Machine learning for interpretation and analysis of microscopic images

• Grant number: 20/12017-9
• Support type: Scholarships in Brazil - Post-Doctorate
• Effective date (Start): October 01, 2020
• Effective date (End): September 30, 2023
• Field of knowledge: Interdisciplinary Subjects
• Principal researcher: Helder Takashi Imoto Nakaya
• Grantee: Mauro César Cafundó de Morais
• Home Institution: Instituto Israelita de Ensino e Pesquisa Albert Einstein (IIEPAE). Sociedade Beneficente Israelita Brasileira Albert Einstein (SBIBAE). São Paulo , SP, Brazil
• Associated research grant: 18/14933-2 - Integrative biology applied to human health, AP.JP2

Abstract

The present work plan aims to extract phenotypic features of the parasite species Trypanosoma, Plasmodium and Leishmania in images derived from microscopy. Such features consist of curvature attributes (minimum, maximum, average, standard deviation, variance, entropy and potential energy), geometric (area, circularity, proportion, perimeter, maximum and minimum distances); and texture statistics (contrast, momentum, second angular momentum, homogeneity, entropy and correlation). These attributes will then be selected through Principal Component Analysis (PCA). Subsequently, the attributes best classified in this way will be used in a geometric approach classification algorithm - Support Vector Machine (SVM). Such a method has the advantage of being computationally cheaper, as it does not require a Graphics Processing Unit (GPU). In this way, we can establish a new, faster and more accurate diagnostic approach to diseases caused by these parasites. (AU)