4 Jun 2024
Hoofcount’s Pedivue was developed with the help of a £250,000 investment from Defra’s Farming Innovation Programme.
Vet Times
Job Title
Hoofcount’s Pedivue has developed a lameness detection device which uses AI.
A ground-breaking lameness detection device that uses artificial intelligence (AI) to allow farmers to detect lameness earlier in their cows has been launched.
Hoofcount, which is a leading UK player in automatic footpaths, has launched Pedivue, marking the completion of a two-year project funded with £250,000 from UK Research and Innovation, part of Defra’s Farming Innovation Programme.
The system was designed using computer vision and machine learning by Hoofcount with the UK Agri-Tech Centre at the South West Dairy Development Centre and in collaboration with farmers and the Centre for Machine Vision (CMV) at the University of the West of England, Bristol.
Hoofcount’s Anthony Marsh said: “Digital dermatitis affects more than 90% of dairy herds in the UK contributing to lameness, which continues to be one of the main, if not biggest, issue within the sector costing an estimated £300 per incident.
“While Hoofcount’s automatic footbath system is helping reduce and control lameness – particularly digital dermatitis on numerous dairy units throughout the UK and overseas – we agreed we needed to develop a device to detect these issues as early as possible in order to enable prompt effective treatment.
“After designing and trialing various options, we soon became aware that machine learning and AI was the way forward to gain best results.”
Mr Marsh continued: “With both financial and technical support, together with guidance and direction, we have been able to develop Pedivue. The system’s results, clarity and accuracy, even in its early stages of development have already gained interest and support from vets and hoof trimmers.”
CMV’s Wenhao Zhang added: “Hoofcount, together with CMV’s help, developed Pedivue’s computer vision and machine learning algorithm by retrofitting a single camera and an AI system to an existing Hoofcount footbath.
“The high-speed camera captured crucial moments of moving hooves, enabling an intelligent filtering algorithm to capture clear, square-on views of hoof soles for AI assessment. By ‘learning’ from extensive hoof data from four farms alongside veterinary expertise, the AI-powered system initially achieved over 80% accuracy in detecting active digital dermatitis lesions.”
