Deformable template models are valuable tools in medicalimage segmentation. Current methods elegantly incorporate global shape and appearance, but can not cope with localized appearance variations and rely on an assu...
详细信息
ISBN:
(纸本)3540229760
Deformable template models are valuable tools in medicalimage segmentation. Current methods elegantly incorporate global shape and appearance, but can not cope with localized appearance variations and rely on an assumption of Gaussian gray value distribution. Furthermore, initialization near the optimal solution is required. We propose a maximum likelihood shape inference that is based on pixel classification, so that local and non-linear intensity variations are dealt with naturally, while a global shape model ensures a consistent segmentation. Optimization by stochastic sampling removes the need for accurate initialization. The method is demonstrated on three different medicalimage segmentation problems: vertebra segmentation in spine radiographs, lung field segmentation in thorax X rays, and delineation of the myocardium of the left ventricle in MRI slices. Accurate results were obtained in all tasks.
The proceedings contain 111 papers from the medicalimagecomputing and computer-assistedintervention - MICCAI 2004 - 7th internationalconference Proceedings, Part I. The topics discussed include: level set methods ...
详细信息
The proceedings contain 111 papers from the medicalimagecomputing and computer-assistedintervention - MICCAI 2004 - 7th internationalconference Proceedings, Part I. The topics discussed include: level set methods in an EM framework for shape classification and estimation;automatic segmentation of neonatal brain MRI;segmentation of 3D probability density fields by surface evolution;improved EM-based tissue segmentation and partial volume effect quantification in multi-sequence brain MRI;a semi-automatic endocardial border detection method for 4D ultrasound data;vessel segmentation using a shape driven flow;learning coupled prior shape and appearance models for segmentation;and local watershed operators for image segmentation.
Ultrasound volumes are corrupted by a multiplicative noise, the speckle, which makes high level analysis difficult. Within each resolution cell a number of elementary scatterers reflects the incident wave toward the s...
详细信息
ISBN:
(纸本)3540229760
Ultrasound volumes are corrupted by a multiplicative noise, the speckle, which makes high level analysis difficult. Within each resolution cell a number of elementary scatterers reflects the incident wave toward the sensor. This paper proposes a method of restoration based on variational principles adapted to ultrasound images statistics. The contribution of this paper is two-fold: we first derive a modified TV scheme to integrate the multiplicative nature of the ultrasound noise and we propose to tune the parameter lambda automatically accordingly to the local noise distribution thanks to the kurtosis information. We present qualitative and quantitative results on various ultrasound volumes.
This paper reports the performance evaluation of a method for visualisation and quantification of intraoperative cortical surface deformations. This method consists in the acquisition of 3D surface meshes of the opera...
详细信息
ISBN:
(纸本)3540229779
This paper reports the performance evaluation of a method for visualisation and quantification of intraoperative cortical surface deformations. This method consists in the acquisition of 3D surface meshes of the operative field directly in the neuronavigator's coordinate system by means of stereoscopic reconstructions, using two cameras attached to the microscope oculars. The locations of about 300 surfaces are compared to the locations of two reference surfaces from a physical phantom: a segmented CT scan with image-to-physical fiducial-based registration, used to compute the overall system performance, and a cloud of points acquired with the neuronavigator's optical localiser, used to compute the intrinsic error of our method. The intrinsic accuracy of our method was shown to be within 1 mm.
A new segmentation scheme is proposed for 3D vascular tree delineation in CTA data sets, which has two essential features. First, the segmentation is carried out locally in a small volume of interest (VOI), second, a ...
详细信息
ISBN:
(纸本)3540229760
A new segmentation scheme is proposed for 3D vascular tree delineation in CTA data sets, which has two essential features. First, the segmentation is carried out locally in a small volume of interest (VOI), second, a global topology estimation is made to initialize a new VOI. The use of local Vol allows that parameter settings for the level set speed function can be optimally set depending on the local image content, which is advantageous especially in vascular tree segmentation where contrast may change significantly, especially in the distal part of the vascular. Moreover, a local approach is significantly faster. A comparison study on five CTA data sets showed that our method has the potential to segment larger part of the vessel tree compared to a similar global level set based segmentation, and in substantially less computation time.
