Research Update CHF 03045-A MY2: Artificial Intelligence and Machine Learning for Diagnosis of Dog Sperm Morphology

Collection of all semen samples required for the study has been completed. Samples from 62 dogs presented to the veterinary hospital at The Ohio State University have been obtained and preserved. At the University of Pennsylvania, slides were prepared, and images acquired using phase-contrast under 1000X magnification. A bank with 11,464 images has been created. Currently, we are working on adapting an algorithm initially developed to identify and crop bovine sperm images to generate images containing a single dog sperm. Once that process is concluded, the images will be individually classified by two specialists and the annotated image bank will be used to train a sperm morphology evaluation algorithm.