The Royal Society Fully Funded PhD Studentship: Deep Learning to improve Augmented Reality Application

Manchester Metropolitan University - Manchester Metropolitan and Image Metrics Ltd

Summary

This is a joint PhD studentship between Manchester Metropolitan and Image Metrics Ltd, funded by The Royal Society. This project will improve the realism of facial appearance try-on technology by developing a novel and light-weight solution for real-time inpainting. This research will investigate deep learning architectures for various augmented reality tasks.

Aims and objectives

Virtual try-on technology assists consumer in making purchase decision when shop online. With the popularity of mobile devices, virtual try-on for faces has been widely used for cosmetic and gaming. In medical practices for face reconstruction surgery, there is a growing need to be able to find the realistic appearance for people with face defects. In addition, the ability to predict a face with occlusion can improve the face recognition rate in security application. This project proposes a new technique that can inpaint the missing face region with realistic appearance. Recent work shows the ability to inpaint face regions however these methods only work on small images and take a long processing time. This project will design a solution using a machine intelligence method that work efficiently on mobile devices. The solution will be tested and deployed in collaboration with Image Metrics Ltd in real-world applications.

The PhD candidate will receive professional development training from the University, the industry partner and attend external training. To ensure the candidate has exposure to commercial setting, s/he will spend 60% of their time at Manchester Metropolitan University and 40% at Image Metrics Ltd (http://image-metrics.com). The joint supervision will broaden the perspective on the research impact and enrich the student experience as s/he gains a wider understanding of applied research and different training environments. The PhD student will have access to training, facilities and expertise in both organisations, which is very valuable particularly in enhancing their employability, ideally becoming a leader in her/his field. With such setting, the student will benefit from different algorithm/software development with an applied or translational dimension.

Specific requirements of the project

Candidates must have a strong motivation for research and excellent programming skills. We welcome applications from candidates with computational science who wish to develop deep learning/AI skills and augmented reality for real-world applications. Qualifications:

  • A high grade undergraduate degree (first class or upper second) in Computer Science/Engineering/Mathematics or other background with strong programming skills
  • A MSc level in Computational Science would be desirable for this post
  • Knowledge of programming, for example C/C++/Java/python, OpenCV or Matlab
  • Good communication and writing skills

Home/EU fees will be covered, a stipend of £14,777 per annum and additional fund for training and conference attendance. The candidate will also receive training on deep learning techniques and augmented reality, alongside aspects of face detection and recognition.

This opportunity is open to Home/EU and International applicants. Please note, International applicants must cover the difference in fees.

Informal enquiries to:

Moi Hoon Yap m.yap@mmu.ac.uk

Click here to apply: https://www2.mmu.ac.uk/research/research-study/scholarships/detail/scieng-mhy-2018-1-the-royal-society-fully-funded-phd-studentship-deep-learning-to-improve-augmented-reality-application.php.

Closing date: 31st July

Interviews: August

Share this PhD
     
  Share by Email   Print this job   More sharing options
We value your feedback on the quality of our adverts. If you have a comment to make about the overall quality of this advert, or its categorisation then please send us your feedback
Advert information

Type / Role:

PhD

Location(s):

Northern England