Ray casting de volum: CPU vs GPU

Abstract

The motivation of this project is to provide interactive displays of volumetric data captured from medical devices. In this project we analyze two strategies based on Ray Casting algorithm: the classic approach in the CPU and optimized approach on the GPU in order to study and analyze the various costs both in computational times. We propose a validation system that displays the results of both the CPU and the GPU so that they can be compared visually and computationally. Moreover, we want to go further and show the results in different ways, such as by using Look up Tables, the use of models with display lighting in the scene, such as Phong Blinn methods or other kind of projection like MIP (Maximum Intensity Projection). This involves the analysis, design and implementation of application able to visualize volumetric data in space, bearing in mind that we want to compare results of the CPU with the GPU.