Bücher der Reihe Aktuelle Forschung Medizintechnik - Latest Research in Medical Engineering

Segmentation of anatomical structures in medical image data is an essential task in clinical practice. Dagmar Kainmueller introduces methods for accurate fully automatic segmentation of anatomical structures in 3D medical image data. The author's core methodological contribution is a novel deformation model that overcomes limitations of state-of-the-art Deformable Surface approaches, hence allowing for accurate segmentation of tip- and ridge-shaped features of anatomical structures. As for practical contributions, she proposes application-specific segmentation pipelines for a range of anatomical structures, together with thorough evaluations of segmentation accuracy on clinical image data. As compared to related work, these fully automatic pipelines allow for highly accurate segmentation of benchmark image data.

Holoscopy is a new tomographic imaging modality that combines techniques of digital holography with Fourier-domain optical coherence tomography (FD-OCT). Dierck Hillmann gives a theoretical introduction to the mathematics and physics of holoscopy and develops an efficient numerical reconstruction procedure. Compared to FD-OCT, holoscopy provides unique advantages by enabling tomographic imaging without a limited depth of focus, but results in an increased numerical cost for reconstruction. In further chapters, the author introduces techniques for FD-OCT that are relevant to holoscopy as well. He demonstrates and compares numerical reconstruction methods for FD-OCT and shows how motion and dispersion artifacts in FD-OCT can be numerically compensated.

Yulia Levakhina gives an introduction to the major challenges of image reconstruction in Digital Tomosynthesis (DT), particularly to the connection of the reconstruction problem with the incompleteness of the DT dataset. The author discusses the factors which cause the formation of limited angle artifacts and proposes how to account for them in order to improve image quality and axial resolution of modern DT. The addressed methods include a weighted non-linear back projection scheme for algebraic reconstruction and novel dual-axis acquisition geometry. All discussed algorithms and methods are supplemented by detailed illustrations, hints for practical implementation, pseudo-code, simulation results and real patient case examples.

Marlitt Erbe provides a detailed introduction into the young research field of Magnetic Particle Imaging (MPI) and field free line (FFL) imaging in particular. She derives a mathematical description of magnetic field generation for FFL imaging in MPI. To substantiate the simulation studies on magnetic FFL generation with a proof-of-concept, the author introduces the FFL field demonstrator, which provides the world's first experimentally generated rotated and translated magnetic FFL field complying with the requirements for FFL reconstruction. Furthermore, she proposes a scanner design of considerably enhanced magnetic field quality and efficiency. The author discusses the influence of magnetic field quality optimization on the image quality achieved using efficient Radon-based reconstruction methods, which arise for a line detection scheme and based on this optimized design, presents a dynamic FFL scanner assembly.

Bernhard Gleich introduces Magnetic Particle Imaging (MPI) including all aspects that are necessary for a complete understanding. MPI is a new imaging modality invented by Bernhard Gleich and Jurgen Weizenecker. The method is capable of imaging the distribution of superparamagnetic iron oxide particles (SPIOs) with high sensitivity, high spatial resolution, and high imaging speed. The author summarizes the results of a number of original papers and countless innovations he has elaborated in the young discipline of magnetic particle imaging.

Various applications in the field of pulmonary image analysis require a registration of CT images of the lung. For example, a registration-based estimation of the breathing motion is employed to increase the accuracy of dose distribution in radiotherapy. Alexander Schmidt-Richberg develops methods to explicitly model morphological and physiological knowledge about respiration in algorithms for the registration of thoracic CT images. The author focusses on two lung-specific issues: on the one hand, the alignment of the interlobular fissures and on the other hand, the estimation of sliding motion at the lung boundaries. He shows that by explicitly considering these aspects based on a segmentation of the respective structure, registration accuracy can be significantly improved.