-
ISMRM Benelux, Rotterdam, Netherlands; 01/2013
-
[show abstract]
[hide abstract]
ABSTRACT: Research in face recognition has continuously been challenged by extrinsic (head pose, lighting conditions) and intrin-sic (facial expression, aging) sources of variability. While many survey papers on face recognition exist, in this paper, we focus on a comparative study of 3-D face recognition under expression varia-tions. As a first contribution, 3-D face databases with expressions are listed, and the most important ones are briefly presented and their complexity is quantified using the iterative closest point (ICP) baseline recognition algorithm. This allows to rank the databases according to their inherent difficulty for face-recognition tasks. This analysis reveals that the FRGC v2 database can be consid-ered as the most challenging because of its size, the presence of expressions and outliers, and the time lapse between the record-ings. Therefore, we recommend to use this database as a reference database to evaluate (expression-invariant) 3-D face-recognition al-gorithms. We also determine and quantify the most important fac-tors that influence the performance. It appears that performance decreases 1) with the degree of nonfrontal pose, 2) for certain ex-pression types, 3) with the magnitude of the expressions, 4) with an increasing number of expressions, and 5) for a higher number of gallery subjects. Future 3-D face-recognition algorithms should be evaluated on the basis of all these factors. As the second con-tribution, a survey of published 3-D face-recognition methods that deal with expression variations is given. These methods are subdi-vided into three classes depending on the way the expressions are handled. Region-based methods use expression-stable regions only, while other methods model the expressions either using an isomet-ric or a statistical model. Isometric models assume the deformation because of expression variation to be (locally) isometric, meaning that the deformation preserves lengths along the surface. Statistical models learn how the facial soft tissue deforms during expressions based on a training database with expression labels. Algorithmic performances are evaluated by the comparison of recognition rates for identification and verification. No statistical significant differ-ences in class performance are found between any pair of classes.
IEEE Transactions on Systems Man and Cybernetics Part C (Applications and Reviews) 11/2012; · 2.01 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Matching 3D faces for recognition is a challenging task caused by the presence of expression variations, missing data, and outliers. In this paper the meshSIFT algorithm and its use for 3D face recognition is presented. This algorithm consists of four major components. First, salient points on the 3D facial surface are detected as mean curvature extrema in scale space. Second, orientations are assigned to each of these salient points. Third, the neighbourhood of each salient point is described in a feature vector consisting of concatenated histograms of shape indices and slant angles. Fourth, the feature vectors of two 3D facial surfaces are reliably matched by comparing the angles in feature space. This results in an algorithm which is robust to expression variations, missing data and outliers. As a first contribution, we demonstrate that the number of matching meshSIFT features is a reliable measure for expression-invariant face recognition, as shown by the rank 1 recognition rate of 93.7% and 89.6% for the Bosphorus and FRGC v2 database, respectively. Next, we demonstrate that symmetrizing the feature descriptors allows comparing two 3D facial surfaces with limited or no overlap. Validation on the data of the "SHREC '11: Face Scans" contest, containing many partial scans, resulted in a recognition rate of 98.6%, clearly outperforming all other participants in the challenge. Finally, we also demonstrate the use of meshSIFT for two other problems related with 3D face recognition: pose normalisation and symmetry plane estimation. For both problems, applying meshSIFT in combination with RANSAC resulted in a correct solution for ±90% of all Bosphorus database meshes (except ±90 • and ±45 • rotations).
Computer Vision and Image Understanding 11/2012; · 1.34 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Diffusion weighted imaging (DWI) allows to delineate neural fibres, based on local, directional information of the diffusion of water. Due to its directional nature, the local information needs to be reoriented upon image transformation, in order to preserve correspondence to the anatomy. In this work, we show that reorientation of the fODF with preservation of volume fractions (PVF) affects both deterministic and probabilistic fibre tracking. We identify the main causes for this, and validate them on synthetic and real brain DWI data. The problem is not with the PVF reorientation itself, but rather with the fODF reconstruction, its use in fibre tracking, and the influence of the seeds.
