Studying the complex thorax breathing motion is an important research topic for medical (e.g. fus... more Studying the complex thorax breathing motion is an important research topic for medical (e.g. fusion of function and anatomy, radiotherapy planning) and engineering (reduction of motion artifacts) questions. In this paper we present first results on investigating the 4D motion of segmented lung surfaces from CT scans at several different breathing states. For this registra-tion task we extend the shape context approach for shape matching by Belongie et al. [2] from 2D shapes to 3D surfaces and apply it to segmented lung data. Resulting point correspondences can be used e.g. for non-rigid thin-plate-spline registration. We describe our experiments on phantom and real thorax data and show our promising results.
In the last years the need for forensic age estimations in living adolescents increased with migr... more In the last years the need for forensic age estimations in living adolescents increased with migration, particularly from countries where birth dates are not reliably documented. To date, the gold standard of dental age estimation is to evaluate the mineralization and eruption stages of the third molars using an orthopantomogram (OPG). Based on published reference values, the stages are converted in an age estimate in years. However, the use of ionizing radiation without medical indication is ethically controversial and not permitted in many countries. Thus, the aim of this study was to investigate if dental MRI can be used for the assessment of dental age with equally good results as when using an OPG. Methods: 27 healthy volunteers (19♀, 8♂, age range 13.56-23.11y, median 18.92) with at least two present third molars underwent an MRI scan of the jaw within 14 days after a clinically indicated OPG. The examinations were performed on a 3T Magnetom scanner (TimTrio, Siemens AG, Erlan...
We present an optical flow deformable registration method which is based on robust measures for d... more We present an optical flow deformable registration method which is based on robust measures for data and regularization terms. We show two specific implementations of the method, where one penalizes gradients in the displacement field in an isotropic fashion and the other one regularizes by weighting the penalization according to the image gradients anisotropically. Our data term consists of the L 1 -norm of the standard optical flow constraint. We show a numerical algorithm that solves the two proposed models in a primal-dual optimization setup. Our algorithm works in a multi-resolution manner and it is applied to the 20 data sets of the EMPIRE10 registration challenge. Our results show room for improvement. Our rather simple model does not penalize non-diffeomorphic transformations, which leads to bad results on one of the evaluation measures, and it seems unsuited for large deformations cases. However, our algorithm is able to perform registrations of data set sizes around 400 3 ...
Pulmonary hypertension (PH) is a chronic disorder of the pulmonary circulation, marked by an elev... more Pulmonary hypertension (PH) is a chronic disorder of the pulmonary circulation, marked by an elevated vascular resistance and pressure. Our objective is to find an automatic, non-invasive method for estimating the pulmonary pressure based on lung vessels from contrast enhanced CT images. We present a pulmonary vessel extraction algorithm which is fast, fully automatic and robust. It uses a coarse airway tree segmentation and a left and right lung labeled volume to restrict the response of an offset medialness vessel enhancement filter. On a data set of 24 patients, we show that quantitative indices derived from the vascular tree are applicable to distinguish patients with and without PH.
2009 IEEE 12th International Conference on Computer Vision Workshops, ICCV Workshops, 2009
Identity-invariant estimation of head pose from still images is a challenging task due to the hig... more Identity-invariant estimation of head pose from still images is a challenging task due to the high variability of facial appearance. We present a novel 3D head pose estimation approach, which utilizes the flexibility and expressibility of a dense generative 3D facial model in combination with a very fast fitting algorithm. The efficiency of the head pose estimation is obtained by a 2D synthesis of the facial input image. This optimization procedure drives the appearance and pose of the 3D facial model. In contrast to many other approaches we are specifically interested in the more difficult task of head pose estimation from still images, instead of tracking faces in image sequences. We evaluate our approach on two publicly available databases (FacePix and USF HumanID) and compare our method to the 3D morphable model and other state of the art approaches in terms of accuracy and speed.
