The Fertiboar technology allows to predict boar semen quality based on testicular ultrasound measures prior to first semen collection.
With it, producers can identify boars with the best semen quality and select these boars for artificial insemination. A machine learning approach is used to identify regions within the testicular pictures that are key to semen production. Those are analysed to calculate parameters describing novel attributes like the homogeneity and echogenicity of the tissue.
In a second step, a supervised learning algorithm is applied to predict the probability that a boar will produce ejaculates of desired quality. Through continuous collection of ultra-sound data and associated semen quality from boar studs, the predictive power of this approach will help producers to continuously improve the quality of the semen, the efficiency of the boar places in artificial centres, and ultimately the fertility rates on the farm due to higher standards in semen quality.