Deep Transfer Learning-Based Fecal Parasitic Recognition Using P3 Prototype Slide Scanner Imagery

To diagnose and treat parasitic illnesses, it is essential to identify and classify fecal parasite eggs. This helps with global public health initiatives. A method employing P3 Prototype Slide Scanner Imagery to train YOLOv8 Transfer Learning was given in the research. Comprehensive training, evaluation, and testing steps were used to gauge the model's performance. Remarkably, the model's mean Average accuracy (mAP) was 81.5%, while its recall and accuracy rates were 85.8% and 74.6%, respectively. Overall, our findings highlight the potential of the YOLOv8 model as an efficient tool in automated fecal parasitic egg identification, with implications for clinical diagnostics and parasitology research. Testing and inference results showed good detection accuracy. Sustained improvement and optimization are necessary to overcome constraints and strengthen the model's resilience for practical use. By using deep learning methods to enhance fecal parasite egg identification and classification, this work advances the discipline of parasitology and makes illness diagnosis and public health interventions easier.