Almost isometric mesh parameterization through abstract domains

Integration of Swin UNETR and statistical shape modeling for a semi-automated segmentation of the knee and biomechanical modeling of articular cartilage

Simulation studies, such as finite element (FE) modeling, provide insight into knee joint mechanics without patient involvement. Generic FE models mimic the biomechanical behavior of the tissue, but overlook variations in geometry, loading, and material properties of a population. Conversely, subject-specific models include these factors, resulting in enhanced predictive precision, but are laborious and time intensive. The present study aimed to enhance subject-specific knee joint FE modeling by incorporating a semi-automated segmentation algorithm using a 3D Swin UNETR for an initial segmentation of the femur and tibia, followed by a statistical shape model (SSM) adjustment to improve surface roughness and continuity. For comparison, a manual FE model was developed through manual segmentation (i.e., the de-facto standard approach). Both FE models were subjected to gait loading and the predicted mechanical response was compared. The semi-automated segmentation achieved a Dice similarity coefficient (DSC) of over 98% for both the femur and tibia. Hausdorff distance (mm) between the semi-automated and manual segmentation was 1.4 mm. The mechanical results (max principal stress and strain, fluid pressure, fibril strain, and contact area) showed no significant differences between the manual and semi-automated FE models, indicating the effectiveness of the proposed semi-automated segmentation in creating accurate knee joint FE models. We have made our semi-automated models publicly accessible to support and facilitate biomechanical modeling and medical image segmentation efforts ( https://data.mendeley.com/datasets/k5hdc9cz7w/1 ).

Impact of data synthesis strategies for the classification of craniosynostosis

Introduction

Photogrammetric surface scans provide a radiation-free option to assess and classify craniosynostosis. Due to the low prevalence of craniosynostosis and high patient restrictions, clinical data are rare. Synthetic data could support or even replace clinical data for the classification of craniosynostosis, but this has never been studied systematically.

Methods

We tested the combinations of three different synthetic data sources: a statistical shape model (SSM), a generative adversarial network (GAN), and image-based principal component analysis for a convolutional neural network (CNN)-based classification of craniosynostosis. The CNN is trained only on synthetic data but is validated and tested on clinical data.

Results

The combination of an SSM and a GAN achieved an accuracy of 0.960 and an F1 score of 0.928 on the unseen test set. The difference to training on clinical data was smaller than 0.01. Including a second image modality improved classification performance for all data sources.

Conclusions

Without a single clinical training sample, a CNN was able to classify head deformities with similar accuracy as if it was trained on clinical data. Using multiple data sources was key for a good classification based on synthetic data alone. Synthetic data might play an important future role in the assessment of craniosynostosis.

A simulation-driven supervised learning framework to estimate brain microstructure using diffusion MRI

We propose a framework to train supervised learning models on synthetic data to estimate brain microstructure parameters using diffusion magnetic resonance imaging (dMRI). Although further validation is necessary, the proposed framework aims to seamlessly incorporate realistic simulations into dMRI microstructure estimation. Synthetic data were generated from over 1,000 neuron meshes converted from digital neuronal reconstructions and linked to their neuroanatomical parameters (such as soma volume and neurite length) using an optimized diffusion MRI simulator that produces intracellular dMRI signals from the solution of the Bloch-Torrey partial differential equation. By combining random subsets of simulated neuron signals with a free diffusion compartment signal, we constructed a synthetic dataset containing dMRI signals and 40 tissue microstructure parameters of 1.45 million artificial brain voxels. To implement supervised learning models we chose multilayer perceptrons (MLPs) and trained them on a subset of the synthetic dataset to estimate some microstructure parameters, namely, the volume fractions of soma, neurites, and the free diffusion compartment, as well as the area fractions of soma and neurites. The trained MLPs perform satisfactorily on the synthetic test sets and give promising in-vivo parameter maps on the MGH Connectome Diffusion Microstructure Dataset (CDMD). Most importantly, the estimated volume fractions showed low dependence on the diffusion time, the diffusion time independence of the estimated parameters being a desired property of quantitative microstructure imaging. The synthetic dataset we generated will be valuable for the validation of models that map between the dMRI signals and microstructure parameters. The surface meshes and microstructures parameters of the aforementioned neurons have been made publicly available.