Procedings of the British Machine Vision Conference 2012, 2012
In this paper we consider the pairwise graph matching problem of finding correspondences between ... more In this paper we consider the pairwise graph matching problem of finding correspondences between two point sets using unary and pairwise potentials, which analyze local descriptor similarity and geometric compatibility. Recently, it was shown that it is possible to learn optimal parameters for the features used in the potentials, which significantly improves results in supervised and unsupervised settings. It was demonstrated that even linear assignments (not considering geometry) with well learned potentials may improve over state-of-the-art quadratic assignment solutions. In this paper we extend this idea by directly learning edge-specific kernels for pairs of nodes. We define the pairwise kernel functions based on a statistical shape model that is learned from labeled training data. Assuming that the setting of graph matching is a priori known, the learned kernel functions allow to significantly improve results in comparison to general graph matching. We further demonstrate the applicability of game theory based evolutionary dynamics as effective and easy to implement approximation of the underlying graph matching optimization problem. Experiments on automatically aligning a set of faces and feature-point based localization of category instances demonstrate the value of the proposed method.
2010 20th International Conference on Pattern Recognition, 2010
We present an approach for unsupervised alignment of an ensemble of images called congealing. Our... more We present an approach for unsupervised alignment of an ensemble of images called congealing. Our algorithm is based on image registration using the mutual information measure as a cost function. The cost function is optimized by a standard gradient descent method in a multiresolution scheme. As opposed to other congealing methods, which use the SSD measure, the mutual information measure is better suited as a similarity measure for registering images since no prior assumptions on the relation of intensities between images are required. We present alignment results on the MNIST handwritten digit database and on facial images obtained from the CVL database.
2009 IEEE 12th International Conference on Computer Vision, 2009
This paper introduces an unsupervised color segmentation method. The underlying idea is to segmen... more This paper introduces an unsupervised color segmentation method. The underlying idea is to segment the input image several times, each time focussing on a different salient part of the image and to subsequently merge all obtained results into one composite segmentation. We identify salient parts of the image by applying affinity propagation clustering to efficiently calculated local color and texture models. Each salient region then serves as an independent initialization for a figure/ground segmentation. Segmentation is done by minimizing a convex energy functional based on weighted total variation leading to a global optimal solution. Each salient region provides an accurate figure/ground segmentation highlighting different parts of the image. These highly redundant results are combined into one composite segmentation by analyzing local segmentation certainty. Our formulation is quite general, and other salient region detection algorithms in combination with any semi-supervised figure/ground segmentation approach can be used. We demonstrate the high quality of our method on the well-known Berkeley segmentation database. Furthermore we show that our method can be used to provide good spatial support for recognition frameworks.
2013 IEEE International Conference on Computer Vision Workshops, 2013
ABSTRACT Integral image data structures are very useful in computer vision applications that invo... more ABSTRACT Integral image data structures are very useful in computer vision applications that involve machine learning approaches based on ensembles of weak learners. The weak learners often are simply several regional sums of intensities subtracted from each other. In this work we present a memory efficient integral volume data structure, that allows reduction of required RAM storage size in such a supervised learning framework using 3D training data. We evaluate our proposed data structure in terms of the tradeoff between computational effort and storage, and show an application for 3D object detection of liver CT data.
The need for forensic age estimations in living adolescents is high mainly due to migration, part... more The need for forensic age estimations in living adolescents is high mainly due to migration, particularly from countries where birth dates are not reliably documented. To date, the gold standard of dental age estimation is the evaluation of the mineralization and eruption stages of the third molars using an orthopantomogram (OPG). However, the use of ionizing radiation without medical indication is ethically controversial and not permitted in many countries.
This paper deals with segmentation of image sequences in an unsupervised manner with the goal of ... more This paper deals with segmentation of image sequences in an unsupervised manner with the goal of getting highly consistent segmentation results from frame-to-frame. We first introduce a segmentation method that uses results of the previous frame as initialization and significantly improves consistency in comparison to a single frame based approach. We also find correspondences between the segmented regions from one frame to the next to further increase consistency. This matching step is based on a modified version of an efficient partial shape matching method which allows identification of similar parts of regions despite topology changes like merges and splits. We use the identified matched parts to define a partial matching cost which is then used as input to pairwise graph matching. Experiments demonstrate that we can achieve highly consistent segmentations for diverse image sequences, even allowing to track manually initialized moving and static objects.
