Structure Of Riemannian Manifold Research Articles

The Covariance Metric in the Blaschke Locus

We prove that the Blaschke locus has the structure of a finite dimensional smooth manifold away from the Teichmüller space and study its Riemannian manifold structure with respect to the covariance metric introduced by Guillarmou, Knieper and Lefeuvre in Guillarmou et al. in (Ergod Theory Dyn Syst 43:974–1022, 2021). We also identify some families of geodesics in the Blaschke locus arising from Hitchin representations for orbifolds and show that they have infinite length with respect to the covariance metric.

Automatic Seizure Detection using Logarithmic Euclidean-Gaussian Mixture Models (LE-GMMs) and Improved Deep Forest Learning.

Automatic seizure detection could facilitate early detection, improve treatment planning, and reduce medical workload. This study describes a novel Logarithmic Euclidean-Gaussian Mixture Models (LE-GMMs) and an improved Deep Forest learning algorithm for epileptic seizure detection. The LE-GMMs could map the Riemannian manifold structure of Gaussian models to linear Euclidean space, which fully exploits the ability of GMMs to distinguish non-seizure and seizure EEG signals. The Multi-Pooling and error Screening Forest (MPSForest) learning method based on Deep Forest uses multi-pooling and out-of-bagging (OOB) error screening to reduce memory load and random tree construction. Firstly, variational modal decomposition (VMD) is applied to decompose electroencephalogram (EEG) signals into five layers, and the first three layers are chosen to construct EEG time-frequency distribution. Then Gaussian Mixture Models are estimated, and the LE-GMMs are constructed to extract valid EEG features. These features are input into the MPSForest model to classify seizure and non-seizure samples. After that, the outputs are subjected to post-processing to get the final seizure detection results, including moving average filtering and the adaptive collar technique. The proposed method achieves average sensitivity of 98.22% and specificity of 98.99% on the UPenn and Mayo Clinic dataset, and for the long-term Freiburg EEG dataset with 21 patients, the sensitivity of 98.47% and specificity of 98.57% are yielded respectively with the false detection rate of 0.24/h. The experimental results show that this proposed method has excellent accuracy in distinguishing non-seizure and seizure EEG signals and holds great potential for clinical research and diagnostics.

Energy-Efficient Power Control and Beamforming for Reconfigurable Intelligent Surface-Aided Uplink IoT Networks

Recently, reconfigurable intelligent surface (RIS), a planar metasurface consisting of a large number of low-cost reflecting elements, has received much attention due to its ability to improve both the spectrum and energy efficiencies by reconfiguring the wireless propagation environment. In this paper, we propose an RIS phase shift and BS beamforming optimization technique that minimizes the uplink transmit power of the RIS-aided IoT network. Key idea of the proposed scheme, referred to as Riemannian conjugate gradient-based joint optimization (RCG-JO), is to jointly optimize the RIS phase shifts and the BS beamforming vectors using the Riemannian conjugate gradient technique. By exploiting the product Riemannian manifold structure of the sets of unit-modulus phase shifts and unit-norm beamforming vectors, we convert the nonconvex uplink power minimization problem into the unconstrained problem and then find out the optimal solution over the product Riemannian manifold. From the performance analysis and numerical evaluations, we demonstrate that the proposed RCG-JO technique achieves 94% reduction of the uplink transmit power over the conventional scheme without RIS.

Riemannian block SPD coupling manifold and its application to optimal transport

In this work, we study the optimal transport (OT) problem between symmetric positive definite (SPD) matrix-valued measures. We formulate the above as a generalized optimal transport problem where the cost, the marginals, and the coupling are represented as block matrices and each component block is a SPD matrix. The summation of row blocks and column blocks in the coupling matrix are constrained by the given block-SPD marginals. We endow the set of such block-coupling matrices with a novel Riemannian manifold structure. This allows to exploit the versatile Riemannian optimization framework to solve generic SPD matrix-valued OT problems. We illustrate the usefulness of the proposed approach in several applications.

Data-driven modeling with fuzzy sets and manifolds

The manifold hypothesis states that the shape of the observed data is relatively simple and that it lies on a low-dimensional manifold embedded in a high-dimensional space. We contribute to the problem of data-driven modeling by treating it as an inverse problem where the model defines a Euclidean space with a Riemannian manifold structure. In particular, our contribution shows that a fuzzy set on a bounded support defines a Riemannian manifold that can be embedded in a multidimensional space where dimension is a model parameter. A noticeable advantage of the proposed approach is its connection with the values of the membership function and independence from the dimension of the data being modeled.Last but not least, we have found various formal representations of the Laplace-Beltrami operator and use its values as an estimate of the quality of the approximate solution to the inverse problem.

