Predicting musculoskeletal loading from video during rehabilitation

Project details

Background

Knowledge of tissue-specific loads on the muscles, bones, tendons, and ligaments of the human musculoskeletal system has implications for the prescription and monitoring of exercises across clinical, sporting, and occupational rehabilitation. These loads can currently be estimated via expensive, challenging, and time-consuming laboratory and/or computer simulation methods but field-based measures of ‘training load’ are mostly limited to whole-body summary metrics such as speed or distance (via Global Positioning Systems), or cumulative accelerations experienced at the upper back (via accelerometer). Recent developments, however, have seen the use of multiple video recordings from different angles to estimate three-dimensional body configuration as an input to musculoskeletal models estimating internal loading.

Your Project

Your project will build upon these developments. Artificial intelligence (AI) methods will be used to estimate three-dimensional body positions (model inputs) from a single video recording during common lower-limb rehabilitation exercises. You will adapt current modelling solutions for use within sporting populations and sporting rehabilitation exercises. Body configurations and musculoskeletal load estimates will be evaluated against current laboratory and computer-based approaches prior to their application to quantify loads during common rehabilitation progressions. The project may extend to include activity recognition of specific exercises within video recordings of larger rehabilitation sessions. Sensitivity analyses will determine the model parameters for which greatest accuracy is necessary and the simplest inputs needed for successful and efficient prediction or scaling of those parameters. Methods may also include data augmentation approaches to generate a great number of synthetic biomechanical datasets from lab-based recordings for the purposes of algorithm training.

Your Group

Loughborough University has been ranked number one in the world for sport-related subjects for six consecutive years (QS World University Rankings). We also boast an outstanding and fast-growing research portfolio in artificial intelligence research. In REF 2021, 100% of research impact from the School of Sport, Exercise and Health Sciences and 100% of the AI research were rated 'world-leading' or 'internationally excellent'. You will be part of our Sports Biomechanics and Motor Control team, one of the world’s largest sports biomechanics research groups, historically specialising in the computer simulation of sporting movements. We are a community based on mutual support and collaboration. You will research, learn, and practise alongside a team of senior academics, Doctoral Researchers, and postgraduate taught students working in similar research areas. Through our Doctoral College, postgraduate courses, group activities, and continuing professional development programmes, there are continual opportunities for building important research skills and networks among your peers and research academics.

Your Supervisors

Dr McErlain-Naylor’s primary ongoing research focus is the estimation of tissue-specific internal biomechanical load during training and/rehabilitation, and its association with adaptation/injury. This research works towards bridging the gaps between research and practice, between lab and field, and between computer simulation and affordable solutions that scale. Dr Cai is within the Department of Computer Science and will co-supervise this project. His research interests include computer vision, motion recognition, and object detection/segmentation.

Entry requirements

Our entry requirements are listed using standard UK undergraduate degree classifications i.e. first-class honours, upper second-class honours and lower second-class honours. To learn the equivalent for your country, please choose it from the drop-down below.

Entry requirements for United Kingdom

Applicants should have, or expect to achieve, at least a 2:1 bachelor’s degree (or equivalent) in sport and exercise science, biomechanics, engineering, computer science, or a related subject. A relevant master’s degree and/or experience in one or more of the following will be an advantage but not essential: 3D motion capture; machine learning approaches; and Python, MATLAB, or similar programming language.

English language requirements

Applicants must meet the minimum English language requirements. Further details are available on the International website.

How to apply

All applications should be made online. Under programme name, select School of Sport, Exercise and Health Sciences. Please quote the advertised reference number: SSEHS/SMNHC.

To avoid delays in processing your application, please ensure that you submit the minimum supporting documents.