Daniel Forsberg

Medical Image Processing on the GPU - Past, Present and Future

Graphics processing units (GPUs) are used today in a wide range of applications, mainly because they can dramatically accelerate parallel computing, are affordable and energy efficient. In the field of medical imaging, GPUs are in some cases crucial for enabling practical use of computationally demanding algorithms. This review presents the past and present work on GPU accelerated medical image processing, and is meant to serve as an overview and introduction to existing GPU implementations. The review covers GPU acceleration of basic image processing operations (filtering, interpolation, histogram estimation and distance transforms), the most commonly used algorithms in medical imaging (image registration, image segmentation and image denoising) and algorithms that are specific to individual modalities (CT, PET, SPECT, MRI, fMRI, DTI, ultrasound, optical imaging and microscopy). The review ends by highlighting some future possibilities and challenges.

Generating patient specific pseudo-CT of the head from MR using atlas-based regression

Radiotherapy planning and attenuation correction of PET images require simulation of radiation transport. The necessary physical properties are typically derived from computed tomography (CT) images, but in some cases, including stereotactic neurosurgery and combined PET/MR imaging, only magnetic resonance (MR) images are available. With these applications in mind, we describe how a realistic, patient-specific, pseudo-CT of the head can be derived from anatomical MR images. We refer to the method as atlas-based regression, because of its similarity to atlas-based segmentation. Given a target MR and an atlas database comprising MR and CT pairs, atlas-based regression works by registering each atlas MR to the target MR, applying the resulting displacement fields to the corresponding atlas CTs and, finally, fusing the deformed atlas CTs into a single pseudo-CT. We use a deformable registration algorithm known as the Morphon and augment it with a certainty mask that allows a tailoring of the influence certain regions are allowed to have on the registration. Moreover, we propose a novel method of fusion, wherein the collection of deformed CTs is iteratively registered to their joint mean and find that the resulting mean CT becomes more similar to the target CT. However, the voxelwise median provided even better results; at least as good as earlier work that required special MR imaging techniques. This makes atlas-based regression a good candidate for clinical use.

A multi-center milestone study of clinical vertebral CT segmentation☆

A multiple center milestone study of clinical vertebra segmentation is presented in this paper. Vertebra segmentation is a fundamental step for spinal image analysis and intervention. The first half of the study was conducted in the spine segmentation challenge in 2014 International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI) Workshop on Computational Spine Imaging (CSI 2014). The objective was to evaluate the performance of several state-of-the-art vertebra segmentation algorithms on computed tomography (CT) scans using ten training and five testing dataset, all healthy cases; the second half of the study was conducted after the challenge, where additional 5 abnormal cases are used for testing to evaluate the performance under abnormal cases. Dice coefficients and absolute surface distances were used as evaluation metrics. Segmentation of each vertebra as a single geometric unit, as well as separate segmentation of vertebra substructures, was evaluated. Five teams participated in the comparative study. The top performers in the study achieved Dice coefficient of 0.93 in the upper thoracic, 0.95 in the lower thoracic and 0.96 in the lumbar spine for healthy cases, and 0.88 in the upper thoracic, 0.89 in the lower thoracic and 0.92 in the lumbar spine for osteoporotic and fractured cases. The strengths and weaknesses of each method as well as future suggestion for improvement are discussed. This is the first multi-center comparative study for vertebra segmentation methods, which will provide an up-to-date performance milestone for the fast growing spinal image analysis and intervention.

Fully automatic measurements of axial vertebral rotation for assessment of spinal deformity in idiopathic scoliosis

Reliable measurements of spinal deformities in idiopathic scoliosis are vital, since they are used for assessing the degree of scoliosis, deciding upon treatment and monitoring the progression of the disease. However, commonly used two dimensional methods (e.g. the Cobb angle) do not fully capture the three dimensional deformity at hand in scoliosis, of which axial vertebral rotation (AVR) is considered to be of great importance. There are manual methods for measuring the AVR, but they are often time-consuming and related with a high intra- and inter-observer variability. In this paper, we present a fully automatic method for estimating the AVR in images from computed tomography. The proposed method is evaluated on four scoliotic patients with 17 vertebrae each and compared with manual measurements performed by three observers using the standard method by Aaro-Dahlborn. The comparison shows that the difference in measured AVR between automatic and manual measurements are on the same level as the inter-observer difference. This is further supported by a high intraclass correlation coefficient (0.971-0.979), obtained when comparing the automatic measurements with the manual measurements of each observer. Hence, the provided results and the computational performance, only requiring approximately 10 to 15 s for processing an entire volume, demonstrate the potential clinical value of the proposed method.

Evaluation and comparison of 3D intervertebral disc localization and segmentation methods for 3D T2 MR data: A grand challenge.

