This 3.5-year PhD research studentship will address creating a SLAM system leveraging both multi-view geometry and machine-learning based recognition to enhance the resulting 3D model with semantic attributes. State-of-the-art SLAM approaches exploit mainly correspondences established across multiple images and utilize the underlying theory of multi-view geometry to generate a 3-dimensional map of the environment over time. Semantic scene understanding plays no or only a minor role in achieving this goal in current SLAM methods. Semantic understanding of the environment enriches the representation of the 3D environment, but can additional provide information leading to better 3D models. One particular focus of this project is to obtain a method for semantic SLAM that operates in live and in real-time in order to make the system applicable to autonomous systems.
You will be a student of the University of Oxford with a 3-6 months industrial internship at Toshiba’s Cambridge Research Laboratory in your 2nd or 3rd year. You will work under supervision of Professor Philip Torr (University of Oxford) and Dr Christopher Zach (Toshiba Research Europe).
- Proven academic excellence (at least first or upper second class undergraduate degree)
- Confidence in mathematical skills (e.g. geometry, linear equations, optimization)
- Strong coding skills in Matlab/C++ (some experience with CUDA is an advantage)
- Motivated and excellent problem solving skills
- Good verbal and written communication
- Prior publications in the field are an advantage
- UK residency (British Citizenship or EU Citizenship with 3 years residency in the UK)
About Torr Vision Group
Formed in 2005 at Oxford Brookes, TVG moved to the University of Oxford in 2013, and is led by Professor Philip Torr. It comprises around 20-25 people. Members of the group have won major awards in all the main conferences in the field including ICCV, CVPR, ECCV, BMVC, NIPS.
The aim of the group is to engage in state of the art research into the mathematical theory of computer vision and artificial intelligence, but to keep the mathematical research relevant to the needs of society. A particular emphasis of the group has been on real time understanding and reconstruction of the world using mobile cameras, such as those on drones, intelligent glasses or other robots.
CRL's Computer Vision Group (CVG) is working on new methods for 3D Computer Vision, Machine Learning, and Optimization applied to visual data. CVG was founded in 2006 and is currently managed by Dr Stephan Liwicki with scientific guidance by Dr Christopher Zach and strategic advice by Professor Roberto Cipolla FREng. Our mission is to conduct fundamental research for future technology and products.
The current focus of the group is 3D reconstruction and machine learning, and members of CVG work on multi-view geometry, SLAM, dense 3D modelling, and deep learning in particular. Since our target applications are in the area of autonomous systems and social infrastructure, we are especially interested in real-time systems.
How to Apply
Please send your CV and covering letter to Prof. Philip Torr at firstname.lastname@example.org.
16th April 2017
Research Internship (PhD Level)
The Toshiba Research Computer Vision Group in Cambridge is looking for interns. We offer a limited number of internships to PhD students in their 2nd year or higher. (Please note that permission to work in the UK is a necessary condition to apply.)
We are a young team of researchers working in exciting areas of computer vision, including 3D reconstruction, recognition, gesture recognition and body capture. The Toshiba lab is closely linked to academia and we encourage scientific publication.
Interns will be working for three months on a well-defined project and will be assigned an experienced researcher as mentor. Current topics of interest include: visual SLAM, 3D reconstruction, and Machine Learning (e.g. CNNs). Strong programming skills are essential.
The Toshiba office is located in the Science Park of the university city of Cambridge. Summer activities such as punting on the river Cam, concerts in college chapels and picnic lunches provide fun during breaks from research.
Evidence of permission to work in the UK is required for this position. Please indicate this in your application. We consider applications throughout the year and internships can take place during a mutually agreeable time period.
To apply send your CV to email@example.com.