Technical details

In order to reassure horse owners and give them greater knowledge about their horses, we employ a number of modern technologies, where 3D printing, IoT, Computer Vision and Deep Learning are key concepts.

The platform is built around an online single-board computer with a camera and modern image recognition and machine learning algorithms to detect events and patterns. We have a fondness for open-source technologies and frameworks. Python, OpenCV and Linux are therefore central to our technology solution.

The server side and data acquisition platform is a modern cloud-based back-end system that is based on modern technologies and platforms, including Kubernetes, Spring Boot, Java and Amazon Aurora. Our front-ends, which are easy-to-use and fast, are developed in modern web technologies such as JavaScript, React.js and Apache Cordova.

We use continuous delivery and launch changes as quickly as possible with the help of automation and devOps.

In order to classify the behaviour patterns of horses, we utilize about ten analysis steps where multiple deep neural networks play a central role. We have created our own tools to manually classify recorded events and we continuously expand our large image database to train our image analysis algorithms