Prof. Dr. Tilo Mentler: Establishment of a laboratory environment for the practical examination of AI-specific topics. In particular, students at Trier University of Applied Sciences will be able to "experience" AI by working on tasks that go beyond the pure use of software frameworks and include hardware-related and design aspects. To this end, AI methodological skills are paired with application skills in demonstrators and individual partner projects, thus creating an interdisciplinary teaching and learning structure.
31. JANUARY 2022
> Treadmill with 3D force measurement
> High-performance computer
> Smartwatches
> 3D Cameras
> VR Headsets
A teaching format for computer science students consists of "drawing" digits (0-9) in the air with a smartwatch worn on the wrist and recording the corresponding acceleration data with the sensors built into the watch. Subsequently, a classification model is trained that classifies the digits based on the patterns in the collected data. Finally, the predictions of the model are tested live.
Another application of AI with a therapeutic reference is the prediction of biomechanical parameters, some of which have a significance for the risk of injury and/or can provide information about changes in the movement pattern of patients in the context of rehabilitation processes. However, due to cost and space reasons, the measurement methods required for this are hardly widespread. The prediction of such biomechanical parameters using video data and AI could make more complex motion analyses much more accessible.
Such an AI application will be implemented as part of the project. For the videos of running movements, a so-called keypoint data set is generated using AI, whereby these keypoints are assigned to "key points" of the body and thus represent the skeleton or its movement in a simplified way. On the basis of this keypoint data, parameters of interest can be predicted with other AI models.
Various aspects can be considered, worked on and evaluated: Computer science students should optimize the implementation of the corresponding AI, for example, as part of an exercise or significantly expand it during a project work, thus increasing its practical suitability. Students of therapy sciences gain insights into the implementation and evaluate the results of AI in terms of their accuracy and thus applicability in therapeutic practice.
Students of computer science are given the opportunity to learn and deepen various aspects of artificial intelligence through a variety of practice-oriented, hardware-related tasks.
One aspect focuses on collecting and processing raw data that comes directly from the project hardware. This allows students to train how to handle real-world data – whether it's evaluating acceleration data from smartwatches or processing video material.
Another competency taught is to apply existing AI models and integrate them into their own projects. Students can easily try out Pose Estimation models, image classifiers, object recognition models and many other types of machine learning models and use them for their own applications.
Particular attention is paid to the most important competence: the independent development of models. For example, students can predict biomechanical parameters or classify acceleration data. The focus is always on practical application.
Students of therapy sciences will be shown practical applications of AI. They are encouraged to use such applications independently and to critically reflect and discuss their potentials and limitations and thus their usability in therapeutic practice.
AI-generated data and results are evaluated and validated to identify strengths and weaknesses as well as areas for improvement.
A strong focus is placed on the biomechanical analysis of movement.
You are leaving the official website of Trier University of Applied Sciences
