3D human pose estimation and shape reconstruction for biomechanics
Saturday 01 August 2020
Closing Date: Review of applications will commence on 1 August 2020 and remain open until filled
Department: Bioengineering, Faculty of Engineering.
Applications are invited for an exciting fully-funded PhD studentship at the Faculty of Engineering, the University of Nottingham.
Research area. The research topic focuses on developing computer vision and machine learning based solutions that enable in-natura markerless motion capture for biomechanical modelling in Biomedical and Sports Engineering. Specifically, it addresses the fundamental research problem of reconstruction of person-specific human pose, kinematics, and surface geometry to enhance our understanding of the non-linear behaviour of human motion, musculoskeletal injury and disease and enable modelling of soft-tissue dynamics and human-object interaction.
The project. The candidate is expected to develop a fast and robust method for inferring and tracking 3D human pose and surface geometry. The method will be mainly based on visual sensing complemented by Inertial and force sensors. The method can use either or both of model-based and learning-based approaches, such as CNN based segmentation, geometric CNNs, or convolutional kernel filter based tracking. The candidate will have access to a newly established state-of-the-art motion capture laboratory.
The candidate. The ideal candidate will have;
a first or upper second class honours or Masters degree in Electrical and Electronic Engineering, Physics, Computer Science, or other relevant and equivalent degree from a quality recognised institution.
a solid background in mathematics and excellent analytical and numerical skills, as well as problem solving skills
strong background in 3D computer vision, pose estimation, shape reconstruction, structure from motion, segmentation, or object detection.
experience in image or video processing and digital signal processing.
strong programming skills in Matlab, C/C++, or Python. Previous hands-on experience with deep learning platforms and agile software development as well as experience of working within industry will be an advantage.
very good written and communication skills and fluency in English.
a driven, independent professional and self-reliant work attitude within a fast-paced & collaborative environment.
The offer. The scholarship on offer (to eligible students) is for a minimum of three years and includes a tax-free stipend of 15,285 per year (for 2020/21) and tuition fees. It is available to students of UK and EU nationality. Applicants must obtain the support of the potential supervisor prior to submitting their application.
Informal enquiries about the project may be addressed to Dr Ami Drory. Please (i) insert your cover letter, CV, copies of academic transcripts, a list of publications, and contact details for two academic referees into a single pdf file. (ii) Name the file with your name as ”firstName_lastName_phd”. (iii) e-mail to: Ami.Drory [ at ] nottingham.ac.uk, with [3D shape reconstruction PhD application – lastName] as the email subject. Applications without academic transcripts or academic referees will not be considered. Applicants are advised to include copies of any publications or examples of their technical writing, such as code projects, project report or dissertation in support of the application.
Application instructions. With the support of the potential supervisor, formal applications are to be made via http://www.nottingham.ac.uk/pgstudy/apply/applyonline.aspx.
Closing date for applications. Review of applications will commence on 1 August 2020 and remain open until filled. A start date is expected to be as soon as practical thereafter.