Generation of a blood vessel image catalog for training machine-learning algorithms

• Grant number: 21/13214-5
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Effective date (Start): March 01, 2022
• Effective date (End): February 28, 2023
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Cesar Henrique Comin
• Grantee: Natália de Carvalho Santos
• Host Institution: Centro de Ciências Exatas e de Tecnologia (CCET). Universidade Federal de São Carlos (UFSCAR). São Carlos , SP, Brazil

Abstract

Changes and ruptures of blood vessels are associated with several diseases. In addition, the characterization of blood vessels is of great importance for understanding the growth and vascularization activity of cell tissues and neurons. Therefore, there is currently a high demand for computational methods that are able to accurately identify and characterize blood vessels in digital images. Although several computational methods have been defined in the literature, the vast majority of the methods were developed from limited databases, having only a few dozen samples. As a result, the methods tend to be specific to the few images used. In this project, we propose to build a comprehensive blood vessel imaging database, having hundreds of high-resolution images. Additionally, we plan to define an original approach for organizing this base. The images will be ordered according to relevant properties, such as contrast and noise level, and also according to characteristics related to the shape of the vessels, such as length and diameter. This organization will enable the systematic verification of the performance of machine learning algorithms in different experimental situations. The generated database will be useful for researchers to develop new algorithms for the study of blood vessels.(AU)