GPU Pro 5

5

VI

Non-separable 2D, 3D, and 4D

Filtering with CUDA

Anders Eklund and Paul Dufort

5.1 Introduction

Filtering is an important step in many image processing applications such as

image denoising (where the goal is to suppress noise, see Figure 5.1), image

registration (where the goal is to align two images or volumes, see Figure 5.2),

and image segmentation (where the goal is to extract certain parts of an image or

volume, see Figure 5.3). In medical imaging, the datasets generated are often 3D

or 4D and contain a large number of samples, making filtering a computationally

demanding operation. A high-resolution magnetic resonance (MR) scan of a

human head ...