3D-2D Distance Maps Conversion Enhances Classification of Craniosynostosis

OBJECTIVE

Diagnosis of craniosynostosis using photogrammetric 3D surface scans is a promising radiation-free alternative to traditional computed tomography. We propose a 3D surface scan to 2D distance map conversion enabling the usage of the first convolutional neural networks (CNNs)-based classification of craniosynostosis. Benefits of using 2D images include preserving patient anonymity, enabling data augmentation during training, and a strong under-sampling of the 3D surface with good classification performance.

METHODS

The proposed distance maps sample 2D images from 3D surface scans using a coordinate transformation, ray casting, and distance extraction. We introduce a CNN-based classification pipeline and compare our classifier to alternative approaches on a dataset of 496 patients. We investigate into low-resolution sampling, data augmentation, and attribution mapping.

RESULTS

Resnet18 outperformed alternative classifiers on our dataset with an F1-score of 0.964 and an accuracy of 98.4%. Data augmentation on 2D distance maps increased performance for all classifiers. Under-sampling allowed 256-fold computation reduction during ray casting while retaining an F1-score of 0.92. Attribution maps showed high amplitudes on the frontal head.

CONCLUSION

We demonstrated a versatile mapping approach to extract a 2D distance map from the 3D head geometry increasing classification performance, enabling data augmentation during training on 2D distance maps, and the usage of CNNs. We found that low-resolution images were sufficient for a good classification performance.

SIGNIFICANCE

Photogrammetric surface scans are a suitable craniosynostosis diagnosis tool for clinical practice. Domain transfer to computed tomography seems likely and can further contribute to reducing ionizing radiation exposure for infants.

A Radiation-Free Classification Pipeline for Craniosynostosis Using Statistical Shape Modeling

BACKGROUND

Craniosynostosis is a condition caused by the premature fusion of skull sutures, leading to irregular growth patterns of the head. Three-dimensional photogrammetry is a radiation-free alternative to the diagnosis using computed tomography. While statistical shape models have been proposed to quantify head shape, no shape-model-based classification approach has been presented yet.

METHODS

We present a classification pipeline that enables an automated diagnosis of three types of craniosynostosis. The pipeline is based on a statistical shape model built from photogrammetric surface scans. We made the model and pathology-specific submodels publicly available, making it the first publicly available craniosynostosis-related head model, as well as the first focusing on infants younger than 1.5 years. To the best of our knowledge, we performed the largest classification study for craniosynostosis to date.

RESULTS

Our classification approach yields an accuracy of 97.8 %, comparable to other state-of-the-art methods using both computed tomography scans and stereophotogrammetry. Regarding the statistical shape model, we demonstrate that our model performs similar to other statistical shape models of the human head.

CONCLUSION

We present a state-of-the-art shape-model-based classification approach for a radiation-free diagnosis of craniosynostosis. Our publicly available shape model enables the assessment of craniosynostosis on realistic and synthetic data.

A microRNA Cluster Controls Fat Cell Differentiation and Adipose Tissue Expansion By Regulating SNCG

The H19X-encoded miR-424(322)/503 cluster regulates multiple cellular functions. Here, it is reported for the first time that it is also a critical linchpin of fat mass expansion. Deletion of this miRNA cluster in mice results in obesity, while increasing the pool of early adipocyte progenitors and hypertrophied adipocytes. Complementary loss and gain of function experiments and RNA sequencing demonstrate that miR-424(322)/503 regulates a conserved genetic program involved in the differentiation and commitment of white adipocytes. Mechanistically, it is demonstrated that miR-424(322)/503 targets γ-Synuclein (SNCG), a factor that mediates this program rearrangement by controlling metabolic functions in fat cells, allowing adipocyte differentiation and adipose tissue enlargement. Accordingly, diminished miR-424(322) in mice and obese humans co-segregate with increased SNCG in fat and peripheral blood as mutually exclusive features of obesity, being normalized upon weight loss. The data unveil a previously unknown regulatory mechanism of fat mass expansion tightly controlled by the miR-424(322)/503 through SNCG.

