Visual recognition and learning algorithms for sign language and action recognition in broadcast video

Grade 7: £30,434 - £37,394 p.a.

We are seeking a full-time Postdoctoral Research Assistant in Computer Vision to join the Visual Geometry Group at the Department of Engineering Science (central Oxford). The post is funded by EPSRC and is fixed-term for 12 months with possible extension.

The aim of the project is to develop visual descriptors and learning algorithms for sign language and action recognition in broadcast video using weak supervision from the subtitles and other meta-information. It is a joint project between the University of Oxford (Andrew Zisserman), the University of Leeds (David Hogg) and the University of Surrey (Richard Bowden).

You will develop and implement novel computer vision and learning algorithms to automate recognition and detection in video. You will also be responsible for developing software infrastructure for viewing and searching the resulting classifications and detections.

The objective of the project is to develop automated tools that allow temporal visual content, such as a human gesturing, using sign language, or interacting with objects or other humans, to be learnt from standard TV broadcast signals or Kinect sensors using the transmitted annotation in the form of subtitles and annotation for the visually impaired as supervision. This requires the development of models of the visual appearance and dynamics of actions, and learning methods which can train such models using the weak supervision provided by the subtitles. Once the models have been learnt they can then be used without supervision, e.g. for sign language interpretation or automatic description of the content of video footage, and during the project demonstrators will be engineered for both of these applications.

Key tasks:

• Development and implementation of novel computer vision and learning algorithms to automate recognition and detection in video.
• Develop software infrastructure for viewing and searching the resulting classifications and detections.
• Writing up the results for conference and journal publications.
• Contributing documented software to the group library.
• Performing occasional demonstrations to industrial or academic visitors.
• Assisting informally with the supervision of graduate students conducting related research

Detailed tasks:

• Develop computer vision and learning algorithms:
• For recognizing British Sign Language using weakly aligned subtitles in broadcast video
• For recognizing human actions and activities from weakly aligned subtitles and transcripts in broadcast video and/or Kinect data
• For efficient learning/detection/tracking in video sequences
• Software development

Additional duties:

• Manage own academic research and administrative activities. This involves small scale project management, to co-ordinate multiple aspects of work to meet deadlines
• Contribute ideas for new research projects
• Develop ideas for generating research income, and present detailed research proposals to senior researchers
• Collaborate in the preparation of scientific reports and journal articles and occasionally present papers and posters 
• Act as a source of information and advice to other members of the group on scientific protocols and experimental techniques
• Represent the research group at external meetings/seminars, either with other members of the group or alone
• Carry out collaborative projects with colleagues in partner institutions, and research groups
• The PDRA may have the opportunity to teach (this includes lecturing, demonstrating, small-group teaching, tutoring of undergraduates and graduate students and supervision of masters projects in collaboration with principal investigators). 

To submit your application for this position, please click on the "Apply through website" button.

Only applications received before 12.00 midday on Friday 5 December 2014 can be considered. You will be required to upload a covering letter/supporting statement, including a brief statement of research interests (describing how past experience and future plans fit with the advertised position), CV and the details of two referees as part of your online application.

Informal enquiries may be addressed to Andrew Zisserman at andrew.zisserman@en.ox.ac.uk

Essential criteria:

• Doctorate, or near completion of doctorate, in Computer Vision or Machine Learning
• Background and expertise in recognition in images or videos
• Background and expertise in machine learning
• Strong publication record at principal computer vision conferences (CVPR, ECCV, ICCV) 
• Good written and spoken communication skills

Desirable criteria:

• Able to program in C++.
• Able to program in Matlab.
• Experience with UNIX.
• Experience with Windows.

Apply here: https://www.recruit.ox.ac.uk/pls/hrisliverecruit/erq_jobspec_version_4.display_form