Classification of Red Blood Cells in the Kendall Space of Reflection Shapes

The classification of red blood cells (RBCs) or erythrocytes into three categories based on their shape, normal, sickle-shaped, and those with other deformations, has proven to be a crucial tool in diagnosing and managing sickle cell disease (SCD). Manual classification techniques have evolved into...

Full description

Saved in:
Bibliographic Details
Main Authors: Ximo Gual-Arnau, Lluïsa Gual-Vayà
Format: Article
Language:English
Published: MDPI AG 2024-12-01
Series:Mathematical and Computational Applications
Subjects:
Online Access:https://www.mdpi.com/2297-8747/29/6/122
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The classification of red blood cells (RBCs) or erythrocytes into three categories based on their shape, normal, sickle-shaped, and those with other deformations, has proven to be a crucial tool in diagnosing and managing sickle cell disease (SCD). Manual classification techniques have evolved into automated tools, with numerous classification methods being applied based on different ways of representing the cells. In this work, we propose a novel methodology for representing RBCs, defined by selecting <i>k</i> landmarks along the cell boundaries and characterizing shapes as points in the Kendall space of reflection shapes, <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msubsup><mo>Ω</mo><mn>2</mn><mi>k</mi></msubsup></semantics></math></inline-formula>. Using this representation, we applied an embedding of the Kendall space <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msubsup><mo>Ω</mo><mn>2</mn><mi>k</mi></msubsup></semantics></math></inline-formula> into a Euclidean space, which allowed for the use of machine learning classification algorithms. We also compared our results with those obtained using other classification methods applied to the same dataset in the literature, highlighting the strong performance of our approach in terms of classification accuracy.
ISSN:1300-686X
2297-8747