Segmentação de imagens de enxertos ósseos utilizando watershed e k-médias

Detalhes bibliográficos
Ano de defesa: 2011
Autor(a) principal: Almeida, Thallys Pereira de
Orientador(a): Não Informado pela instituição
Banca de defesa: Não Informado pela instituição
Tipo de documento: Dissertação
Tipo de acesso: Acesso aberto
Idioma: por
Instituição de defesa: Universidade Federal da Paraí­ba
BR
Informática
Programa de Pós Graduação em Informática
UFPB
Programa de Pós-Graduação: Não Informado pela instituição
Departamento: Não Informado pela instituição
País: Não Informado pela instituição
Palavras-chave em Português:
Link de acesso: https://repositorio.ufpb.br/jspui/handle/tede/6102
Resumo: The applicability of bone grafts in the recovery of bone tissue is an area that has been extensively studied and is constantly evolving. In the last years, several techniques for applying these grafts have been developed. Thus, there is a need of creating an automated method of analysis, objective, impartial and efficient evaluation of the results of these techniques. Several studies using image segmentation have been conducted in several areas. Among them we can cite the medical field, where the segmentation can be applied to define the image regions such as tumors, cells, glands, organs, tissue, cells, among others. Several segmentation methods were developed specifically to detect each one of these elements. Segmentations like these were traditionally performed manually. In some cases, the objects in an image were detected of a purely visual method, and no numerical data could accurately determinate the size or other properties of the picture. Evaluations like this become subjective to the interpretation of each researcher, and could bring impartial and unbiased results. Thus, the use of a computational method for carrying out this type of analysis is essential. This work presents a morphometric analysis method of the bone formation on bone grafts applied to mice that received implants of biomaterials. The technique employs a combination of segmentation algorithms k-means, and watershed in order to perform a segmentation based on the color represented by the L*a*b*. For identification of the region that corresponds to new bone formation, an analysis was performed using the HSL color system. Several experiments were performed with a large number of images and satisfactory results were obtained. The performance of the segmentation method on images of bones of rodents has been thoroughly evaluated by researches, who considered the process efficient and compatible with results obtained by experts.