Principal Investigator: Parminder Basran

Co-PI: Kilian Weinberger, Matthias Wieland, Ian Porter, Julio Giordano

DESCRIPTION (provided by applicant): 

In the dairy industry, milk production is primarily affected by lameness, fertility, and the health of dairy cows. Most significant of these health factors is mastitis, or the infection of pathogenic bacteria entering the mammary gland through the teat canal, which results in markedly reduced milk production and a reduced immune system. Sustaining the integrity of the teat canal and its adjacent tissues is critical to resist infection. Machine milking causes changes to the teat tissue that impact teat canal integrity, tissue pliability, and thus, the teats’ defense mechanisms against mastitis pathogens. Machine milking-induced alterations include short-term and long-term changes. Short-term changes refer to teat tissue responses to a single milking and are differentiated into congestion and edema. Long-term changes are the adaptation of the teat tissue over several weeks and are classified into changes of teat end callosity thickness and roughness. Assessment of short- and long-term teat health conditions are normally conducted through physical examinations of dairy cows: costly and time-consuming manual methods that are prone to human errors. We lack objective and automated methods to efficiently assess teat tissue condition in dairy cows leaving large knowledge gaps in what is one of the most principal factors of udder health management. Here, we propose to develop two independent methods for automated digital assessment of teat tissue condition with machine learning and bypassing the need for physical measurements. Our global hypothesis is that manual teat tissue condition assessments in dairy cows can be replaced by machine learning methods. In the requested 2-year funding period, we will address the global hypothesis specifically by the following two objectives: Objective 1: Develop a multi-modality machine learning system to predict short-term changes to the teat tissue. Objective 2: Develop an image and video based deep learning classifier to predict long-term changes to the teat tissue.

Our overarching long-term goal is to improve the efficiency and accuracy of health monitoring strategies in dairy cows through automation and machine learning. We expect that our results will have a substantial impact on proposed milk harvesting strategies and encourage dairy producers to advance udder health, animal well-being, and the sustainability of their farms.