Horse Activity Recognition Pipeline
Problem Statement:
The wellbeing of animals can be improved through online animal monitoring. Animal activity can be recognized on a collar that is attached to the animal. The classified activities provide information regarding the animal and its environment.
This project aims to implement and assess the performance of an animal monitoring pipeline using a real-world horse dataset.
Task:
Firstly, a state of the art deep learning model must be implemented to classify 3 to 5 different activities with adequate performance.
The classified activities must be stored in a database and visualized on a dashboard.
Students must keep in mind that eventually, the classifier must be portable to an embedded system, which implies resource restrictions in processing, memory, and battery life.
The classifier pipeline must be developed using python and pandas to allow the results to be used in related projects. Students are free in the other design choices.
Work:
15% Theory, 70% Simulations, 15%Writing
Contact:
Jacob Kamminga, j.w.kamminga@utwente.nl, room ZI5011