Paper Key : IRJ************328
Author: Devanila.j,Dr.ilamchezhian.j,Sarala Devi.v
Date Published: 12 Apr 2024
Abstract
In this project, we propose a robust and efficient approach for blood cell detection and counting utilizing YOLOv8, a state-of-the-art object detection algorithm. The input comprises a collection of microscopic images representing blood samples, with the primary objective of identifying and quantifying three distinct classes: Red Blood Cells (RBC), White Blood Cells (WBC), and platelets. YOLOv8, known for its superior real-time object detection capabilities, is employed to simultaneously localize and classify these blood cell types within complex and heterogeneous microscopic images. Our methodology involves training the YOLOv8 model on a carefully curated dataset containing annotated examples of RBCs, WBCs, and platelets. Subsequently, the trained model is applied to unseen images, demonstrating its proficiency in accurately detecting and counting each class. The proposed approach not only facilitates efficient blood cell analysis but also holds promise for automating the labor-intensive task of manual cell counting in medical diagnostics, thereby contributing to the advancement of clinical research and healthcare applications.
DOI Requested