Tetrahedral Volume Reconstruction in X-Ray Tomography using GPU Architecture

Michele A. Quinto (1) , Dominique Houzet (2) , Fanny Buyens (1)

1 LIST (CEA) - Laboratoire d'Intégration des Systèmes et des Technologies
2 GIPSA-AGPIG - GIPSA - Architecture, Géométrie, Perception, Images, Gestes

In this paper, we propose the use of the graphics processor unit (GPU) to accelerate a ray-tracing method in the framework of X-ray tomographic image reconstruction. We first describe an innovative iterative reconstruction method we have developed based on a tetrahedral volume with conjugate gradient. We do not use voxels here but instead tetrahedrons to increase the quality of reconstruction and the reduction of data as thus we need less resolution of the volume to fit the object reconstructed. This is an important point to use the GPU. We present here the algorithms adapted to the GPU and the results obtained compared to CPU.

Domaines: Architectures Matérielles [cs.AR]; Calcul parallèle, distribué et partagé [cs.DC]

Michele A. Quinto, Dominique Houzet, Fanny Buyens. Tetrahedral Volume Reconstruction in X-Ray Tomography using GPU Architecture. DASIP 2013 - Conference on Design and Architectures for Signal and Image Processing, Oct 2013, Cagliari, Italy. pp.334-339. ⟨hal-00938933⟩