Ventral striatal islands of Calleja neurons control grooming in mice

The striatum comprises multiple subdivisions and neural circuits that differentially control motor output. The islands of Calleja (IC) contain clusters of densely packed granule cells situated in the ventral striatum, predominantly in the olfactory tubercle (OT). Characterized by expression of the D3 dopamine receptor, the IC are evolutionally conserved, but have undefined functions. Here, we show that optogenetic activation of OT D3 neurons robustly initiates self-grooming in mice while suppressing other ongoing behaviors. Conversely, optogenetic inhibition of these neurons halts ongoing grooming, and genetic ablation reduces spontaneous grooming. Furthermore, OT D3 neurons show increased activity before and during grooming and influence local striatal output via synaptic connections with neighboring OT neurons (primarily spiny projection neurons), whose firing rates display grooming-related modulation. Our study uncovers a new role of the ventral striatum's IC in regulating motor output and has important implications for the neural control of grooming.

A new giant sauropod, Australotitan cooperensis gen. et sp. nov., from the mid-Cretaceous of Australia

A new giant sauropod, Australotitan cooperensis gen. et sp. nov., represents the first record of dinosaurs from the southern-central Winton Formation of the Eromanga Basin, Australia. We estimate the type locality to be 270–300 m from the base of the Winton Formation and compare this to the semi-contemporaneous sauropod taxa, Diamantinasaurus matildaeHocknull et al., 2009, Wintonotitan wattsiHocknull et al., 2009 and Savannasaurus elliottorumPoropat et al., 2016. The new titanosaurian is the largest dinosaur from Australia as represented by osteological remains and based on limb-size comparisons it reached a size similar to that of the giant titanosaurians from South America. Using 3-D surface scan models we compare features of the appendicular skeleton that differentiate Australotitan cooperensis gen. et sp. nov. as a new taxon. A key limitation to the study of sauropods is the inability to easily and directly compare specimens. Therefore, 3-D cybertypes have become a more standard way to undertake direct comparative assessments. Uncoloured, low resolution, and uncharacterized 3-D surface models can lead to misinterpretations, in particular identification of pre-, syn- and post-depositional distortion. We propose a method for identifying, documenting and illustrating these distortions directly onto the 3-D geometric surface of the models using a colour reference scheme. This new method is repeatable for researchers when observing and documenting specimens including taphonomic alterations and geometric differences. A detailed comparative and preliminary computational phylogenetic assessment supports a shared ancestry for all four Winton Formation taxa, albeit with limited statistical support. Palaeobiogeographical interpretations from these resultant phylogenetic hypotheses remain equivocal due to contrary Asian and South American relationships with the Australian taxa. Temporal and palaeoenvironmental differences between the northern and southern-central sauropod locations are considered to explain the taxonomic and morphological diversity of sauropods from the Winton Formation. Interpretations for this diversity are explored, including an eco-morphocline and/or chronocline across newly developed terrestrial environments as the basin fills.

The quantitative influence of current treatment options on patellofemoral stability in patients with trochlear dysplasia and symptomatic patellofemoral instability - a finite element simulation

BACKGROUND

Trochlear dysplasia is highly associated with patellofemoral instability. The goal of conservative and surgical treatment is to stabilize the patella while minimizing adverse effects. However, there is no literature investigating the quantitative influence of different treatment options on patellofemoral stability in knees with trochlear dysplasia. We created and exploited a range of finite element models to address this gap in knowledge.

METHODS

MRI data of 5 knees with trochlear dysplasia and symptomatic patellofemoral instability were adapted into this previously established model. Vastus medialis obliquus strengthening as well as double-bundle medial patellofemoral ligament reconstruction and the combination of medial patellofemoral ligament reconstruction and trochleoplasty were simulated. The force necessary to dislocate the patella by 10 mm and fully dislocate the patella was calculated in different flexion angles.