Communications in Computer and Information Science, 2010
The Active Appearance Model (AAM) is a widely used approach for model based vision showing excell... more The Active Appearance Model (AAM) is a widely used approach for model based vision showing excellent results. But one major drawback is that the method is not robust against occlusions. Thus, if parts of the image are occluded the method converges to local minima and the obtained results are unreliable. To overcome this problem we propose a robust AAM fitting strategy. The main idea is to apply a robust PCA model to reconstruct the missing feature information and to use the thus obtained image as input for the standard AAM fitting process. Since existing methods for robust PCA reconstruction are computationally too expensive for real-time processing we applied a more efficient method: Fast-Robust PCA (FR-PCA). In fact, by using our FR-PCA the computational effort is drastically reduced. Moreover, more accurate reconstructions are obtained. In the experiments, we evaluated both, the FR-PCA model on the publicly available ALOI database and the whole robust AAM fitting chain on facial images. The results clearly show the benefits of our approach in terms of accuracy and speed when processing disturbed data (i.e., images containing occlusions).
Lecture Notes in Computational Vision and Biomechanics, 2015
Automatic segmentation of 3D vertebrae is a challenging task in medical imaging. In this paper, w... more Automatic segmentation of 3D vertebrae is a challenging task in medical imaging. In this paper, we introduce a total variation (TV) based framework that incorporates an a priori model, i.e., a vertebral mean shape, image intensity and edge information. The algorithm was evaluated using leave-one-out cross validation on a data set containing ten computed tomography scans and ground truth segmentations provided for the CSI MICCAI 2014 spine and vertebrae segmentation challenge. We achieve promising results in terms of the Dice Similarity Coefficient (DSC) of 0.93 ± 0.04 averaged over the whole data set.
2014 IEEE 11th International Symposium on Biomedical Imaging (ISBI), 2014
ABSTRACT The determination of an individual’s legal majority age is becoming increasingly importa... more ABSTRACT The determination of an individual’s legal majority age is becoming increasingly important in forensic practice. Established age estimation methods are based on 2D X-rays, but suffer from problems due to projective imaging and exposure to ionizing radiation, which, without proper medical or criminal indication, is ethically questionable and legally prohibited in many countries. We propose an automatic 3D method for the determination of legal maturity from MR images based on the ossification of the radius bone. Age estimation is performed by a linear regression model of the epiphyseal gap volume over the known ground truth age of training data. Results are comparable with the established Greulich/Pyle (GP) and Tanner/Whitehouse (TW) methods, but do not involve harmful radiation.
Nonlinear image registration is an initial step for a large number of medical image analysis appl... more Nonlinear image registration is an initial step for a large number of medical image analysis applications. Optical flow based intensity registration is often used for dealing with intra-modality applications involving motion differences. In this work we present an energy functional which uses a novel, second-order regularization prior of the displacement field. Compared to other methods our scheme is robust to non-Gaussian noise and does not penalize locally affine deformation fields in homogeneous areas. We propose an efficient and stable numerical scheme to find the minimizer of the presented energy. We implemented our algorithm using modern consumer graphics processing units and thereby increased the execution performance dramatically. We further show experimental evaluations on clinical CT thorax data sets at different breathing states and on dynamic 4D CT cardiac data sets.
Principal Component Analysis (PCA) is a powerful and widely used tool in Computer Vision and is a... more Principal Component Analysis (PCA) is a powerful and widely used tool in Computer Vision and is applied, e.g., for dimensionality reduction. But as a drawback, it is not robust to outliers. Hence, if the input data is corrupted, an arbitrarily wrong representation is obtained. To overcome this problem, various methods have been proposed to robustly estimate the PCA coefficients, but these methods are computationally too expensive for practical applications. Thus, in this paper we propose a novel fast and robust PCA (FR-PCA), which drastically reduces the computational effort. Moreover, more accurate representations are obtained. In particular, we propose a two-stage outlier detection procedure, where in the first stage outliers are detected by analyzing a large number of smaller subspaces. In the second stage, remaining outliers are detected by a robust least-square fitting. To show these benefits, in the experiments we evaluate the FR-PCA method for the task of robust image reconstruction on the publicly available ALOI database. The results clearly show that our approach outperforms existing methods in terms of accuracy and speed when processing corrupted data.
