2d3 is seeking to fill two computer vision scientist positions.  One position will be based out of its Oxford office and the other out of its Southern California office.  The US position requires US citizenship.

Please contact Ben Ochoa during CVPR.  He will arrange to meet excellent candidates during the conference.  Following is the full job description.

The Company

Since 1999, 2d3 has supplied a range of computer vision products, all based on ‘3D from the moving image’, mainly to the film and television production industries.  The company’s primary product, Boujou, the world’s leading automatic camera tracker, is now used by most film and video post-production companies to generate visual effects requiring accurate registration of real and virtual images. 

In 2006, 2d3 expanded its efforts into the aerial imaging market.  The group has developed a set of technologies covering a wide range of real-time and off-line computer vision capabilities for processing of aerial motion imagery. 

2d3 is part of OMG plc, a company with a 25 year history of developing, manufacturing, and selling products for 3-dimensional tracking in medical, industrial, government, and defense markets.  OMG exports over 80% of its sales with customers in over 50 countries.  Since 2001, OMG has been listed on the London Stock Exchange.

The Job

The new appointee will be a computer vision scientist with relevant research experience, gained in academic, public sector, or commercial environments.  He or she will join a team with similar experience and act as technical lead on one or more projects covering augmented reality, real time structure from motion, automatic object recognition, vision hybrid navigation and targeting, tracking and collision avoidance, and terrain and urban modelling.

 2d3 has close working relationships and a number of active collaborative projects with leading academic computer vision groups in UK, Europe, and US. Members of the company regularly attend and present at the major international computer vision conferences. Within the constraints of commercial and national security, 2d3 encourages the publication of peer-reviewed or invited papers and presentations.

2d3 uses Matlab for prototyping but the majority of 2d3’s extensive internal computer vision libraries are implemented in speed-optimised, multi-threaded C++.  Windows or Linux environments are used according to project and customer requirements.

 The key areas of R&D for 2d3 include, but are not limited to:

•  Image-based tracking (stabilization, SFM, SLAM, multi-sensor systems, etc)
•  Object detection, recognition, & classification (with/without tracking, single/multi-view)
•  Automatic mosaicing (using a range of matching methods)
•  Photometric optimisation & matching
•  Camera calibration & image rectification
• Novel view synthesis
• Super-resolution & compression artifact reduction
• Hyperspectral image fusion
• Terrain & townscape 3D modeling

In addition, if you have the following key skills, they will set you apart from other candidates:

• C and C++
• OpenGL
• XML
• Boost
• IPP

In addition to your skills, you will also be required to satisfy the following minimum requirements:

• Fluency in English, both written and spoken
• Doctoral degree  - Computer Science or Electrical Engineering with direct experience in applied Computer Vision techniques
• 5 years of full time, professional software development and engineering experience
• Must have been published in a generally accepted academic trade publication for novel application of vision science theory to a real world problem. 
• Excellent analytic problem solving and critical thinking skills