MICCAI 2012 Workshop on Computational Diffusion MRI, Nice, France; 10/2012
-
Journal of Orthopaedic Research 06/2012; 30(12):2054-6. · 2.81 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Accurate measurement of facial sexual dimorphism is useful to understanding facial anatomy and specifically how faces influence, and have been influenced by, sexual selection. An important facial aspect is the display of bilateral symmetry, invoking the need to investigate aspects of symmetry and asymmetry separately when examining facial shape. Previous studies typically employed landmarks that provided only a sparse facial representation, where different landmark choices could lead to contrasting outcomes. Furthermore, sexual dimorphism is only tested as a difference of sample means, which is statistically the same as a difference in population location only. Within the framework of geometric morphometrics, we partition facial shape, represented in a spatially dense way, into patterns of symmetry and asymmetry, following a two-factor anova design. Subsequently, we investigate sexual dimorphism in symmetry and asymmetry patterns separately, and on multiple aspects, by examining (i) population location differences as well as differences in population variance-covariance; (ii) scale; and (iii) orientation. One important challenge in this approach is the proportionally high number of variables to observations necessitating the implementation of permutational and computationally feasible statistics. In a sample of gender-matched young adults (18-25 years) with self-reported European ancestry, we found greater variation in male faces than in women for all measurements. Statistically significant sexual dimorphism was found for the aspect of location in both symmetry and asymmetry (directional asymmetry), for the aspect of scale only in asymmetry (magnitude of fluctuating asymmetry) and, in contrast, for the aspect of orientation only in symmetry. Interesting interplays with hypotheses in evolutionary and developmental biology were observed, such as the selective nature of the force underpinning sexual dimorphism and the genetic independence of the structural patterns of fluctuating asymmetry. Additionally, insights into growth patterns of the soft tissue envelope of the face and underlying skull structure can also be obtained from the results.
Journal of Anatomy 06/2012; 221(2):97-114. · 2.37 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Piecewise rigid registration is a fundamental problem in medical imaging involving intra-patient pose differences in multiple medical images, mostly due to articulated motion. In this paper, we propose a method to extract multiple rigid transformations in 2D medical images in the presence of outliers. First, points of interest in the images are extracted and matched with the SIFT algorithm. Secondly, multiple rigid motions are sampled and clustered by the mean shift algorithm in the special Euclidean group SE(2), a smooth manifold of 2-D rigid transformation matrices. The method proposed is evaluated for intra-subject registrations of knee fluoroscopy images, demonstrating a mean angular and trans-lational error on the estimated motion of 0.39 • and 6.65 pixels, respectively.
International Symposium on Biomedical Imaging; 05/2012
-
[show abstract]
[hide abstract]
ABSTRACT: 18F-fluorodeoxyglucose (FDG) positron emission tomography (PET) has become the de facto standard for current clinical therapy follow up evaluations. In pursuit of robust biomarkers for predicting early therapy response, an efficient marker quantification procedure is certainly a necessity. Among various PET derived markers, the clinical investigations indicated that the total lesion metabolic activity (TLA) of a tumor lesion has a good prognostic value in several longitudinal studies. We utilize a fuzzy multi-class modeling using a stochastic expectation maximization (SEM) algorithm to fit a finite mixture model (FMM) to the PET image. We then propose a direct estimation formula for TLA and SUVmean from this multi-class statistical model. In order to evaluate our proposition, a realistic liver lesion is simulated and reconstructed. All results were evaluated with reference to the ground truth knowledge. Our experimental study conveys that the proposed method is robust enough to handle background heterogeneities in realistic scenarios.
Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention. 01/2012; 15(Pt 1):107-14.
-
[show abstract]
[hide abstract]
ABSTRACT: The study of typical morphological variations using quantitative, morphometric descriptors has always interested biologists in general. However, unusual examples of form, such as abnormalities are often encountered in biomedical sciences. Despite the long history of morphometrics, the means to identify and quantify such unusual form differences remains limited.
A theoretical concept, called dysmorphometrics, is introduced augmenting current geometric morphometrics with a focus on identifying and modelling form abnormalities. Dysmorphometrics applies the paradigm of detecting form differences as outliers compared to an appropriate norm. To achieve this, the likelihood formulation of landmark superimpositions is extended with outlier processes explicitly introducing a latent variable coding for abnormalities. A tractable solution to this augmented superimposition problem is obtained using Expectation-Maximization. The topography of detected abnormalities is encoded in a dysmorphogram.
We demonstrate the use of dysmorphometrics to measure abrupt changes in time, asymmetry and discordancy in a set of human faces presenting with facial abnormalities.
The results clearly illustrate the unique power to reveal unusual form differences given only normative data with clear applications in both biomedical practice & research.