Model-Centric Data Manifold: The Data Through the Eyes of the Model

We discover that deep ReLU neural network classifiers can see a low-dimensional Riemannian manifold structure on data. Such structure comes via the local data matrix, a variation of the Fisher information matrix, where the role of the model parameters is taken by the data variables. We obtain a foliation of the data domain and we show that the dataset on which the model is trained lies on a leaf, the data leaf, whose dimension is bounded by the number of classification labels. We validate our results with some experiments with the MNIST dataset: paths on the data leaf connect valid images, while other leaves cover noisy images.

Nested Grassmannians for Dimensionality Reduction with Applications

In the recent past, nested structure of Riemannian manifolds has been studied in the context of dimensionality reduction as an alternative to the popular principal geodesic analysis (PGA) technique, for example, the principal nested spheres. In this paper, we propose a novel framework for constructing a nested sequence of homogeneous Riemannian manifolds. Common examples of homogeneous Riemannian manifolds include the spheres, the Stiefel manifolds, and the Grassmann manifolds. In particular, we focus on applying the proposed framework to the Grassmann manifolds, giving rise to the nested Grassmannians (NG). An important application in which Grassmann manifolds are encountered is planar shape analysis. Specifically, each planar (2D) shape can be represented as a point in the complex projective space which is a complex Grassmann manifold. Some salient features of our framework are: (i) it explicitly exploits the geometry of the homogeneous Riemannain manifolds and (ii) the nested lower-dimensional submanifolds need not be geodesic. With the proposed NG structure, we develop algorithms for the supervised and unsupervised dimensionality reduction problems respectively. The proposed algorithms are compared with PGA via simulation studies and real data experiments and are shown to achieve a higher ratio of expressed variance compared to PGA.<br>The code is available at <a href='https://github.com/cvgmi/NestedGrassmann'>https://github.com/cvgmi/NestedGrassmann</a>

Manifold Optimization for High-Accuracy Spatial Location Estimation Using Ultrasound Waves

This paper designs a high accuracy spatial location estimation method using ultrasound waves by exploiting the fixed geometry of the transmitters. Assuming an equilateral triangle antenna configuration, where three antennas are placed as the vertices of an equilateral triangle, the spatial location problem can be formulated as a non-convex optimization problem whose interior is shown to admit a Riemannian manifold structure. The investigation of the geometry of the newly introduced manifold, i.e. the manifold of all equilateral triangles in R^3, allows the design of highly efficient optimization algorithms. Simulation results are presented to compare the performance of the proposed approach against popular methods from the literature. The results suggest that the proposed Riemannian-based methods outperform the state-of-the-art methods. Furthermore, the proposed Riemannian methods require much smaller computation time as compared with popular generic non-convex approaches.

Metric connections with parallel skew-symmetric torsion

A geometry with parallel skew-symmetric torsion is a Riemannian manifold carrying a metric connection with parallel skew-symmetric torsion. Besides the trivial case of the Levi-Civita connection, geometries with non-vanishing parallel skew-symmetric torsion arise naturally in several geometric contexts, e.g. on naturally reductive homogeneous spaces, nearly Kähler or nearly parallel G2-manifolds, Sasakian and 3-Sasakian manifolds, or twistor spaces over quaternion-Kähler manifolds with positive scalar curvature. In this paper we study the local structure of Riemannian manifolds carrying a metric connection with parallel skew-symmetric torsion. On every such manifold one can define a natural splitting of the tangent bundle which gives rise to a Riemannian submersion over a geometry with parallel skew-symmetric torsion of smaller dimension endowed with some extra structure. We show how previously known examples of geometries with parallel skew-symmetric torsion fit into this pattern, and construct several new examples. In the particular case where the above Riemannian submersion has the structure of a principal bundle, we give the complete local classification of the corresponding geometries with parallel skew-symmetric torsion.

Dilated Convolutional Neural Networks for Sequential Manifold-valued Data.

Efforts are underway to study ways via which the power of deep neural networks can be extended to non-standard data types such as structured data (e.g., graphs) or manifold-valued data (e.g., unit vectors or special matrices). Often, sizable empirical improvements are possible when the geometry of such data spaces are incorporated into the design of the model, architecture, and the algorithms. Motivated by neuroimaging applications, we study formulations where the data are sequential manifold-valued measurements. This case is common in brain imaging, where the samples correspond to symmetric positive definite matrices or orientation distribution functions. Instead of a recurrent model which poses computational/technical issues, and inspired by recent results showing the viability of dilated convolutional models for sequence prediction, we develop a dilated convolutional neural network architecture for this task. On the technical side, we show how the modules needed in our network can be derived while explicitly taking the Riemannian manifold structure into account. We show how the operations needed can leverage known results for calculating the weighted Fréchet Mean (wFM). Finally, we present scientific results for group difference analysis in Alzheimer's disease (AD) where the groups are derived using AD pathology load: here the model finds several brain fiber bundles that are related to AD even when the subjects are all still cognitively healthy.

