Innovatieve tools voor de ondersteuning van preventieve en curatieve klauwverzorging bij melkvee. (CLAWCARE)

The aim of this VLAIO-LA project is to reduce claw problems in dairy cows through an affordable intelligent app that can process thermal camera images (on a mobile phone for example). This will require development of a great deal of diverse knowledge. Claw problems are one of the most important health disorders in dairy cattle. They have serious economic consequences for the farm and they have a negative impact on the cow's welfare because a claw problem can lead to lameness. The project partners, being UGent (coordinator), ILVO, Inagro, Hooibeekhoeve and HoGent want to significantly improve the current farm practice of detection and treatment. The game-changer is the use of mobile infrared cameras with thermography as a first-line tool that can be handled by the farmer. These have become more affordable. It often turns out that a claw injury can already be found via the heat camera, while no problems can be detected visually. However, the biggest work is to develop a corresponding self-learning image processing algorithm for the automatic detection of all kinds of claw problems.

We record a database of thermal claw images using mobile thermal (infrared) cameras to develop a self-learning image processing algorithm. This database of infrared images, a collaboration between ILVO and UGent, will be used to (1) validate whether they can be used to more accurately determine the location of claw injuries and (2) to investigate whether certain types of claw problems can be automatically detected and classified using machine learning techniques. So everything ends up in a self-learning image processing algorithm. ILVO specifically puts its shoulders to the comparison of images from a powerful, expensive thermal camera with images from the smaller, more affordable cameras that can be connected to the mobile phone. The aim is to find correlations between the expensive and the inexpensive. The first step is, for example, to have horn distinguished from skin by the algorithm. The next step is to learn to recognize injuries correctly in the images of the powerful thermal camera. The final phase will be a software tool that is reliable and affordable for claw carers and dairy farmers. This tool determines how deep they may and must cut to cure the claw problem. The researchers work closely with dairy farmers, professional claw trimmers and veterinarians.

The importance of claw health is still too often being underestimated. The self-learning software will help to break through this "blind spot". We expect that the software tool will contribute to the general awareness and increase in knowledge about the importance of claw health, hopefully with a decrease in prevalence of claw problems as a result.