Applications are invited for two Engineering Doctorates (EngD) in the Centre for Digital Entertainment (CDE), a collaboration between the University of Bath and Bournemouth University, with Venatrack http://www.venatrack.com.
Venatrack is the leading company in the development of 3D tracking systems for sport applications. It has successfully developed systems for tracking and 3D video viewing for the broadcasting and sport industry. Within the company, this research will be based in the video processing pipeline.
EPSRC EngD-studentships, currently £15,090 per year (for 4 years). PLUS £3000 per year top-up from sponsoring company. Tuition fees paid PLUS funding for conference travel and research materials available.
Please see http://www.digital-entertainment.org/ for more information.
PROJECT 1) Object Extraction and Identification for Real Time 3D Tracking - focus on researching novel object extraction and recognition strategies for football players in a real-time and muti-camera environment.
Video Object Detection and Identification is a branch of image processing that uses object's features and changes in images to extract and identify pixels that represent entities that can be used for 3D tracking or 3D viewing. Existing techniques on this area obtain a reasonable performance in simple controlled conditions or selected sets of images. However, in practice, there are many situations where the performance can be reduced significantly. This is particularly important in outdoor applications where lighting variations, noise and camera angles can drastically change.
KEY AIMS: research of new video techniques that can increase the reliability of object detection in difficult environments. In particular to reduce the extraction and identification problems caused by background noise and variable lighting conditions in football video sequences.
--------------------------------------------------------------------------------------
PROJECT 2) Multiple-View Video Processing for Real Time 3D Video
This project is part of the video processing pipeline and it will focus on researching multi-camera 3D object video for football players.
The development of high quality 3D video requires a precise delineation of the objects in the scene. However, in addition to noise and calibration precision, there are practical factors in large outdoor environments that limit the accuracy of the image processing results. Thus, the generation of high quality 3D video still relies on manual processing.
KEY AIMS: research image processing techniques that can improve the quality and processing speed of multiple images for 3D video generation. These images are obtained from strong calibrated cameras taken from widely different viewpoints. There is a particular interest on investigating temporal coherence and the use of player's models to assist the identification of limbs and silhouette details necessary to achieve an accurate delineation of football players.
REQUIREMENTS: First class or 2:1 degree in electrical engineering, physics, mathematics, computer science or a related subject. The ideal candidate should have a background and interest in image processing, signal processing, computer vision and pattern recognition (project 1) or 3D graphics modelling (project 2). A very strong interest in research and good written and oral communication skills are essential. A good knowledge of C/C++ is desirable.
To apply, please complete the application form on http://www.digital-entertainment.org/ by 31 October 2011.
-- Katja Haferburg Centre for Digital Entertainment (Bath branch) Department of Computer Science University of Bath BA2 7AY T: 01225 386465 E: k.haferburg@bath.ac.uk The Centre for Digital Entertainment http://www.digital-entertainment.org
Keine Kommentare:
Kommentar veröffentlichen