FINDINGS

Our model predicts a significant increase of patellofemoral stability at the investigated flexion angles (0°-45°) for a dislocation of 10 mm and a full dislocation after medial patellofemoral ligament reconstruction and the combination of medial patellofemoral ligament reconstruction and trochleoplasty compared to trochleodysplastic (P = 0.01) and healthy knees (P = 0.01-0.02). Vastus medialis obliquus strengthening has a negligible effect on patellofemoral stability.

INTERPRETATIONS

This is the first objective quantitative biomechanical evidence supporting the place of medial patellofemoral ligament reconstruction and medial patellofemoral ligament reconstruction combined with trochleoplasty in patients with symptomatic patellofemoral instability and trochlear dysplasia type B. Vastus medialis obliquus strengthening has a negligible effect on patellar stability at a low total quadriceps load of 175 N.

Patellofemoral instability in trochleodysplastic knee joints and the quantitative influence of simulated trochleoplasty - A finite element simulation

BACKGROUND

Patellofemoral instability is a debilitating condition mainly affecting young patients and has been correlated with trochlear dysplasia. It can occur when the patella is insufficiently guided through its range of motion. Currently, there is no literature describing patellofemoral stability in trochleodysplastic knees and the effect of isolated trochleoplasty on patellofemoral stability.

METHODS

The effect of isolated trochleoplasty in trochleodysplastic knees of patients with symptomatic patellofemoral instability was investigated using a quasi-static finite element model. MRI data of five healthy knees were segmented, meshed and a finite element analysis was performed in order to validate the model. A second validation was performed by comparing simulated patellofemoral kinematics to in-vivo values obtained from upright- weight bearing CT scans. Subsequently, five trochleodysplastic knees were modelled before and after simulated trochleoplasty. The force necessary to dislocate the patella by 10 mm and to fully dislocate the patella was calculated in various knee flexion angles between 0 and 45°.

FINDINGS

The developed models successfully predicted outcome values within the range of reference values from literature. Lateral stability was significantly lower in trochleodysplastic knees compared to healthy knees. Trochleoplasty was determined to significantly increase the force necessary to dislocate the patella in trochleodysplastic knees to comparable values as in healthy knees.

INTERPRETATION

This is the first study to investigate lateral patellofemoral stability in patients with symptomatic patellofemoral instability and dysplasia of the trochlear groove. We confirm that patellofemoral stability is significantly lower in trochleodysplastic knees than in healthy knees. Trochleoplasty increases patellofemoral stability to levels similar to healthy.

The Natural Historian's Guide to the CT Galaxy: Step-by-Step Instructions for Preparing and Analyzing Computed Tomographic (CT) Data Using Cross-Platform, Open Access Software

The decreasing cost of acquiring computed tomographic (CT) data has fueled a global effort to digitize the anatomy of museum specimens. This effort has produced a wealth of open access digital three-dimensional (3D) models of anatomy available to anyone with access to the Internet. The potential applications of these data are broad, ranging from 3D printing for purely educational purposes to the development of highly advanced biomechanical models of anatomical structures. However, while virtually anyone can access these digital data, relatively few have the training to easily derive a desirable product (e.g., a 3D visualization of an anatomical structure) from them. Here, we present a workflow based on free, open source, cross-platform software for processing CT data. We provide step-by-step instructions that start with acquiring CT data from a new reconstruction or an open access repository, and progress through visualizing, measuring, landmarking, and constructing digital 3D models of anatomical structures. We also include instructions for digital dissection, data reduction, and exporting data for use in downstream applications such as 3D printing. Finally, we provide Supplementary Videos and workflows that demonstrate how the workflow facilitates five specific applications: measuring functional traits associated with feeding, digitally isolating anatomical structures, isolating regions of interest using semi-automated segmentation, collecting data with simple visual tools, and reducing file size and converting file type of a 3D model.

