Tomografia computadorizada aplicada a otólitos em 3D

Abstract

Otoliths are called "black box" because they are sources of biological information that can be applied to the evaluation and management of fish stocks. And, one of the ways to study them is from analysis of images. However, 2D images have already been extensively explored, despite losing information because otoliths are a 3D object. Thus, the present study aimed to apply computed tomography to analyze otoliths sagittae. From 15 configurations of the tomographic parameters, it was concluded that average voltages (70 and 80 kV) with aluminum filters with 0.5 mm thickness are optimal for the acquisition of tomographic images of otoliths. These images were then explored internally and externally, with applications in growth ring counts, ontogeny, species comparisons, and, in the future, layered segmentation using the density difference. In addition, internal cavities were found in individuals of the genus Thunnus. The curvature of the surface of the otoliths was also analyzed in order to test the capacity of discrimination between species, ontogenetic variation and as a tool of segmentation of sulcus acusticus. It was found that the Mean and RMS curvature attributes can be used as a unique signature by species, besides discerning peaks and valleys in the relief of otoliths. And, from this, relate its complexity to the metabolism of individuals, as a reflection of age. In addition, with 3D segmentation, it will be possible to describe not only the area of the sulcus, but also its depth and volume. Therefore, this thesis will serve as a basis for future studies of 3D image analysis of otoliths and some applications.