Theoretical Biology and Medical Modelling 01/2012; 9:5. · 1.86 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: In this paper an automated method is presented for the localization of cephalometric landmarks in craniofacial cone-beam computed
tomography images. This method makes use of a statistical sparse appearance and shape model obtained from training data. The
sparse appearance model captures local image intensity patterns around each landmark. The sparse shape model, on the other
hand, is constructed by embedding the landmarks in a graph. The edges of this graph represent pairwise spatial dependencies
between landmarks, hence leading to a sparse shape model. The edges connecting different landmarks are defined in an automated
way based on the intrinsic topology present in the training data. A maximum a posteriori approach is employed to obtain an
energy function. To minimize this energy function, the problem is discretized by considering a finite set of candidate locations
for each landmark, leading to a labeling problem. Using a leave-one-out approach on the training data the overall accuracy
of the method is assessed. The mean and median error values for all landmarks are equal to 2.41mm\textrm{mm} and 1.49mm\textrm{mm}, respectively, demonstrating a clear improvement over previously published methods.
09/2011: pages 249-256;
-
[show abstract]
[hide abstract]
ABSTRACT: Biomechanical parameters of gait such as muscle's moment arm length (MAL) and muscle-tendon length are known to be sensitive to anatomical variability. Nevertheless, most studies rely on rescaled generic models (RGMo) constructed from averaged data of cadaveric measurements in a healthy adult population. As an alternative, deformable generic models (DGMo) have been proposed. These models integrate a higher level of subject-specific detail by applying characteristic deformations to the musculoskeletal geometry. In contrast, musculoskeletal models based on magnetic resonance (MR) images (MRMo) reflect the involved subject's characteristics in every level of the model. This study investigated the effect of the varying levels of subject-specific detail in these three model types on the calculated hip MAL during gait in a pediatric population of seven cerebral palsy subjects presenting aberrant femoral geometry. Our results show large percentage differences in calculated MAL between RGMo and MRMo. Furthermore, the use of DGMo did not uniformly reduce inter-model differences in calculated MAL. The magnitude of these percentage differences stresses the need to take these effects into account when selecting the level of subject-specific detail one wants to integrate in musculoskeletal. Furthermore, the variability of these differences between subjects and between muscles makes it very difficult to a priori estimate their importance for a biomechanical analysis of a certain muscle in a given subject.
Journal of biomechanics 02/2011; 44(7):1346-53. · 2.66 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Rescaling generic models is the most frequently applied approach in generating biomechanical models for inverse kinematics. Nevertheless it is well known that this procedure introduces errors in calculated gait kinematics due to: (1) errors associated with palpation of anatomical landmarks, (2) inaccuracies in the definition of joint coordinate systems. Based on magnetic resonance (MR) images, more accurate, subject-specific kinematic models can be built that are significantly less sensitive to both error types. We studied the difference between the two modelling techniques by quantifying differences in calculated hip and knee joint kinematics during gait. In a clinically relevant patient group of 7 pediatric cerebral palsy (CP) subjects with increased femoral anteversion, gait kinematic were calculated using (1) rescaled generic kinematic models and (2) subject-specific MR-based models. In addition, both sets of kinematics were compared to those obtained using the standard clinical data processing workflow. Inverse kinematics, calculated using rescaled generic models or the standard clinical workflow, differed largely compared to kinematics calculated using subject-specific MR-based kinematic models. The kinematic differences were most pronounced in the sagittal and transverse planes (hip and knee flexion, hip rotation). This study shows that MR-based kinematic models improve the reliability of gait kinematics, compared to generic models based on normal subjects. This is the case especially in CP subjects where bony deformations may alter the relative configuration of joint coordinate systems. Whilst high cost impedes the implementation of this modeling technique, our results demonstrate that efforts should be made to improve the level of subject-specific detail in the joint axes determination.
Gait & posture 02/2011; 33(2):158-64. · 2.58 Impact Factor
-
Machine Learning in Medical Imaging - Second International Workshop, MLMI 2011, Held in Conjunction with MICCAI 2011, Toronto, Canada, September 18, 2011. Proceedings; 01/2011
-
Machine Learning in Medical Imaging - Second International Workshop, MLMI 2011, Held in Conjunction with MICCAI 2011, Toronto, Canada, September 18, 2011. Proceedings; 01/2011
-
Proceedings of the 8th IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2011, March 30 - April 2, 2011, Chicago, Illinois, USA; 01/2011
-
[show abstract]
[hide abstract]
ABSTRACT: In the field of diffusion weighted imaging (DWI), it is common to fit one of many available models to the acquired data. A hybrid diffusion imaging (HYDI) approach even allows to reconstruct different models and measures from a single dataset. Methods for DWI atlas construction (and registration) are as plenty as the number of available models. Therefore, it would be nice if we were able to perform atlas building before model reconstruction. In this work, we present a method for atlas construction of DWI data in q-space: we developed a new multi-subject multi-channel diffeomorphic matching algorithm, which is combined with a recently proposed DWI retransformation method in q-space. We applied our method to HYDI data of 10 healthy subjects. From the resulting atlas, we also reconstructed some advanced models. We hereby demonstrate the feasibility of q-space atlas building, as well as the quality, advantages and possibilities of such an atlas.
Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention. 01/2011; 14(Pt 2):166-73.
-
Remco C. Veltkamp,
Stefan van Jole,
Hassen Drira,
Boulbaba Ben Amor,
Mohamed Daoudi,
Huibin Li,
Liming Chen,
Peter Claes,
Dirk Smeets,
Jeroen Hermans,
Dirk Vandermeulen, Paul Suetens
Eurographics Workshop on 3D Object Retrieval 2011, Llandudno, UK, April 10, 2011. Proceedings; 01/2011
-
[show abstract]
[hide abstract]
ABSTRACT: In this paper the problem of pairwise model-to-scene point set registration is considered. Three contributions are made. Firstly, the relations between correspondence-based and some information-theoretic point cloud registration algorithms are formalized. Starting from the observation that the outlier handling of existing methods relies on heuristically determined models, a second contribution is made exploiting aforementioned relations to derive a new robust point set registration algorithm. Representing model and scene point clouds by mixtures of Gaus-sians, the method minimizes their Kullback-Leibler divergence both w.r.t. the registration transformation parameters and w.r.t. the scene's mixture coefficients. This results in an Expectation-Maximization Iterative Closest Point (EM-ICP) approach with a parameter-free outlier model that is optimal in information-theoretical sense. While the current (CUDA) implementation is limited to the rigid registration case, the underlying theory applies to both rigid and non-rigid point set registration. As a by-product of the registration algorithm's theory, a third contribution is made by suggesting a new point cloud Kernel Density Estimation approach which relies on maximizing the resulting distribution's entropy w.r.t. the kernel weights. The rigid registration algorithm is applied to align different patches of the publicly available Stanford Dragon and Stanford Happy Budha range data. The results show good performance regarding accuracy, robustness and convergence range.
The 24th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2011, Colorado Springs, CO, USA, 20-25 June 2011; 01/2011
-
[show abstract]
[hide abstract]
ABSTRACT: Local motion within intra-patient biomedical images can be compensated by using elastic image registration. The application of B-spline based elastic registration during interventional treatment is seriously hampered by its considerable computation time. The graphics processing unit (GPU) can be used to accelerate the calculation of such elastic registrations by using its parallel processing power, and by employing the hardwired tri-linear interpolation capabilities in order to efficiently perform the cubic B-spline evaluation. In this article it is shown that the similarity measure and its derivatives also can be calculated on the GPU, using a two pass approach. On average a speedup factor 50 compared to a straight-forward CPU implementation was reached.
Computer methods and programs in biomedicine 10/2010; 103(2):104-12. · 1.14 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Many short- and intermediate-term radiological and clinical studies on cervical arthroplasty with the Bryan Cervical Disc have been published, providing, most of the time, satisfactory results.
To prospectively assess the intermediate and long-term radiographic characteristics of disk replacement surgery with the Bryan Cervical Disc and to correlate these results with clinical outcome.
Range of motion was measured with a validated tool. Intervertebral disk degeneration was assessed with a quantitative scoring system. Heterotopic ossification was evaluated with a previously published scoring system. Device stability was investigated by measuring subsidence and anteroposterior migration. General clinical patient outcome was assessed with the Odom classification system.
Eighty-nine patients were initially included in this prospective long-term study. One patient was reoperated on at the index level and 4 were reoperated on at an adjacent level; those patients were not further analyzed. The mobility at the treated level was preserved in > or = 85% of our cases. The insertion of the prosthesis did not lead to an increase in mobility at the adjacent levels. The degeneration score increased at both adjacent levels. Heterotopic ossification was present in 34% to 39% of the patients, depending on the follow-up point. No cases of anteroposterior migration or subsidence were found. More than 82% of all patients had a good to excellent clinical outcome in the long run.
The device maintains preoperative motion at the index and adjacent levels, seems to protect against acceleration of adjacent-level degeneration as seen after anterior cervical discectomy and fusion, and remains securely anchored in the adjacent bone mass in the long run. Heterotopic ossification was frequently seen. The vast majority of all patients had a good to excellent clinical outcome.
Neurosurgery 09/2010; 67(3):679-87; discussion 687. · 2.79 Impact Factor
