George McMaster began by milking eight cows and watching for signs of illness. In Louise Cullen’s 26 October BBC report, his County Antrim herd numbers 300 and an AI camera adds another view of the animals leaving the milking parlour. His son Jack says it flags cows they would not have recognised as lame during everyday observation.[1]
That is a farmer’s account of changed attention. Understanding what the camera adds starts with the movement record it makes.[1][2]
A 2023 University of Liverpool study describes the CattleEye system tracking an animal’s outline across video frames, assigning its identity using enrolled coat-pattern and head-shape information, then recording moving reference points. A neural network converts those changes into a mobility score between zero and 100. Higher scores indicate poorer movement. The paper’s minimum usable sequence was 40 frames over two seconds.[2]
That study compared 6,040 paired scoring records across three farms. It also checked a smaller set against foot lesions found during trimming. The score was an indicator to investigate: it identified 52% of cows with the study’s potentially painful lesions, at 81% specificity. It did not provide a complete view of every animal’s foot health.[2]
Putting the alert into the care routine
An English farm’s randomized trial appeared online on 3 July, then in October’s Journal of Dairy Science. Of 419 cows beyond their first lactation, 211 had additional weekly selection for treatment using CattleEye; 208 received usual care. A score of at least 50, or a 20-point rise over two weeks, could put an eligible cow on the trimming list.[3]
Severe lameness occurred in 4 of 201 assessed intervention cows, versus 16 of 203 controls: 2.0% against 7.9%. The intervention involved more trimming. Overall lameness prevalence, milk yield and culling did not differ significantly.[3]
The single-farm study’s researcher knew group assignments; farm staff and the foot trimmer did not. CattleEye helped discuss the study design under shared Innovate UK funding, but did not collect or analyse the data.[3]
The useful change was a different treatment list, followed by foot care.[3]
Sources & further reading
- From agriculture to AI-griculture: How tech is changing farming Louise Cullen / BBC News · 26 Oct 2026 · Accessed 28 Oct 2026
- Initial validation of an intelligent video surveillance system for automatic detection of dairy cattle lameness Alkiviadis Anagnostopoulos and colleagues / Frontiers in Veterinary Science · 13 Jun 2023 · Accessed 28 Oct 2026
- A randomized controlled trial evaluating the use of an intelligent, fully automated 2D imaging system to detect lame cows and control lameness Nektarios Siachos and colleagues / Journal of Dairy Science · 3 Jul 2026 · Accessed 28 Oct 2026