Cryptosporidiosis, caused by Cryptosporidium spp., is a significant zoonotic parasitic disease impacting neonatal calves, leading to severe diarrhea, dehydration, and substantial economic losses in the livestock industry. Rapid and accurate detection of Cryptosporidium oocysts is crucial for effective disease management and control, contributing to both animal and human health under a One Health perspective. This study investigates the application of state-of-the-art object detection algorithms, YOLOv10 and YOLOv11, for the automated identification of Cryptosporidium oocysts in microscopic images of calf fecal samples. A dataset of 406 annotated images was used to train and evaluate these models using metrics including precision, recall, and mean Average Precision (mAP). YOLOv11 demonstrated superior precision (88.94 %), indicating a reduced false-positive rate, which is critical for avoiding unnecessary treatments and for the accurate assessment of prevalence in epidemiological studies. Conversely, YOLOv10 exhibited higher recall (92.57 %), ensuring high sensitivity for screening purposes where minimizing false negatives is paramount. These findings highlight the potential of advanced object detection as a rapid, scalable, and cost-effective AI-assisted screening and support tool for Cryptosporidium. This automated system has the potential to standardize diagnostic procedures, facilitate high-throughput microscopy-based surveillance, and improve our understanding of oocyst shedding dynamics in infected animals.
扫码关注我们
求助内容:
应助结果提醒方式:
