024 seconds