®®®® SIIA Público

Título del libro: Proceedings Of The 21st International Symposium On Medical Information Processing And Analysis, Sipaim 2025
Título del capítulo: Deep learning for Chagas disease: YOLOv8-based automatic detection of amastigote nests

Autores UNAM:
NIDIYARE HEVIA MONTIEL;
Autores externos:

Idioma:

Año de publicación:
2025
Palabras clave:

Automatic Detection; Biomedical imaging; Chagas disease; Clinical manifestation; Deep learning; Global public health; Histopathological images; Objects detection; Trypanosoma cruzi parasites; YOLOv8


Resumen:

© 2025 IEEE.Chagas disease, caused by the Trypanosoma cruzi parasite, remains a global public health problem, with cases reported even outside its traditional endemic areas. Its primary clinical manifestation occurs at the cardiac level, characterized by the formation of amastigote nests. This work presents a deep learning algorithm for automatic detection of amastigote nests in cardiac histopathological images obtained from a murine model. For the nests? detection, a YOLOv8-based architecture was trained on histological images performing a 5-fold cross-validation, while the final tests were carried out with data unseen by the YOLOv8 from four image groups of different post-infection days. In training stage, YOLOv8 achieved 97.9% in accuracy, 98.6±0.3% in precision, and 97.2±0.8% in recall, while in the final test, the best result identified was 78.6% in accuracy, 76.4% in precision, and 81.1% in recall on images at 100× magnification. These metrics reveal the model?s ability to generalize well and its high sensitivity to unseen data. The results of this study show that the proposed algorithm is capable of identifying amastigote nests, which may be useful in clinical areas to understand the damage caused by Trypanosoma cruzi infection.


Entidades citadas de la UNAM: