Tese de Doutorado de Lizeth Andrea Castellanos Beltrán

Aluna: Lizeth Andrea Castellanos Beltrán
Orientadora: Profª. Drª. Carla Maria Dal Sasso Freitas

Título: Characterization of Structures in Confocal Images Datasets Obtained from Bile Ducts
Linha de Pesquisa: Computação Gráfica e Visualização de Dados
 
Data: 14/07/2020
Horário: 13h30

Esta banca ocorrerá, excepcionalmente, de forma totalmente remota.
Acesso público disponibilizado através do link: https://mconf.ufrgs.br/webconf/inf-talks

Banca Examinadora:

– Profª. Drª. Olga Regina Pereira Bellon (UFPR – por videoconferência)
– Profª. Drª. Isabel Harb Manssour (PUCRS – por videoconferência)
– Prof. Dr. Eduardo Simões Lopes Gastal (UFRGS – por videoconferência)

Presidente da Banca: Profª. Drª. Carla Maria Dal Sasso Freitas

Abstract: Confocal microscopy is a useful tool for acquiring 3D datasets of fluorescent specimens. Some applications of these datasets include the study of cell organelles and tissue changes. In hepatology, researchers have been using confocal microscopy for investigating the microanatomy of bile ducts. Bile ducts are complex tubular structures consisting of many juxtaposed microstructures with distinct characteristics. The analysis of microscopic morphological changes in the bile ducts is important in the study of biliary diseases. Since confocal images are difficult to segment because of the noise introduced during the specimen preparation, traditional quantitative analyses used in other medical datasets are difficult to perform on confocal microscopy data and require extensive user intervention. Thus, the visual exploration and analysis of bile ducts pose challenges for hepatology research, requiring different methods. In this work, we provide methods for characterizing structures in confocal images datasets obtained from bile ducts. In our motivating case study, the characterization of such structures is likely to help hepatologists to distinguish specimens affected by biliary atresia, a disease that requires a liver transplant to avoid premature death. Our data consists of 3D image datasets containing several slices of mouse bile ducts organized as two fluorescence channels.  The red channel contains a network of small vessels named Peribiliary Vascular Plexus (PVP), and the green channel contains the internal bile duct with Peribiliary Glands (PBGs). Our approach for characterizing the bile ducts structures includes image processing techniques for enhancing the volumetric data, a clustering method as a tool for segmentation, and fractal dimension analysis for supporting characterization. We use anisotropic diffusion to enhance the confocal images. For clustering, we have explored the density-based spatial clustering for applications with noise (DBSCAN) algorithm, using gradient information for guiding the clustering. We obtained a better visualization of the most prominent vessels and internal structures. We also explore the fractal properties of the datasets based on the computation of fractal dimension, which we found useful for extracting quantitative information aiming at characterizing relevant structures.

Keywords: confocal microscopy, clustering, DBSCAN, fractal dimension.