&NA; The evaluation of changes in Intervertebral Discs (IVDs) with 3D Magnetic Resonance (MR) Imaging (MRI) can be of interest for many clinical applications. This paper presents the evaluation of both IVD localization and IVD segmentation methods submitted to the Automatic 3D MRI IVD Localization and Segmentation challenge, held at the 2015 International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI2015) with an on‐site competition. With the construction of a manually annotated reference data set composed of 25 3D T2‐weighted MR images acquired from two different studies and the establishment of a standard validation framework, quantitative evaluation was performed to compare the results of methods submitted to the challenge. Experimental results show that overall the best localization method achieves a mean localization distance of 0.8 mm and the best segmentation method achieves a mean Dice of 91.8%, a mean average absolute distance of 1.1 mm and a mean Hausdorff distance of 4.3 mm, respectively. The strengths and drawbacks of each method are discussed, which provides insights into the performance of different IVD localization and segmentation methods. HighlightsEstablish a standard framework with 25 manually annotated 3D T2 MRI data for an objective comparison of intervertebral disc (IVD) localization and segmentation methods.Investigate strengths and limitations of a representative selection of the state‐of‐the‐art IVD localization and segmentation methods with a challenge setup.Results achieved by the best algorithms in this study set new frontiers for IVD localization and segmentation from MR data. Graphical abstract Figure. No caption available.

Atlas-Based Registration for Accurate Segmentation of Thoracic and Lumbar Vertebrae in CT Data

Segmentation of the vertebrae in the spine is of relevance to many medical applications related to the spine. This paper describes a method based upon atlas-based registration for achieving an accurate segmentation of the thoracic and the lumbar vertebrae in the spine as imaged by computed tomography. The method has been evaluated on ten data sets provided as a part of the segmentation challenge hosted by the 2nd MICCAI workshop on Computational Methods and Clinical Applications for Spine Imaging. An average point-to-surface error of \(1.05\,\pm \,0.65\) mm and a mean DICE coefficient of \(0.94\,\pm \,0.03\) were obtained when comparing the computed segmentations with ground truth segmentations. These results are highly competitive when compared to the results of earlier presented methods.

Model-based registration for assessment of spinal deformities in idiopathic scoliosis

Detailed analysis of spinal deformity is important within orthopaedic healthcare, in particular for assessment of idiopathic scoliosis. This paper addresses this challenge by proposing an image analysis method, capable of providing a full three-dimensional spine characterization. The proposed method is based on the registration of a highly detailed spine model to image data from computed tomography. The registration process provides an accurate segmentation of each individual vertebra and the ability to derive various measures describing the spinal deformity. The derived measures are estimated from landmarks attached to the spine model and transferred to the patient data according to the registration result. Evaluation of the method provides an average point-to-surface error of 0.9 mm ± 0.9 (comparing segmentations), and an average target registration error of 2.3 mm ± 1.7 (comparing landmarks). Comparing automatic and manual measurements of axial vertebral rotation provides a mean absolute difference of 2.5° ± 1.8, which is on a par with other computerized methods for assessing axial vertebral rotation. A significant advantage of our method, compared to other computerized methods for rotational measurements, is that it does not rely on vertebral symmetry for computing the rotational measures. The proposed method is fully automatic and computationally efficient, only requiring three to four minutes to process an entire image volume covering vertebrae L5 to T1. Given the use of landmarks, the method can be readily adapted to estimate other measures describing a spinal deformity by changing the set of employed landmarks. In addition, the method has the potential to be utilized for accurate segmentations of the vertebrae in routine computed tomography examinations, given the relatively low point-to-surface error.

Atlas-Based Segmentation of the Thoracic and Lumbar Vertebrae

Segmentation of the vertebrae in the spine is of relevance to many medical applications. To this end, the 2nd MICCAI workshop on Computational Methods and Clinical Applications for Spine Imaging organized a segmentation challenge. This paper briefly presents one of the participating methods along with achieved results. The employed method is based upon atlas-based segmentation, where a number of atlases of the spine are registered to the target data set. The labels of the deformed atlases are combined using label fusion to obtain the final segmentation of the target data set. An average DICE score of \( 0.94 \pm 0.03 \) was achieved on the training data set.

Adaptive anisotropic regularization of deformation fields for non-rigid registration using the morphon framework

Image registration is a crucial task in many applications and applied in a variety of different areas. In addition to the primary task of image alignment, the deformation field is valuable when studying structural/volumetric changes in the brain. In most applications a regularizing term is added to achieve a smoothly varying deformation field. This can sometimes cause conflicts in situations of local complex deformations. In this paper we present a new regularizer, which aims at handling local complex deformations while maintaining an overall smooth deformation field. It is based on an adaptive anisotropic regularizer and its usefulness is demonstrated by two examples, one synthetic and one with real MRI data from a pre- and post-op situation with normal pressure hydrocephalus.

Improving registration using multi-channel diffeomorphic demons combined with certainty maps

The number of available imaging modalities increases both in clinical practice and in clinical studies. Even though data from multiple modalities might be available, image registration is typically only performed using data from a single modality. In this paper, we propose using certainty maps together with multi-channel diffeomorphic demons in order to improve both accuracy and robustness when performing image registration. The proposed method is evaluated using DTI data, multiple region overlap measures and a fiber bundle similarity metric.