Detail

Tomo-GPU v1: an enviroment dedicated to reinforcement characterization in composite materials

In the scope of Tomo-GPU FCT/MCTES-funded project, we developed the first version of a problem solving environment (PSE) dedicated to the characterization of reinforcement population in composite materials. The PSE is based on the SCIRun framework and includes, besides visualization facilities, several modules that perform highly-demanding computations related with the processing of 3D tomographic images, namely:
– image cleaning, including segmentation, hysteresis and percolation: besides the author prototypes below, this includes work by Magda Encarnação and Bruno Preto
– image labeling. This is work developed by Bruno Preto in MSc
thesis and corresponds to the identification of 3D objects in a black and white 3D tomographic image
The high-demanding computational processing in the above modules uses GPGPUs; the software developed is based on OpenCL