During the past decade, computer-aided image processing has evolved rapidly due to modern processors and graphics processing units. Computer imaging got a major role in several fields of clinical diagnostics. Due to the increased processing power, the development of new, more powerful algorithms has now become possible, that will make clinical diagnostics a lot easier and will lead to a significant speed up as well. The goal of my thesis is to provide the researchers of this field with a tool that can help them run, benchmark the speed, and compare the output of medical imaging methods with ease, helping them develop faster and more efficient algorithms.
The completed tasks and the problems I solved are described in the following subsections of this thesis: first, I provide a short summary of the existing tools with similar functionality, then I present the framework I designed and implemented, and explain the engineering decisions behind it. Next, I provide a brief summary of the magnetic resonance imaging and its characteristic noise, and present a few algorithms that can compensate it. Then I design and implement a graphical user interface, that provides a user-friendly way to run benchmarks. Finally, through an example, I explain the usage and significance of my work.
Thanks to the modularity and the flexibility of the environment, extending it with new features and algorithms is easy and hopefully with the help of this tool, faster and more efficient algorithms will be published in the near future.