Design of segmentation algorithms to recognize interested cells in microscopy biological images

Fish fecundity is one of the most relevant parameters for estimating reproductive potential of fish stocks used for assessing stock status to guarantee a sustainable fisheries management. Fecundity is the number of matured eggs that each female fish can spawn each year. The stereological method is the most accurate technique to estimate fecundity using histological images of fish ovaries, in which matured oocytes must be measured and counted. This thesis propose a multi-scale Canny filter (MSCF) algorihm to recognize the outlines of cells. It also develop the graphical software STERapp, which includes the MSCF algorithm and other machine learning technique to help the quantitative analysis of images in the fishering labs. STERapp saves between 40% to 70% of time in the fecundity estimation.

keywords: image segmentation, fecundity estimation, machine learning, Support Vector Machine, Microscopic image, Fish gonad