Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
DELTA
This project investigates the possibility to use low budget sensors such as webcams and IMUs to measure movement of stroke patients and quantify the movement quality. This low cost approach will allow to scale the solutions and bring instrumented solutions into clinical application. Integral part of this project is to develop and validate algorithms, create user-friendly apps and translate the new technology into clinical application. This project is a collaboration between ETH and cereneo foundation and is thus based in Zurich and Vitznau/Hertenstein.
Movement quality describes the relationship between compensatory movement strategies and the movement that a healthy person would perform to accomplish a specific task. Movement quality is rarely assessed with current standardized clinical tests, and when it is, it is mainly visually scored by an on-site therapist. The goal of this project is to quantify movement quality using measurements. To ensure that our developed assessment tool can be used across the entire continuum of care (clinic and at home), we rely on low-cost mobile sensors such as commercially available webcams combined with state-of-the-art computer vision. An instrumented assessment tool that the vast majority of people have access to and know how to use would greatly improve the objectivity and frequency of movement quality assessments. We are currently finishing data collection from chronic stroke patients with a variety of sensors (multiple cameras, depth cameras, motion capture, IMUs). The participants perform tasks such as the drinking task or Box and Block test. The goal is to find the sensor system configuration which has the best trade-off between cost, usability and accuracy in movement representation. Depending on your interest, background and type of project (bachelor-, master thesis / intern) there are different projects available (see Job description).
Movement quality describes the relationship between compensatory movement strategies and the movement that a healthy person would perform to accomplish a specific task. Movement quality is rarely assessed with current standardized clinical tests, and when it is, it is mainly visually scored by an on-site therapist. The goal of this project is to quantify movement quality using measurements. To ensure that our developed assessment tool can be used across the entire continuum of care (clinic and at home), we rely on low-cost mobile sensors such as commercially available webcams combined with state-of-the-art computer vision. An instrumented assessment tool that the vast majority of people have access to and know how to use would greatly improve the objectivity and frequency of movement quality assessments. We are currently finishing data collection from chronic stroke patients with a variety of sensors (multiple cameras, depth cameras, motion capture, IMUs). The participants perform tasks such as the drinking task or Box and Block test. The goal is to find the sensor system configuration which has the best trade-off between cost, usability and accuracy in movement representation. Depending on your interest, background and type of project (bachelor-, master thesis / intern) there are different projects available (see Job description).
As part of the project, you will be integrated into the clinical routine and have the chance to work with medical professionals on all aspects of neurorehabilitation. Besides working on your data analysis or algorithm development you will be involved in the patient measurements, which means you will collect parts of the data you use yourself.
Open positions:
1. Markerless Motion capture
We already have a basic implementation of markerless motion capture (motion capture with a single or multiple Webcams). There is still some room for improvement of the algorithm. Next, we want to package the algorithm into a user friendly app to make it useable in clinical context (like freemocap.org). Optimizing the algorithm, packaging into a Computer App and helping with translation of the technology into the clinic will be main work packages here.
2. From sensor recordings to OpenSim
Translation from different movement data (Mocap, IMU, keypoints from pose estimation algorithms) into an OpenSim upper body model. Goal of this project is to compare the kinematics of the model, based on different sensor recording data and compare it to Mocap (gold standard).
3. IMU Analysis
The basic model to calculate kinematics and movement quality measures from IMU data is implemented. However the algorithms do not provide yet an IMU standalone solution. Improving the algorithm and applying it to the data set will be main work package here.
4. New measures of movement quality
Movement Quality is currently defined through discrete kinematic measures that have been empirically identified. However there are new methods to define the movement quality in a more data driven way by comparing kinematics from able-bodied participants to stroke patients. Researching these new methods, implementing them and applying it to the data set will be main work package here.
As part of the project, you will be integrated into the clinical routine and have the chance to work with medical professionals on all aspects of neurorehabilitation. Besides working on your data analysis or algorithm development you will be involved in the patient measurements, which means you will collect parts of the data you use yourself. Open positions:
1. Markerless Motion capture We already have a basic implementation of markerless motion capture (motion capture with a single or multiple Webcams). There is still some room for improvement of the algorithm. Next, we want to package the algorithm into a user friendly app to make it useable in clinical context (like freemocap.org). Optimizing the algorithm, packaging into a Computer App and helping with translation of the technology into the clinic will be main work packages here.
2. From sensor recordings to OpenSim Translation from different movement data (Mocap, IMU, keypoints from pose estimation algorithms) into an OpenSim upper body model. Goal of this project is to compare the kinematics of the model, based on different sensor recording data and compare it to Mocap (gold standard).
3. IMU Analysis The basic model to calculate kinematics and movement quality measures from IMU data is implemented. However the algorithms do not provide yet an IMU standalone solution. Improving the algorithm and applying it to the data set will be main work package here.
4. New measures of movement quality Movement Quality is currently defined through discrete kinematic measures that have been empirically identified. However there are new methods to define the movement quality in a more data driven way by comparing kinematics from able-bodied participants to stroke patients. Researching these new methods, implementing them and applying it to the data set will be main work package here.
• Literature research (20%)
• Method development(20%)
• Data collection (10%)
• Data analysis (30%)
• Communication and writing (20%)
• Literature research (20%) • Method development(20%) • Data collection (10%) • Data analysis (30%) • Communication and writing (20%)
• Background in Engineering, Human Movement, Neuroscience, informatics or a similar field
• Basic Coding skills (Python/Matlab) are required
• Experience in Motion Capture (not mandatory but the more you know the better)
• Data analysis
• Reliability and structured working
• Enthusiastic, friendly, and communicative approach to work
• Background in Engineering, Human Movement, Neuroscience, informatics or a similar field • Basic Coding skills (Python/Matlab) are required • Experience in Motion Capture (not mandatory but the more you know the better) • Data analysis • Reliability and structured working • Enthusiastic, friendly, and communicative approach to work