Dieter A. Hahn

Segmentation of kidneys using a new active shape model generation technique based on non-rigid image registration

Active shape models (ASMs) are widely used for applications in the field of image segmentation. Building an ASM requires to determine point correspondences for input training data, which usually results in a set of landmarks distributed according to the statistical variations. State-of-the-art methods solve this problem by minimizing the description length of all landmarks using a parametric mapping of the target shape (e.g. a sphere). In case of models composed of multiple sub-parts or highly non-convex shapes, these techniques feature substantial drawbacks. This article proposes a novel technique for solving the crucial correspondence problem using non-rigid image registration. Unlike existing approaches the new method yields more detailed ASMs and does not require explicit or parametric formulations of the problem. Compared to other methods, the already built ASM can be updated with additional prior knowledge in a very efficient manner. For this work, a training set of 3-D kidney pairs has been manually segmented from 41 CT images of different patients and forms the basis for a clinical evaluation. The novel registration based approach is compared to an already established algorithm that uses a minimum description length (MDL) formulation. The presented results indicate that the use of non-rigid image registration to solve the point correspondence problem leads to improved ASMs and more accurate segmentation results. The sensitivity could be increased by approximately 10%. Experiments to analyze the dependency on the user initialization also show a higher sensitivity of 5-15%. The mean squared error of the segmentation results and the ground truth manually classified data could also be reduced by 20-34% with respect to varying numbers of training samples.

Automatic Parameter Selection for Multimodal Image Registration

Over the past ten years similarity measures based on intensity distributions have become state-of-the-art in automatic multimodal image registration. An implementation for clinical usage has to support a plurality of images. However, a generally applicable parameter configuration for the number and sizes of histogram bins, optimal Parzen-window kernel widths or background thresholds cannot be found. This explains why various research groups present partly contradictory empirical proposals for these parameters. This paper proposes a set of data-driven estimation schemes for a parameter-free implementation that eliminates major caveats of heuristic trial and error. We present the following novel approaches: a new coincidence weighting scheme to reduce the influence of background noise on the similarity measure in combination with Max-Lloyd requantization, and a tradeoff for the automatic estimation of the number of histogram bins. These methods have been integrated into a state-of-the-art rigid registration that is based on normalized mutual information and applied to CT-MR, PET-MR, and MR-MR image pairs of the RIRE 2.0 database. We compare combinations of the proposed techniques to a standard implementation using default parameters, which can be found in the literature, and to a manual registration by a medical expert. Additionally, we analyze the effects of various histogram sizes, sampling rates, and error thresholds for the number of histogram bins. The comparison of the parameter selection techniques yields 25 approaches in total, with 114 registrations each. The number of bins has no significant influence on the proposed implementation that performs better than both the manual and the standard method in terms of acceptance rates and target registration error (TRE). The overall mean TRE is 2.34 mm compared to 2.54 mm for the manual registration and 6.48 mm for a standard implementation. Our results show a significant TRE reduction for distortion-corrected magnetic resonance images.

Quantification of Thyroid Volume Using 3-D Ultrasound Imaging

Ultrasound (US) is among the most popular diagnostic techniques today. It is non-invasive, fast, comparably cheap, and does not require ionizing radiation. US is commonly used to examine the size, and structure of the thyroid gland. In clinical routine, thyroid imaging is usually performed by means of 2-D US. Conventional approaches for measuring the volume of the thyroid gland or its nodules may therefore be inaccurate due to the lack of 3-D information. This work reports a semi-automatic segmentation approach for the classification, and analysis of the thyroid gland based on 3-D US data. The images are scanned in 3-D, pre-processed, and segmented. Several pre-processing methods, and an extension of a commonly used geodesic active contour level set formulation are discussed in detail. The results obtained by this approach are compared to manual interactive segmentations by a medical expert in five representative patients. Our work proposes a novel framework for the volumetric quantification of thyroid gland lobes, which may also be expanded to other parenchymatous organs.

Towards Real-time Guidewire Detection and Tracking in the Field of Neuroradiology

Two-dimensional roadmapping is considered state-of-the-art in guidewire navigation during endovascular interventions. This paper presents a methodology for extracting the guidewire from a sequence of 2-D roadmap images in almost real time. The detected guidewire can be used to improve its visibility on noisy fluoroscopic images or to do a back projection of the guidewire into a registered 3-D vessel tree. A lineness filter based on the Hessian matrix is used to detect only those line structures in the image that lie within the vessel tree. Loose wire fragments are properly linked by a novel connection method fulfilling clinical processing requirements. We show that Dijkstras algorithm can be applied to efficiently compute the optimal connection path. The entire guidewire is finally approximated by a B-spline curve in a least-squares manner. The proposed method is both integrated into a commercial clinical prototype and evaluated on five different patient data sets containing up to 249 frames per image series.

Unbiased rigid registration using transfer functions

The evaluation of tumor growth as regression under therapy is an important clinical issue. Rigid registration of sequentially acquired 3D-images has proven its value for this purpose. Existing approaches to rigid image registration use the whole volume for the estimation of the rigid transform. Non-rigid soft tissue deformation, however, will imply a bias to the registration result, because local deformations cannot be modeled by rigid transforms. Anatomical substructures, like bones or teeth, are not affected by these deformations, but follow a rigid transform. This important observation is incorporated in the proposed registration algorithm. The selection of anatomical substructure is done by manual interaction of medical experts adjusting the transfer function of the volume rendering software. The parameters of the transfer function are used to identify the voxels that are considered for registration. A rigid transform is estimated by a quaternion gradient descent algorithm based on the intensity values of the specified tissue classes. Commonly used voxel intensity measures are adjusted to the modified registration algorithm. The contribution describes the mathematical framework of the proposed registration method and its implementation in a commercial software package. The experimental evaluation includes the discussion of different similarity measures, the comparison of the proposed method to established rigid registration techniques and the evaluation of the efficiency of the new method. We conclude with the discussion of potential medical applications of the proposed registration algorithm.

