A Two-stage AI Framework to Detect and Classify White Blood Cells for Supporting Diseases Diagnosis in Veterinary Medicine

In veterinary medicine, the analysis of blood smears is crucial for diagnosing diseases such as systemic inflammatory response syndrome (SIRS) and sepsis, necessitating the identification and classification of white blood cells. Traditionally, this analysis is performed manually by observers, a process that is not only time-consuming and labor-intensive but also prone to variability in results between different observers. To address these challenges, this study introduces a two-stage framework that automates the detection and classification of white blood cells in smear images. Utilizing the YOLO-v8 model to detect all intact cells and the DenseNet model for classifying six distinct cell types, the framework aims to streamline the diagnostic process. Experimental results for the proposed two-stage framework demonstrate a mAP@50 of 0.964 for white blood cells detection and an accuracy of 0.836 for classification, surpassing conventional single-object detection models in both detection accuracy and classification efficacy.