Augmented Reality Markerless Multi-Image Outdoor Tracking System for the Historical Buildings on Parliament Hill

Augmented Reality (AR) applications have experienced extraordinary growth recently, evolving into a well-established method for the dissemination and communication of content related to cultural heritage—including education. AR applications have been used in museums and gallery exhibitions and virtual reconstructions of historic interiors. However, the circumstances of an outdoor environment can be problematic. This paper presents a methodology to develop immersive AR applications based on the recognition of outdoor buildings. To demonstrate this methodology, a case study focused on the Parliament Buildings National Historic Site in Ottawa, Canada has been conducted. The site is currently undergoing a multiyear rehabilitation program that will make access to parts of this national monument inaccessible to the public. AR experiences, including simulated photo merging of historic and present content, are proposed as one tool that can enrich the Parliament Hill visit during the rehabilitation. Outdoor AR experiences are limited by factors, such as variable lighting (and shadows) conditions, caused by changes in the environment (objects height and orientation, obstructions, occlusions), the weather, and the time of day. This paper proposes a workflow to solve some of these issues from a multi-image tracking approach.

Designing and validating a PVA liver phantom with respiratory motion for needle-based interventions

Purpose

The purpose is to design and validate an anthropomorphic polyvinyl alcohol (PVA) liver phantom with respiratory motion to simulate needle-based interventions. Such a system can, for example, be used as a validation tool for novel needles.

Methods

Image segmentations of CT scans of four patients during inspiration and expiration were used to measure liver and rib displacement. An anthropomorphic liver mold based on a CT scan was 3D printed and filled with 5% w/w PVA-to-water, undergoing two freeze–thaw cycles, in addition to a 3D-printed compliant rib cage. They were both held in place by a PVA abdominal phantom. A sinusoidal motion vector, based on the measured liver displacement, was applied to the liver phantom by means of a motion stage. Liver, rib cage and needle deflection were tracked by placing electromagnetic sensors on the phantom. Liver and rib cage phantom motion was validated by comparison with the CT images of the patients, whereas needle deflection was compared with the literature.

Results

CT analysis showed that from the state of expiration to inspiration, the livers moved predominantly toward the right (mean: 2 mm, range: − 11 to 11 mm), anterior (mean: 15 mm, range: 9–21 mm) and caudal (mean: 16 mm, range: 6–24 mm) direction. The mechatronic design of the liver phantom gives the freedom to set direction and amplitude of the motion and was able to mimic the direction of liver motion of one patient. Needle deflection inside the phantom increased from 1.6 to 3.8 mm from the initial expiration state to inspiration.

Conclusions

The developed liver phantom allows for applying different motion patterns and shapes/sizes and thus allows for patient-specific simulation of needle-based interventions. Moreover, it is able to mimic appropriate respiratory motion and needle deflection as observed in patients.

Electronic supplementary material

The online version of this article (10.1007/s11548-019-02029-6) contains supplementary material, which is available to authorized users.

Good Random Multi-Triangulation of Surfaces

We introduce the Hierarchical Poisson Disk Sampling Multi-Triangulation (HPDS-MT) of surfaces, a novel structure that combines the power of multi-triangulation (MT) with the benefits of Hierarchical Poisson Disk Sampling (HPDS). MT is a general framework for representing surfaces through variable resolution triangle meshes, while HPDS is a well-spaced random distribution with blue noise characteristics. The distinguishing feature of the HPDS-MT is its ability to extract adaptive meshes whose triangles are guaranteed to have good shape quality. The key idea behind the HPDS-MT is a preprocessed hierarchy of points, which is used in the construction of a MT via incremental simplification. In addition to proving theoretical properties on the shape quality of the triangle meshes extracted by the HPDS-MT, we provide an implementation that computes the HPDS-MT with high accuracy. Our results confirm the theoretical guarantees and outperform similar methods. We also prove that the Hausdorff distance between the original surface and any (extracted) adaptive mesh is bounded by the sampling distribution of the radii of Poisson-disks over the surface. Finally, we illustrate the advantages of the HPDS-MT in some typical problems of geometry processing.

