Identification of Stomatocytes through Microscopic Image Analysis of Blood Smears

The analysis of microscopic images of blood smears remains crucial in medical diagnostics, aiming to reveal abnormalities related to blood cells, particularly red blood cells. These abnormalities, whether morphological or colorimetric, allow for the precise detection of both common and rare diseases, as certain anomalies are clear indicators of specific pathologies. Stomatocytes, which are the focus of our study, are red blood cells exhibiting membrane defects that, to a certain extent, lead to increased permeability to sodium and potassium. These abnormal erythrocytes generally exhibit a morphology that is overall similar to that of normal red blood cells. However, the central pale area takes on a slit-like or elliptical shape instead of the typical round form. This specific feature, which distinguishes them from other cells, is indicative of a pathology known as stomatocytosis, which may be either congenital or acquired (such as in alcoholic cirrhosis or acute alcohol toxicity). Its diagnosis relies on a series of costly biological tests. However, the blood smear remains the essential examination due to the specific morphological characteristics of stomatocytes. This paper proposes a semi-automated characterization method for the clear identification of stomatocytes in blood smear images. Developed within the MATLAB environment, the method combines K-means pixel-based classification with algorithms designed to isolate the central pallor of the stomatocyte, followed by the extraction of distinctive features enabling its differentiation. The results obtained are highly promising, as stomatocytes in blood smear images are accurately identified using the proposed approach. Thus, the identification of stomatocytes is based on compactness, eccentricity characterized by the difference between the major and minor axes, as well as the proportion of red and white pixels.