United KingdomAgriculture and food production2021 – ongoingGenus (PIC)
FertiBoar ultrasound analysis to predict boar semen quality
PIC's FertiBoar software predicted poor semen quality in young boars from ultrasound images, with over 90% of poor boars identified
Problem
PIC cited a 2014 study in which about 37% of boars sent to artificial insemination studs failed minimum semen quality standards. Poor boars were found only after months of housing and testing.
Approach
PIC and the IFN Schönow institute spent four years developing FertiBoar. Machine learning analyses testicular ultrasound images and tissue measurements to predict semen quality before a boar's first collection. The software was trained on images from over 1,500 boars and more than 500 ejaculate analyses. PIC began implementation in June 2021, with the first tested animals delivered in September 2021.
Results
Takeaway
The accuracy figure comes from PIC's own model testing. The savings PIC calculated are a worked example for a hypothetical stud, not measured results.
Categories
- Industry
- Agriculture and food production
- Business function
- Production and qualityIT and R&D
- Innovation type
- Process innovation
- AI technology
- Computer visionPredictive analytics and ML
- Adoption
- Partner co-development
- Evidence
- Company's own claim
Sources
Source summaries
PIC's article explains how FertiBoar predicts semen quality from ultrasound scans and how it was trained. It gives the accuracy figure and a hypothetical savings calculation for a 300-head stud. It reports the June 2021 start of implementation.
Full details are in the original sources.
