PamiTC Job Board - Posting Details

Title: 2+ PostDocs in Vision/GraphicsPosted: July 10, 2013
Company/Institution: University College London (UCL)
Location: London, UK
Department: Computer Science


UCL Vision/Graphics: 3D + Recognition + Synthesis

Acquiring Complete and Editable Outdoor Models from Video and Images
Larger Project: The overall objective of this multi-team project is to develop interactive computer graphics and vision technology that leverages machine learning and procedural modeling, with possible applications with our special effects partners. There are two 3-year postdoc positions at UCL, and four more at the University of Bath. The UCL posts are with Gabriel Brostow and Tim Weyrich respectively - see below to apply. The University of Bath-based posts are here:, with group leaders Peter Hall, Matt Brown, Darren Cosker, and Phil Willis + Brian Wyvill.
Each postdoc will be responsible for carrying out research in Computer Vision and Graphics generally, and on the smart capture, synthesis and editing of real-world shapes and phenomena.

The postdocs are for at least 24 months (maximum 36), to start September 2013 (or very nearly).

At UCL, each postdoc will focus on either

Landscapes (Weyrich):

Aims include extracting and learning characteristic properties of real landscapes from ground-based and aerial images, devising a procedural landscape generation system that creates virtual landscapes and ecosystems that mimic the acquired landscape. A key challenge will be how to maintain direct editability of the procedurally generated output, for use in a content creation pipeline. The work will straddle aspects of computer vision, computer graphics and machine learning.


Smart Capture (Brostow):

Aims include tools for creative users to quickly and accurately capture surface detail on indoor/outdoor "sets," and to synthesize/complete not just texture, but shapes. Our new methods will learn from experience, ask the user questions, and give feedback: What is this surface? Does it look like +this+ from the other side? Can you move the camera closer/further away? These are hard problems, made tractable (but harder in their own right) by having users in the semi-supervised loop.

Programming experience desired: high proficiency in one or more of Matlab / C++ / Python
Experience in the following areas is a plus:
- 3D Acquisition
- Camera Calibration
- Machine Learning
- Texture Synthesis
- Procedural Modeling
- Content Creation / Visual Effects
- User Interface Development (e.g. Qt)

Publication Record: Previous research should appear as publications in CVPR, ECCV, SIGGRAPH, PAMI, CHI, NIPS.

About us: UCL is in central London, and is one of the top 3 groups for Vision/Graphics/Learning in Europe. Our people are fun and productive. Enough said :)

Application Instructions:
The projected closing date for applications to receive full consideration is 1st August 2013. There is one official application process at UCL for both our posts. You must apply through the official web-page ( to be considered. Emailing me does NOT count.

In the meantime, you can get in touch with myself (see my homepage here) or Tim Weyrich to express your interest. Informal enquiries about the post are welcome. Please use the word "postdocOAK" in the Subject when contacting us.

The application page is ready, with vacancy reference number 1346587, accessible via where you should search for Ref No 1346587, or by searching for job titles with "Research Associate in Computer Graphics and Vision."