Smartphone Detection of Abnormal Equine Behavior
Type of DegreePhD Dissertation
DepartmentComputer Science and Software Engineering
MetadataShow full item record
This research was motivated by two key facts: There are over 9.2 million horses in the United States and over 64% of adults own a smartphone. Blending these two ideas led to the genesis of this research by asking the question: “Can an unmodified, off-the-shelf smartphone be used to detect and categorize behavior of an equine in a controlled setting?” This research used computer vision techniques and extended game-based modeling to describe patterns of behavior that are considered normal, to determine when observed behavior falls outside those patterns, and to diagnose the possible cause of the anomaly. The research resulted in a proof-of-feasibility system that demonstrated use of a smartphone to differentiate normal behavior from abnormal behavior – pawing, in this case – of an equine while in a stall.