This paper introduces a new technique to detect the coronary arteries as well as other heart's peripheral vessels. After finding the location of the myocardium through a graph theoretic segmentation method, the al...
详细信息
ISBN:
(纸本)3540229760
This paper introduces a new technique to detect the coronary arteries as well as other heart's peripheral vessels. After finding the location of the myocardium through a graph theoretic segmentation method, the algorithm models the heart with a biaxial ellipsoid. For each point of this ellipsoid, we compute the collection of intensities that are normal to the surface. This collection is then filtered to detect the cardiovascular structures. Ultimately, the vessels centerline points are detected using a vessel tracking algorithm, and linked together to form a complete coronary artery tree.
Non-rigid registration of multimodal images is a challenging problem. One approach, maximization of mutual information, has been shown to be effective for registering certain image modalities and is currently consider...
详细信息
ISBN:
(纸本)3540229760
Non-rigid registration of multimodal images is a challenging problem. One approach, maximization of mutual information, has been shown to be effective for registering certain image modalities and is currently considered the standard against which all other techniques are measured. In this paper, we propose an alternative representation of an image based on local phases rather than intensities;we then show how mutual information can be extended to this representation. Local phase acts as a description of local image structure, enabling mutual phase information to detect complex image relationships that are hard or impossible to detect using mutual intensity information. Typical results are presented, comparing the performance of phase and intensity mutual information methods on simulated MR and ultrasound images.
A method fully utilizing multiscale line filter responses is presented to estimate the point spread function (PSF) of a CT scanner and diameters of small tubular structures based on the PSF. The estimation problem is ...
详细信息
ISBN:
(纸本)3540229760
A method fully utilizing multiscale line filter responses is presented to estimate the point spread function (PSF) of a CT scanner and diameters of small tubular structures based on the PSF. The estimation problem is formulated as a least square fitting of a sequence of multiscale responses obtained at each medial axis point to the precomputed multiscale response curve for the ideal line model. The method was validated through phantom experiments and demonstrated to accurately measure small-diameter structures which are significantly overestimated by conventional methods based on the full width half maximum (FWHM) and zero-crossing edge detection.
The problem in 2D/3D intensity-based registration of computed tomography (CT) images is to find a pose such that a digitally reconstructed radiograph (DRR) of the 3D image matches a given 2D image. Generating DRRs, a ...
详细信息
ISBN:
(纸本)3540229779
The problem in 2D/3D intensity-based registration of computed tomography (CT) images is to find a pose such that a digitally reconstructed radiograph (DRR) of the 3D image matches a given 2D image. Generating DRRs, a computationally intensive process, can be accelerated by precomputation [1], or by using custom hardware that accelerates volume rendering. Recent algorithmic advances for 3D texture mapping using off-the-shelf video cards have further improved volume rendering. We propose to generate DRRs generated in real time, without any precomputation, for 2D/3D registration.
The use of the C-arm fluoroscope for surgical navigation in various computerassisted Orthopaedic Surgery Systems (CAOS) has been an important success of research into CAOS technology. To use the fluoroscope for quant...
详细信息
ISBN:
(纸本)3540229779
The use of the C-arm fluoroscope for surgical navigation in various computerassisted Orthopaedic Surgery Systems (CAOS) has been an important success of research into CAOS technology. To use the fluoroscope for quantitative surgical navigation involves calibrating its 2D images and tracking the spatial position of the fluoroscope's image beam. This allows 3D reconstruction of anatomy from a series of 2D fluoroscopic images. This paper presents a new technique for determining the C-arm position and calibrating the image beam. This technique is based on a small imaging phantom that is placed close to the patient. This paper also briefly describes the CAOS system developed at Hull that uses this imaging phantom and reports on in vivo and in vitro studies.
暂无评论