Innovative technology to select for optimal farrowing characteristics

Published on Feb. 9, 2023

Innovative technology to select for optimal farrowing characteristics

Through R&D innovation, the Hendrix Genetics Swine/Hypor team has developed specialized programming to capture data on farrowing time for individual sows. This technology is now being utilized at the Hypor Bon Accord nucleus farm to select sows with improved litter characteristics.

Farrowing time has a major impact on the livability and long-term health of the piglets and the sow. Much like laboring for humans, farrowing time can have a huge range depending on the individual sow. Duration of farrowing has been shown to have a negative impact on the proportion of stillborn piglets, i.e., a longer farrowing process increases the risk of stillborn piglets.

In some instances, farrowing can last over 24 hours, going from morning until well into the night. For this reason, it can be difficult to capture data on farrowing time unless farm laborers are present with the sow throughout the process. For this reason, we have developed specialized software for capturing data on farrowing.

Benefits of machine learning

What is machine learning? Machine learning is the use of computer systems that can use algorithms and statistical models to find patterns in data. For our purposes, the goal was to develop a program that could automatically capture and analyze data on farrowing piglets.

To capture images of the farrowing process, smart cameras were installed over the farrowing pens at Bon Accord. Using object detection technology, images are analyzed, and the program can automatically identify when each piglet is born. With machine learning technology, we can efficiently collect data on farrowing time for individual sows. Data is also collected on the time between birthing each piglet . There are future considerations for this technology that have exciting possibilities. For example, we can efficiently identify live or stillborn piglets, or we can identify piglet uniformity and select sows that have more uniform piglets and lower pre-weaning mortality.

farrowing image 1
Figure 1: 7 piglets detected
image of farrowing 2
Figure 2: 9 piglets detected

Selecting for sustainable litter characteristics

This data has important implications for achieving sustainable litter characteristics, particularly for our 16-15-14 standard (16 total born, 15 born alive, and 14 weaned). The sows that farrow their piglets quickly, with consistency between each piglet, are more likely to have litters with higher livability. This is because the last piglets in the litter to be born spend less time in the birth canal and are more likely to be born alive.

Click here to learn more about our sustainable standards.

Innovative selection methods, such as machine learning, drive continuous improvement in genetic performance. With enhanced methods for phenotyping, we can select sows that produce healthy, uniform piglets. For the commercial swine producer, this translates to better health and welfare for the sow, piglets that thrive after birth and weaning, and getting the best economic return in your operations. With these combined benefits, it all adds up to sustainable swine breeding for the future of the industry.

Related articles

, by
Hypor to acquire KCAP multiplier and agreement of sale of product through Cornbelt Livestock Services

Missouri, US - February 14th, 2024Hypor strengthens its presence in the US swine market with strategic acquisition of KC Air Pork LLC (KCAP) breeding stock. Additionally, Hypor and Cornbelt Livestock…

, by
Cooperativa Virgen del Rosario and Hendrix Genetics Announce Sale of Gratal Farm and Long-Term Partnership Agreement

Huesca, Spain - February 1st, 2024Cooperativa Virgen del Rosario and Hendrix Genetics are pleased to announce the signing of a groundbreaking agreement on December 1st, 2023, marking the sale of Grat…

“She is a really independent sow”: A Hypor customer success story

Christelle and Jean-Pierre OLIVIER are parent stock producer based in France.