Design and Evaluation of Parallel Processing Techniques for 3D Liver Segmentation and Volume Rendering

Vol. 5 No. 4

Year: 2011

Issue: April-June

Title: Design and Evaluation of Parallel Processing Techniques for 3D Liver Segmentation and Volume Rendering 

Author Name: Mohammed Goryawala, Magno R. Guillen , Armando Barreto , Ruchir Bhatt , Anthony McGoron , Malek Adjouadi 

Synopsis: 

3D reconstruction is a task that exerts a heavy computational load. This study describes the performance results obtained on testing MatLab parallel processing toolbox to execute a three-dimensional (3-D) liver reconstruction. The FFT algorithm was tested using the parallel tool box, by changing system platform, number of workers,image size and number of images. A second set was executed keeping the hardware fixed and changing the operating system to obtain unbiased results. The third experiment set was to assess the effect of parallelization applied to a newly developed 3-D liver reconstruction algorithm. Results showed a reduction of processing time from 4.5 hours to almost 1 hour, yielding a 78% (3.5 hours) saving in computational time due to the multicore deployment. The results show that the leveraging of multicore platforms can speed up considerably the processing of medical images through the use of parallel computing tools in MatLab.