high performance computing on graphics processing units: hgpu.org

In this thesis a superresolution method for geometry reconstruction is presented. Superresolution methods have already been used in order to improve texture resolution and RGB images. In a lesser degree they have also been used to improve the quality of depth maps. The objective of this thesis is to validate the hypothesis that superresolution methods can also be used in the geometry reconstruction. In order to do that an algorithm has been implemented and a superresolution operator has been derived and plugged in the algorithm. Results show that the algorithm produces better quality reconstructions than the original approach without superresolution. Not only upsampling but also deblurring help to obtain better reconstructions. With our method smoother results are obtained without losing geometric details.