How it works
Reading radiographs can be a repetitive task for the radiographer.
To address this problem, a CAD (Computer-Aided Detection) system was developed using GPGPU (General Purpose Graphics Processing Unit) image-processing techniques.
This system is a test bench for image analysis that assists a radiologist in the diagnosis of the tuberculosis structures in a chest radiograph, by using Local Binary Pattern texture classification methods.
The parallelisation potential of a GPU is exploited to accelerate the image-processing techniques utilised in the system, increasing patient throughput.
The winners of the South African leg of Microsoft's Imagine Cup will represent SA at the worldwide Microsoft Imagine Cup finals, in Sydney, Australia.
This is according to Microsoft, which adds that University of Johannesburg (UJ) students Joshua Leibstein, Michael Cilliers and their mentor, Duncan Coulter, from the Academy of Software Engineering and Computer Science at UJ, came first in the South African local finals, in December 2011, with their tuberculosis detection project.
Team Asclepius will be representing SA in the Software Design category and has also been nominated for the Health Awareness Award.
Team Asclepius' project involved detecting tuberculosis in chest radiographs using image-processing techniques.
According to Microsoft, the Imagine Cup aims at encouraging students around the world to apply themselves to technology innovations that can make a difference.
Currently in its 10th year, the Imagine Cup has seen a cumulative 1.65 million students from more than 190 countries participate in the competition.
Share