A practical salient region feature based 3D multi-modality registration method for medical images

We present a novel representation of 3D salient region features and its integration into a hybrid rigid-body registration framework. We adopt scale, translation and rotation invariance properties of those intrinsic 3D features to estimate a transform between underlying mono- or multi-modal 3D medical images. Our method combines advantageous aspects of both feature- and intensity-based approaches and consists of three steps: an automatic extraction of a set of 3D salient region features on each image, a robust estimation of correspondences and their sub-pixel accurate refinement with outliers elimination. We propose a region-growing based approach for the extraction of 3D salient region features, a solution to the problem of feature clustering and a reduction of the correspondence search space complexity. Results of the developed algorithm are presented for both mono- and multi-modal intra-patient 3D image pairs (CT, PET and SPECT) that have been acquired for change detection, tumor localization, and time based intra-person studies. The accuracy of the method is clinically evaluated by a medical expert with an approach that measures the distance between a set of selected corresponding points consisting of both anatomical and functional structures or lesion sites. This demonstrates the robustness of the proposed method to image overlap, missing information and artefacts. We conclude by discussing potential medical applications and possibilities for integration into a non-rigid registration framework.

3D annotation and manipulation of medical anatomical structures

Although the medical scanners are rapidly moving towards a three-dimensional paradigm, the manipulation and annotation/labeling of the acquired data is still performed in a standard 2D environment. Editing and annotation of three-dimensional medical structures is currently a complex task and rather time-consuming, as it is carried out in 2D projections of the original object. A major problem in 2D annotation is the depth ambiguity, which requires 3D landmarks to be identified and localized in at least two of the cutting planes. Operating directly in a three-dimensional space enables the implicit consideration of the full 3D local context, which significantly increases accuracy and speed. A three-dimensional environment is as well more natural optimizing the users comfort and acceptance. The 3D annotation environment requires the three-dimensional manipulation device and display. By means of two novel and advanced technologies, Wii Nintendo Controller and Philips 3D WoWvx display, we define an appropriate 3D annotation tool and a suitable 3D visualization monitor. We define non-coplanar setting of four Infrared LEDs with a known and exact position, which are tracked by the Wii and from which we compute the pose of the device by applying a standard pose estimation algorithm. The novel 3D renderer developed by Philips uses either the Z-value of a 3D volume, or it computes the depth information out of a 2D image, to provide a real 3D experience without having some special glasses. Within this paper we present a new framework for manipulation and annotation of medical landmarks directly in three-dimensional volume.

Difference imaging of inter- and intra-ictal SPECT images for the localization of seizure onset in epilepsy

The comparison of inter- with intra-ictal SPECT images plays an important role during the diagnosis and treatment of epilepsy patients. Although there is already commercial software available to address this problem using complex clinical workflows, this article describes a different way of looking at this issue. During the examination various issues arise from differing tracer concentrations, patient movement between the acquisitions at different times and also the lack of morphological information. The goal of the presented work is therefore to present an approach that is on the one hand easy to use for the physician and on the other hand both reliable and robust enough to cope with the previously mentioned challenges. The proposed algorithm introduces methods that have already been applied successfully in digital subtraction angiography (DSA). The work comprises of several steps for the intensity normalization, image registration, difference imaging and the incorporation of an MR image for the spatial localization. As a result, information is provided about differences within the cerebral blood flow (CBF) and active brain areas between the intra- and inter-ictal states. Very new to the field of SPECT brain imaging is the application of non-rigid registration techniques. This helps to drastically reduce the artifacts within the difference images due to a bias of the standard rigid registration. Acquired results from a collective of 11 patients show that this additional feature helps to further improve the image quality.

A nonlinear projection scheme for fast rigid registration

Intensity based rigid registration algorithms commonly employed for medical image fusion are based on the iterative optimization of a pixel-by-pixel distance measure defined on the images. As medical images grow larger in size due to advanced scanner technology, evaluating such similarity measures is no longer computationally efficient. In order to overcome the inherent limitations of the standard approach we propose a new, nonlinear projection scheme that enables a very fast evaluation of the distance between two images. Current state-of-the-art projection schemes decompose a six dimensional search space into three dimensional subspaces. The proposed approach, however, yields a complete decomposition into ID subspaces. The optimization on these subspaces is highly efficient and does not require a reprojection. This scheme is therefore suitable for 2D and 3D registrations, and it is able to cope with subvolume matching problems. Furthermore, the use of modern graphics hardware allows for a highly efficient implementation. Experiments show that computation times can be reduced to less than 10 seconds with the proposed approach for 2563 sized volumes.

Regression Forest-Based Organ Detection in Normalized PET Images

The detection of organs from full-body PET images is a chal- lenging task due to the high noise and the limited amount of anatomical information of PET imaging. The knowledge of organ locations can sup- port many clinical applications like image registration or tumor detec- tion. This paper is the first to propose an organ localization framework tailored on the challenges of PET. The algorithm involves intensity nor- malization, feature extraction and regression forests. Linear and non- linear intensity normalization methods are compared theoretically and experimentally. From the normalized images, long-range spatial con- text visual features are extracted. A regression forest predicts the organ bounding boxes. Experiments show that percentile normalization is the best preprocessing method. The algorithm is evaluated on 25 clinical images with a spatial resolution of 5 mm. With 13.8 mm mean absolute bounding box error, it achieves state-of-the-art results.