@inbook{
author = "Peulić, Aleksandar S. and Milanković, Ivan and Mijailović, Nikola V. and Filipović, Nenad",
year = "2019",
abstract = "Image segmentation is one of the most common procedures in medical imaging applications. It is also very important task in breast cancer detection. Breast cancer detection procedure based on mammography can be divided into several stages. The first stage is the extraction of region of interest from a breast image, after which the identification of suspicious mass regions, their classification, and comparison with the existing image database follows. It is often the case that already existing image databases have large sets of data for which processing requires a lot of time, thus the acceleration of each of the processing stages in breast cancer detection is a very important issue. Image filtering is also one of the most common and important tasks in image processing applications. It is, in most cases, preprocessing procedure for 3D visualization of an image stack. In order to achieve high-quality 3D visualization of a 2D image stack, it is of particular importance that all the images of the input stack are clear and sharp, thus their filtering should be executed carefully. There are also many algorithms for 3D visualization, so it is important to choose the right one which will execute fast enough and produce satisfied quality. In this chapter, the implementation of the already existing algorithm for region-of-interest-based image segmentation for mammogram images on High-Performance Reconfigurable DataFlow Computers (HPRDC) is proposed. As a DataFlow Engine (DFE) of such HPRDC, Maxeler’s acceleration card is used. The experiments for examining the acceleration of that algorithm on the Reconfigurable DataFlow Computers (RDC) are performed with two types of mammogram images with different resolutions. There were, also, several DFE configurations and each of them gave different acceleration of algorithm execution. Those accelerations are presented and experimental results have been shown good acceleration. Also, image processing using a mean filtering algorithm combined with thresholding and binarization algorithms and 3D visualization of murine lungs using marching cubes method are explained. These algorithms are mapped on the Maxeler’s DFE to significantly increase calculation speed. Optimal algorithm calculation speed was up to 20-fold baseline calculation speed.",
publisher = "Springer Cham",
journal = "Exploring the DataFlow Supercomputing Paradigm: Example Algorithms for Selected Applications, Part of the book series: Computer Communications and Networks (CCN)",
booktitle = "Biomedical Images Processing Using Maxeler DataFlow Engines",
pages = "197-227",
doi = "10.1007/978-3-030-13803-5_7"
}