Studying the complex thorax breathing motion is an important research topic for medical (e.g. fus... more Studying the complex thorax breathing motion is an important research topic for medical (e.g. fusion of function and anatomy, radiotherapy planning) and engineering (reduction of motion artifacts) questions. In this paper we present first results on investigating the 4D motion of segmented lung surfaces from CT scans at several different breathing states. For this registra-tion task we extend the shape context approach for shape matching by Belongie et al. [2] from 2D shapes to 3D surfaces and apply it to segmented lung data. Resulting point correspondences can be used e.g. for non-rigid thin-plate-spline registration. We describe our experiments on phantom and real thorax data and show our promising results.
In the last years the need for forensic age estimations in living adolescents increased with migr... more In the last years the need for forensic age estimations in living adolescents increased with migration, particularly from countries where birth dates are not reliably documented. To date, the gold standard of dental age estimation is to evaluate the mineralization and eruption stages of the third molars using an orthopantomogram (OPG). Based on published reference values, the stages are converted in an age estimate in years. However, the use of ionizing radiation without medical indication is ethically controversial and not permitted in many countries. Thus, the aim of this study was to investigate if dental MRI can be used for the assessment of dental age with equally good results as when using an OPG. Methods: 27 healthy volunteers (19♀, 8♂, age range 13.56-23.11y, median 18.92) with at least two present third molars underwent an MRI scan of the jaw within 14 days after a clinically indicated OPG. The examinations were performed on a 3T Magnetom scanner (TimTrio, Siemens AG, Erlan...
We present an optical flow deformable registration method which is based on robust measures for d... more We present an optical flow deformable registration method which is based on robust measures for data and regularization terms. We show two specific implementations of the method, where one penalizes gradients in the displacement field in an isotropic fashion and the other one regularizes by weighting the penalization according to the image gradients anisotropically. Our data term consists of the L 1 -norm of the standard optical flow constraint. We show a numerical algorithm that solves the two proposed models in a primal-dual optimization setup. Our algorithm works in a multi-resolution manner and it is applied to the 20 data sets of the EMPIRE10 registration challenge. Our results show room for improvement. Our rather simple model does not penalize non-diffeomorphic transformations, which leads to bad results on one of the evaluation measures, and it seems unsuited for large deformations cases. However, our algorithm is able to perform registrations of data set sizes around 400 3 ...
Pulmonary hypertension (PH) is a chronic disorder of the pulmonary circulation, marked by an elev... more Pulmonary hypertension (PH) is a chronic disorder of the pulmonary circulation, marked by an elevated vascular resistance and pressure. Our objective is to find an automatic, non-invasive method for estimating the pulmonary pressure based on lung vessels from contrast enhanced CT images. We present a pulmonary vessel extraction algorithm which is fast, fully automatic and robust. It uses a coarse airway tree segmentation and a left and right lung labeled volume to restrict the response of an offset medialness vessel enhancement filter. On a data set of 24 patients, we show that quantitative indices derived from the vascular tree are applicable to distinguish patients with and without PH.
2009 IEEE 12th International Conference on Computer Vision Workshops, ICCV Workshops, 2009
Identity-invariant estimation of head pose from still images is a challenging task due to the hig... more Identity-invariant estimation of head pose from still images is a challenging task due to the high variability of facial appearance. We present a novel 3D head pose estimation approach, which utilizes the flexibility and expressibility of a dense generative 3D facial model in combination with a very fast fitting algorithm. The efficiency of the head pose estimation is obtained by a 2D synthesis of the facial input image. This optimization procedure drives the appearance and pose of the 3D facial model. In contrast to many other approaches we are specifically interested in the more difficult task of head pose estimation from still images, instead of tracking faces in image sequences. We evaluate our approach on two publicly available databases (FacePix and USF HumanID) and compare our method to the 3D morphable model and other state of the art approaches in terms of accuracy and speed.
