Obesity is an increasing problem in the western world and triggers diseases like cancer, type two diabetes, and
cardiovascular diseases. In recent years, magnetic resonance imaging (MRI) has become a clinically viable method
to measure the amount and distribution of adipose tissue (AT) in the body. However, analysis of MRI images
by manual segmentation is a tedious and time-consuming process. In this paper, we propose a semi-automatic
method to quantify the amount of different AT types from whole-body MRI data with less user interaction.
Initially, body fat is extracted by automatic thresholding. A statistical shape model of the abdomen is then
used to differentiate between subcutaneous and visceral AT. Finally, fat in the bone marrow is removed using
morphological operators. The proposed method was evaluated on 15 whole-body MRI images using manual
segmentation as ground truth for adipose tissue. The resulting overlap for total AT was 93.7% ± 5.5 with a
volumetric difference of 7.3% ± 6.4. Furthermore, we tested the robustness of the segmentation results with regard
to the initial, interactively defined position of the shape model. In conclusion, the developed method proved
suitable for the analysis of AT distribution from whole-body MRI data. For large studies, a fully automatic
version of the segmentation procedure is expected in the near future.
Due to noise and artifacts often encountered in medical images, segmenting objects in these is one of the most
challenging tasks in medical image analysis. Model-based approaches like statistical shape models (SSMs) incorporate
prior knowledge that supports object detection in case of in-complete evidence from image data. In this paper, we present
a method to increase information of the object's shape in problematic image areas by incorporating mutual shape
information from other entities in the image. This is done by using a common shape space of multiple objects as
additional restriction. Two different approaches to implement mutual shape information are presented. Evaluation was
performed on nine cardiac images by simultaneous segmentation of the epi- and endocardium of the left heart ventricle
using the proposed methods. The results show that the segmentation quality is improved with both methods. For the
better one, the average surface distance error is approx. 40% lower.
The Fontan operation is a surgical treatment for patients with severe congenital heart diseases, where a biventricular correction of the heart can't be achieved. In these cases, a uni-ventricular system is established. During the last step of surgery a tunnel segment is placed to connect the inferior caval vein directly with the pulmonary artery, bypassing the right atrium and ventricle. Thus, the existing ventricle works for the body circulation, while the venous blood is passively directed to the pulmonary arteries. Fontan tunnels can be placed intra- and extracardially. The location, length and shape of the tunnel must be planned accurately. Furthermore, if the tunnel is placed extracardially, it must be positioned between other anatomical structures without constraining them. We developed a software system to support planning of the tunnel location, shape, and size, making pre-operative preparation of the tunnel material possible. The system allows for interactive placement and adjustment of the tunnel, affords a three-dimensional visualization of the virtual Fontan tunnel inside the thorax, and provides a quantification of the length, circumferences and diameters of the tunnel segments. The visualization and quantification can be used to plan and prepare the tunnel material for surgery in order to reduce the intra-operative time and to improve the fit of the tunnel patch.
The Medical Imaging Interaction Toolkit (MITK) and the eXtensible Imaging Platform (XIP) both aim at
facilitating the development of medical imaging applications, but provide support on different levels. MITK
offers support from the toolkit level, whereas XIP comes with a visual programming environment.
XIP is strongly based on Open Inventor. Open Inventor with its scene graph-based rendering paradigm was
not specifically designed for medical imaging, but focuses on creating dedicated visualizations. MITK has a
visualization concept with a model-view-controller like design that assists in implementing multiple, consistent
views on the same data, which is typically required in medical imaging. In addition, MITK defines a unified means
of describing position, orientation, bounds, and (if required) local deformation of data and views, supporting
e.g. images acquired with gantry tilt and curved reformations. The actual rendering is largely delegated to the
Visualization Toolkit (VTK).
This paper presents an approach of how to integrate the visualization concept of MITK with XIP, especially
into the XIP-Builder. This is a first step of combining the advantages of both platforms. It enables experimenting
with algorithms in the XIP visual programming environment without requiring a detailed understanding of Open
Inventor. Using MITK-based add-ons to XIP, any number of data objects (images, surfaces, etc.) produced by
algorithms can simply be added to an MITK DataStorage object and rendered into any number of slice-based
(2D) or 3D views. Both MITK and XIP are open-source C++ platforms. The extensions presented in this paper
will be available from www.mitk.org.
Statistical shape models have become a fast and robust method for segmentation of anatomical structures in medical image volumes. In clinical practice, however, pathological cases and image artifacts can lead to local deviations of the detected contour from the true object boundary. These deviations have to be corrected manually. We present an intuitively applicable solution for surface interaction based on Gaussian deformation kernels. The method is evaluated by two radiological experts on segmentations of the liver in contrast-enhanced CT images and of the left heart ventricle (LV) in MRI data. For both applications, five datasets are segmented automatically using deformable shape models, and the resulting surfaces are corrected manually. The interactive correction step improves the average surface distance against ground truth from 2.43mm to 2.17mm for the liver, and from 2.71mm to 1.34mm for the LV. We expect this method to raise the acceptance of automatic segmentation methods in clinical application.