In Vitro Fertilization (IVF) is the most common fertility treatment. In an IVF treatment, embryologists inspect embryos for subjective quality assessment on a daily basis to select one for implantation. There are several biological studies that confirm the correlations between morphological properties of an embryo’s internal structure and its potential in leading to successful implantation. Automated assessment of embryo’s quality enables a more in-depth understanding of such characteristics and their impact on a positive pregnancy outcome. Moreover, automated quality assessment eliminates subjectivity by selecting embryos with the highest implantation potentials. Automatic monitoring and objective quality assessment of human embryo can potentially improve the outcome of IVF process. This can be achieved through unbiased computer-based approaches that can automatically identify various components of an embryo at different growth stages and quantify their characteristics. This dissertation aims to design and develop tools and methodologies for automatic analysis of temporal and morphological aspects of the human embryo’s in vitro development process. Some of these features and components include the number, centroid locations, boundaries of blastomeres, and segmenting regions corresponding to Zona Pellucida, Trophectoderm and Inner Cell Mass. This dissertation takes a crucial step toward achieving automatic embryo quality assessment. The major contributions of this dissertation include the proposal of a novel cell counting and localization method for blastomeres, the development of the first semantic segmentation for blastocyst components outperforming state-of-the-art methods, and the design of the first system to predict implantation outcome from a single blastocyst image outperforming expert embryologists. Experiments are carried out using various criteria to verify the performance of the proposed methods. Furthermore, the methods developed in this Ph.D. dissertation can be utilized to validate various theoretical assumptions about the relationship between morphological and temporal features of the main components of an embryo and the implantation outcome.
Copyright is held by the author.
This thesis may be printed or downloaded for non-commercial research and scholarly purposes.
Supervisor or Senior Supervisor
Thesis advisor: Saeedi, Parvaneh
Member of collection