Procedings of the British Machine Vision Conference 2012, 2012
In this paper we consider the pairwise graph matching problem of finding correspondences between ... more In this paper we consider the pairwise graph matching problem of finding correspondences between two point sets using unary and pairwise potentials, which analyze local descriptor similarity and geometric compatibility. Recently, it was shown that it is possible to learn optimal parameters for the features used in the potentials, which significantly improves results in supervised and unsupervised settings. It was demonstrated that even linear assignments (not considering geometry) with well learned potentials may improve over state-of-the-art quadratic assignment solutions. In this paper we extend this idea by directly learning edge-specific kernels for pairs of nodes. We define the pairwise kernel functions based on a statistical shape model that is learned from labeled training data. Assuming that the setting of graph matching is a priori known, the learned kernel functions allow to significantly improve results in comparison to general graph matching. We further demonstrate the applicability of game theory based evolutionary dynamics as effective and easy to implement approximation of the underlying graph matching optimization problem. Experiments on automatically aligning a set of faces and feature-point based localization of category instances demonstrate the value of the proposed method.
2010 20th International Conference on Pattern Recognition, 2010
We present an approach for unsupervised alignment of an ensemble of images called congealing. Our... more We present an approach for unsupervised alignment of an ensemble of images called congealing. Our algorithm is based on image registration using the mutual information measure as a cost function. The cost function is optimized by a standard gradient descent method in a multiresolution scheme. As opposed to other congealing methods, which use the SSD measure, the mutual information measure is better suited as a similarity measure for registering images since no prior assumptions on the relation of intensities between images are required. We present alignment results on the MNIST handwritten digit database and on facial images obtained from the CVL database.
2009 IEEE 12th International Conference on Computer Vision, 2009
This paper introduces an unsupervised color segmentation method. The underlying idea is to segmen... more This paper introduces an unsupervised color segmentation method. The underlying idea is to segment the input image several times, each time focussing on a different salient part of the image and to subsequently merge all obtained results into one composite segmentation. We identify salient parts of the image by applying affinity propagation clustering to efficiently calculated local color and texture models. Each salient region then serves as an independent initialization for a figure/ground segmentation. Segmentation is done by minimizing a convex energy functional based on weighted total variation leading to a global optimal solution. Each salient region provides an accurate figure/ground segmentation highlighting different parts of the image. These highly redundant results are combined into one composite segmentation by analyzing local segmentation certainty. Our formulation is quite general, and other salient region detection algorithms in combination with any semi-supervised figure/ground segmentation approach can be used. We demonstrate the high quality of our method on the well-known Berkeley segmentation database. Furthermore we show that our method can be used to provide good spatial support for recognition frameworks.
2013 IEEE International Conference on Computer Vision Workshops, 2013
ABSTRACT Integral image data structures are very useful in computer vision applications that invo... more ABSTRACT Integral image data structures are very useful in computer vision applications that involve machine learning approaches based on ensembles of weak learners. The weak learners often are simply several regional sums of intensities subtracted from each other. In this work we present a memory efficient integral volume data structure, that allows reduction of required RAM storage size in such a supervised learning framework using 3D training data. We evaluate our proposed data structure in terms of the tradeoff between computational effort and storage, and show an application for 3D object detection of liver CT data.
The need for forensic age estimations in living adolescents is high mainly due to migration, part... more The need for forensic age estimations in living adolescents is high mainly due to migration, particularly from countries where birth dates are not reliably documented. To date, the gold standard of dental age estimation is the evaluation of the mineralization and eruption stages of the third molars using an orthopantomogram (OPG). However, the use of ionizing radiation without medical indication is ethically controversial and not permitted in many countries.
This paper deals with segmentation of image sequences in an unsupervised manner with the goal of ... more This paper deals with segmentation of image sequences in an unsupervised manner with the goal of getting highly consistent segmentation results from frame-to-frame. We first introduce a segmentation method that uses results of the previous frame as initialization and significantly improves consistency in comparison to a single frame based approach. We also find correspondences between the segmented regions from one frame to the next to further increase consistency. This matching step is based on a modified version of an efficient partial shape matching method which allows identification of similar parts of regions despite topology changes like merges and splits. We use the identified matched parts to define a partial matching cost which is then used as input to pairwise graph matching. Experiments demonstrate that we can achieve highly consistent segmentations for diverse image sequences, even allowing to track manually initialized moving and static objects.