Molecular Surface Remeshing with Local Region Refinement

Molecular surface mesh generation is a prerequisite for using the boundary element method (BEM) and finite element method (FEM) in implicit-solvent modeling. Molecular surface meshes typically have small angles, redundant vertices, and low-quality elements. In the implicit-solvent modeling of biomolecular systems it is usually required to improve the mesh quality and eliminate low-quality elements. Existing methods often fail to efficiently remove low-quality elements, especially in complex molecular meshes. In this paper, we propose a mesh refinement method that smooths the meshes, eliminates invalid regions in a cut-and-fill strategy, and improves the minimal angle. We compared our method with four different state-of-the-art methods and found that our method showed a significant improvement over state-of-the-art methods in minimal angle, aspect ratio, and other meshing quality measurements. In addition, our method showed satisfactory results in terms of the ratio of regular vertices and the preservation of area and volume.

Asteroid modeling for testing spacecraft approach and landing

Spacecraft exploration of asteroids presents autonomous-navigation challenges that can be aided by virtual models to test and develop guidance and hazard-avoidance systems. Researchers have extended and applied graphics techniques to create high-resolution asteroid models to simulate cameras and other spacecraft sensors approaching and descending toward asteroids. A scalable model structure with evenly spaced vertices simplifies terrain modeling, avoids distortion at the poles, and enables triangle-strip definition for efficient rendering. To create the base asteroid models, this approach uses two-phase Poisson faulting and Perlin noise. It creates realistic asteroid surfaces by adding both crater models adapted from lunar terrain simulation and multiresolution boulders. The researchers evaluated the virtual asteroids by comparing them with real asteroid images, examining the slope distributions, and applying a surface-relative feature-tracking algorithm to the models.

Interactive applications for sketch-based editable polycube map

In this paper, we propose a sketch-based editable polycube mapping method that, given a general mesh and a simple polycube that coarsely resembles the shape of the object, plus sketched features indicating relevant correspondences between the two, provides a uniform, regular, and user-controllable quads-only mesh that can be used as a basis structure for subdivision. Large scale models with complex geometry and topology can be processed efficiently with simple, intuitive operations. We show that the simple, intuitive nature of the polycube map is a substantial advantage from the point of view of the interface by demonstrating a series of applications, including kit-basing, shape morphing, painting over the parameterization domain, and GPU-friendly tessellated subdivision displacement, where the user is also able to control the number of patches in the base mesh by the construction of the base polycube.

Efficient optimization of common base domains for cross parameterization

Given a set of corresponding user-specified anchor points on a pair of models having similar features and topologies, the cross parameterization technique can establish a bijective mapping constrained by the anchor points. In this paper, we present an efficient algorithm to optimize the complexes and the shape of common base domains in cross parameterization for reducing the distortion of the bijective mapping. The optimization is also constrained by the anchor points. We investigate a new signature, Length-Preserved Base Domain (LPBD), for measuring the level of stretch between surface patches in cross parameterization. This new signature well balances the accuracy of measurement and the computational speed. Based on LPBD, a set of metrics are studied and compared. The best ones are employed in our domain optimization algorithm that consists of two major operators, boundary swapping and patch merging. Experimental results show that our optimization algorithm can reduce the distortion in cross parameterization efficiently.

An interactive local flattening operator to support digital investigations on artwork surfaces

Analyzing either high-frequency shape detail or any other 2D fields (scalar or vector) embedded over a 3D geometry is a complex task, since detaching the detail from the overall shape can be tricky. An alternative approach is to move to the 2D space, resolving shape reasoning to easier image processing techniques. In this paper we propose a novel framework for the analysis of 2D information distributed over 3D geometry, based on a locally smooth parametrization technique that allows us to treat local 3D data in terms of image content. The proposed approach has been implemented as a sketch-based system that allows to design with a few gestures a set of (possibly overlapping) parameterizations of rectangular portions of the surface. We demonstrate that, due to the locality of the parametrization, the distortion is under an acceptable threshold, while discontinuities can be avoided since the parametrized geometry is always homeomorphic to a disk. We show the effectiveness of the proposed technique to solve specific Cultural Heritage (CH) tasks: the analysis of chisel marks over the surface of a unfinished sculpture and the local comparison of multiple photographs mapped over the surface of an artwork. For this very difficult task, we believe that our framework and the corresponding tool are the first steps toward a computer-based shape reasoning system, able to support CH scholars with a medium they are more used to.