The evaluation of an artificial intelligence system for estrus detection in sows

Area of Science:

• Animal Science
• Artificial Intelligence
• Reproductive Biology

Background:

• Accurate estrus detection in sows is critical for timely insemination and optimal reproductive outcomes.
• Technological advancements, including AI-powered sensor and camera systems, offer new methods for estrus monitoring.

Purpose of the Study:

• To evaluate the effectiveness of an AI system in supporting farmers to optimize insemination timing.
• To assess the impact of the AI system on key reproductive performance parameters in sows.

Main Methods:

• A study was conducted across three Belgian sow farms over 1.5 years before and after AI system implementation.
• Reproductive data, including farrowing rate (FR), repeat-breeder percentage (RB), farrowing rate after first insemination (FRFI), and total born piglets (NTBP), were analyzed.
• Weaning-to-estrus interval (WEI), estrus duration (ED), and inseminations per estrus were also assessed using AI-collected data.

Main Results:

• Farm A showed significant improvements in FR (+4.3%), RB (-3.75%), FRFI (+6.2%), and NTBP (+1.06 piglets).
• Farm B experienced a significant decrease in NTBP (-0.48 piglets) potentially due to low insemination doses.
• Farm C saw a significant increase in NTBP (+0.45 piglets), with the AI system providing shorter WEI and less variable ED compared to farmer assessments. Inseminations per estrus decreased in Farm C.

Conclusions:

• The AI system demonstrates potential to enhance sow reproductive performance and refine estrus assessment.
• The system can aid in reducing the number of inseminations required per estrus cycle.
• Farm-specific factors like management and genetics significantly influence outcomes, necessitating tailored implementation strategies.