STABLE - Smart Behaviour Learning for Horses
|Coordinator||Högskolan i Halmstad - Center for Applied Intelligent Systems Research (CAISR)|
|Funding from Vinnova||SEK 126 665|
|Project duration||November 2018 - September 2019|
|Venture||Personal mobility between societal sectors|
|Call||Funding for staff exchange and artificial intelligence (AI)|
Purpose and goal
The project´s goal is to create an intelligent monitoring system of horse behaviour and anomaly detection by applying state of the art neural network techniques on the data stream provided by a web camera based monitoring system. This will provide the horse owner with an improved summary of the horses behaviour, and provide an early identification of possible problems. The proposed solution will learn a horse´s normal behaviour from webcam image data, and will be able to autonomously detect deviations from this norm.
Expected results and effects
Expected results: A way to quantify the level of normal behaviour for a horse, and device a method to detect deviations from the normal level. Provide a working prototype. Ultimately, we are working with a company who is looking to implement an intelligent monitoring system. The result of the research must therefore also be implemented to test its viability outside a research environment. A paper describing the system and results.
Planned approach and implementation
The following steps have been proposed: Collect relevant data Determine and extract relevant features. Adapt the data for anomaly detection. Train and test anomaly detection on single horses. Discuss the results with experts and horse owners to identify why deviations were detected.