Ongeveer 22 uur geleden - Rijksuniversiteit Groningen (RUG) - Groningen
- you will design and perform laboratory experiments with marine micro-algal cultures - you will explore new techniques to measure carbon uptake and photosynth…
The goal of this project is to develop algorithms that analyze videoclips of human movement in sports (e.g. sprinting, jumping, kicking a ball in soccer), evaluate the quality of the performed movement, and automatically produce recommendations for improving the movement.
The goal of this project is to develop algorithms that analyze videoclips of human movement in sports (e.g. sprinting, jumping, kicking a ball in soccer), evaluate the quality of the performed movement, and automatically produce recommendations for improving the movement. The PhD student will explore state-of-the-art machine learning (mainly deep learning) approaches and develop prototypes for movement analysis. The algorithms will be used in (mobile) applications in which a user can load a videoclip, submit it to the algorithms for analysis, and receive the analysis results with recommendations for improvement.
The to-be-developed algorithms have a wide range of potential application domains including, but not limited to, soccer, athletics, rowing, skating, gymnastics, and swimming. The algorithms solve the problem that access to human experts that can provide instructions on how to best improve a movement is highly limited. Seen from a broader perspective, this project helps to increase the overall training level of a sport, which will have a positive effect on the feeling of well-being of the players/athletes, on the amateur level of a sport and in extension on the professional level of that sport.
The applications will be developed by a commercial partner and will include additional functionality providing the user with information on how the movement can be improved, for example by suggesting a number of training exercises. The embedding of the project within JADS and its entrepreneurial ecosystem allows the candidate to perform fundamental machine learning research on video analysis inside a stimulating academic and entrepreneurial environment.
The project is a collaboration of the Jheronimus Academy of Data Science (JADS), 's-Hertogenbosch (campus Mariënburg), Tilburg University (TiU), Eindhoven University of Technology (TU/e) and commercial partners. The project also cooperates with the Department of Human Movement Sciences of VU Amsterdam.
The research will be conducted under supervision of Prof. Dr. Eric Postma and Prof. Dr. Ir. Wim Nuijten. It will deliver both long-term results (understanding of machine learning in video analysis) and mid-term results (algorithms, papers, and best practices).
The successful candidate is expected to:
The PhD student will be employed at Tilburg University.