Communications in Computer and Information Science, 2010
The Active Appearance Model (AAM) is a widely used approach for model based vision showing excell... more The Active Appearance Model (AAM) is a widely used approach for model based vision showing excellent results. But one major drawback is that the method is not robust against occlusions. Thus, if parts of the image are occluded the method converges to local minima and the obtained results are unreliable. To overcome this problem we propose a robust AAM fitting strategy. The main idea is to apply a robust PCA model to reconstruct the missing feature information and to use the thus obtained image as input for the standard AAM fitting process. Since existing methods for robust PCA reconstruction are computationally too expensive for real-time processing we applied a more efficient method: Fast-Robust PCA (FR-PCA). In fact, by using our FR-PCA the computational effort is drastically reduced. Moreover, more accurate reconstructions are obtained. In the experiments, we evaluated both, the FR-PCA model on the publicly available ALOI database and the whole robust AAM fitting chain on facial images. The results clearly show the benefits of our approach in terms of accuracy and speed when processing disturbed data (i.e., images containing occlusions).
Lecture Notes in Computational Vision and Biomechanics, 2015
Automatic segmentation of 3D vertebrae is a challenging task in medical imaging. In this paper, w... more Automatic segmentation of 3D vertebrae is a challenging task in medical imaging. In this paper, we introduce a total variation (TV) based framework that incorporates an a priori model, i.e., a vertebral mean shape, image intensity and edge information. The algorithm was evaluated using leave-one-out cross validation on a data set containing ten computed tomography scans and ground truth segmentations provided for the CSI MICCAI 2014 spine and vertebrae segmentation challenge. We achieve promising results in terms of the Dice Similarity Coefficient (DSC) of 0.93 ± 0.04 averaged over the whole data set.
2014 IEEE 11th International Symposium on Biomedical Imaging (ISBI), 2014
ABSTRACT The determination of an individual’s legal majority age is becoming increasingly importa... more ABSTRACT The determination of an individual’s legal majority age is becoming increasingly important in forensic practice. Established age estimation methods are based on 2D X-rays, but suffer from problems due to projective imaging and exposure to ionizing radiation, which, without proper medical or criminal indication, is ethically questionable and legally prohibited in many countries. We propose an automatic 3D method for the determination of legal maturity from MR images based on the ossification of the radius bone. Age estimation is performed by a linear regression model of the epiphyseal gap volume over the known ground truth age of training data. Results are comparable with the established Greulich/Pyle (GP) and Tanner/Whitehouse (TW) methods, but do not involve harmful radiation.
Nonlinear image registration is an initial step for a large number of medical image analysis appl... more Nonlinear image registration is an initial step for a large number of medical image analysis applications. Optical flow based intensity registration is often used for dealing with intra-modality applications involving motion differences. In this work we present an energy functional which uses a novel, second-order regularization prior of the displacement field. Compared to other methods our scheme is robust to non-Gaussian noise and does not penalize locally affine deformation fields in homogeneous areas. We propose an efficient and stable numerical scheme to find the minimizer of the presented energy. We implemented our algorithm using modern consumer graphics processing units and thereby increased the execution performance dramatically. We further show experimental evaluations on clinical CT thorax data sets at different breathing states and on dynamic 4D CT cardiac data sets.
Principal Component Analysis (PCA) is a powerful and widely used tool in Computer Vision and is a... more Principal Component Analysis (PCA) is a powerful and widely used tool in Computer Vision and is applied, e.g., for dimensionality reduction. But as a drawback, it is not robust to outliers. Hence, if the input data is corrupted, an arbitrarily wrong representation is obtained. To overcome this problem, various methods have been proposed to robustly estimate the PCA coefficients, but these methods are computationally too expensive for practical applications. Thus, in this paper we propose a novel fast and robust PCA (FR-PCA), which drastically reduces the computational effort. Moreover, more accurate representations are obtained. In particular, we propose a two-stage outlier detection procedure, where in the first stage outliers are detected by analyzing a large number of smaller subspaces. In the second stage, remaining outliers are detected by a robust least-square fitting. To show these benefits, in the experiments we evaluate the FR-PCA method for the task of robust image reconstruction on the publicly available ALOI database. The results clearly show that our approach outperforms existing methods in terms of accuracy and speed when processing corrupted data.
Uploads
Papers by Martin Urschler