15 January 2010
The eScience Centre receives a grant for high productivity tools for Industrial Image analysis
Professor Brian Vinter at the Escience Centre, Copenhagen University has received a grant of DKK 5,4 million, which will open up for the employment of 3-4 new PhD students.
The project which has a total budget of DKK 7,8 million includes besides DIKU, Danish Meat Research Institute in Roskilde, Deformalyze and Institute of Computing Technology, Chinese Academy of Sciences.
The project will start up in February 2010 and will run until August 2013. The purpose of the project is to develop a high speed and high productivity tool in connection with industrial processing of meat production. The means is computer based image analysis, the driver is a parallel connected supercomputer, that can speed up the calculation processes and the overall purpose with the project is to reduce time consuming feed-back processes in the industrial meat production.
The project - an overview
Computer based image analysis is very popular in the industry and many companies use many resources in developing algorithms for the analysis and recognition of pictures in connection with production.
The development of these algorithms often takes place in an environment that gives high productivity. It is applied in a Matlab where the researcher after having worked out an idea can make a prototype only few days later and thereby tests if the idea works or not.
When one problem is solved the researcher goes on to the next problem and when the complete task is solved the company has a fully working prototype developed with a high productive tool. By now the researcher can convert the algorithms from the high productive version in to a high speed version that can meet the demands for real time you will experience during one production. This process takes long time, many months, even years and may result in a usable solution as it is not fast enough for its purpose.
The project aims at developing high productivity tools, which gives high performance so the researcher can avoid converting algorithms with the high speed tool. Currently, the project is facilitating a long and very costly conversion of a larger CT image analysis system for Danish Meat Research Institute.
For further information please contact Brian Vinter, firstname.lastname@example.org, phone: 35 32 14 27