The Geometry of the Generalized Gamma Manifold and an Application to Medical Imaging

The Fisher information metric provides a smooth family of probability measures with a Riemannian manifold structure, which is an object in information geometry. The information geometry of the gamma manifold associated with the family of gamma distributions has been well studied. However, only a few results are known for the generalized gamma family that adds an extra shape parameter. The present article gives some new results about the generalized gamma manifold. This paper also introduces an application in medical imaging that is the classification of Alzheimer’s disease population. In the medical field, over the past two decades, a growing number of quantitative image analysis techniques have been developed, including histogram analysis, which is widely used to quantify the diffuse pathological changes of some neurological diseases. This method presents several drawbacks. Indeed, all the information included in the histogram is not used and the histogram is an overly simplistic estimate of a probability distribution. Thus, in this study, we present how using information geometry and the generalized gamma manifold improved the performance of the classification of Alzheimer’s disease population.

A Kernel Regression Procedure in the 3D Shape Space with an Application to Online Sales of Children’s Wear

This paper is focused on kernel regression when the response variable is the shape of a 3D object represented by a configuration matrix of landmarks. Regression methods on this shape space are not trivial because this space has a complex finite-dimensional Riemannian manifold structure (non-Euclidean). Papers about it are scarce in the literature, the majority of them are restricted to the case of a single explanatory variable, and many of them are based on the approximated tangent space. In this paper, there are several methodological innovations. The first one is the adaptation of the general method for kernel regression analysis in manifold-valued data to the three-dimensional case of Kendall’s shape space. The second one is its generalization to the multivariate case and the addressing of the curse-of-dimensionality problem. Finally, we propose bootstrap confidence intervals for prediction. A simulation study is carried out to check the goodness of the procedure, and a comparison with a current approach is performed. Then, it is applied to a 3D database obtained from an anthropometric survey of the Spanish child population with a potential application to online sales of children’s wear.

A Riemannian trust‐region method for low‐rank tensor completion

SummaryThe goal of tensor completion is to fill in missing entries of a partially known tensor (possibly including some noise) under a low‐rank constraint. This may be formulated as a least‐squares problem. The set of tensors of a given multilinear rank is known to admit a Riemannian manifold structure; thus, methods of Riemannian optimization are applicable. In our work, we derive the Riemannian Hessian of an objective function on the low‐rank tensor manifolds using the Weingarten map, a concept from differential geometry. We discuss the convergence properties of Riemannian trust‐region methods based on the exact Hessian and standard approximations, both theoretically and numerically. We compare our approach with Riemannian tensor completion methods from recent literature, both in terms of convergence behavior and computational complexity. Our examples include the completion of randomly generated data with and without noise and the recovery of multilinear data from survey statistics.

Human Motion Segmentation via Robust Kernel Sparse Subspace Clustering.

Studies on human motion have attracted a lot of attentions. Human motion capture data, which much more precisely records human motion than videos do, has been widely used in many areas. Motion segmentation is an indispensable step for many related applications, but current segmentation methods for motion capture data do not effectively model some important characteristics of motion capture data, such as Riemannian manifold structure and containing non-Gaussian noise. In this paper, we convert the segmentation of motion capture data into a temporal subspace clustering problem. Under the framework of sparse subspace clustering, we propose to use the geodesic exponential kernel to model the Riemannian manifold structure, use correntropy to measure the reconstruction error, use the triangle constraint to guarantee temporal continuity in each cluster and use multi-view reconstruction to extract the relations between different joints. Therefore, exploiting some special characteristics of motion capture data, we propose a new segmentation method, which is robust to non-Gaussian noise, since correntropy is a localized similarity measure. We also develop an efficient optimization algorithm based on block coordinate descent method to solve the proposed model. Our optimization algorithm has a linear complexity while sparse subspace clustering is originally a quadratic problem. Extensive experiment results both on simulated noisy data set and real noisy data set demonstrate the advantage of the proposed method.Studies on human motion have attracted a lot of attentions. Human motion capture data, which much more precisely records human motion than videos do, has been widely used in many areas. Motion segmentation is an indispensable step for many related applications, but current segmentation methods for motion capture data do not effectively model some important characteristics of motion capture data, such as Riemannian manifold structure and containing non-Gaussian noise. In this paper, we convert the segmentation of motion capture data into a temporal subspace clustering problem. Under the framework of sparse subspace clustering, we propose to use the geodesic exponential kernel to model the Riemannian manifold structure, use correntropy to measure the reconstruction error, use the triangle constraint to guarantee temporal continuity in each cluster and use multi-view reconstruction to extract the relations between different joints. Therefore, exploiting some special characteristics of motion capture data, we propose a new segmentation method, which is robust to non-Gaussian noise, since correntropy is a localized similarity measure. We also develop an efficient optimization algorithm based on block coordinate descent method to solve the proposed model. Our optimization algorithm has a linear complexity while sparse subspace clustering is originally a quadratic problem. Extensive experiment results both on simulated noisy data set and real noisy data set demonstrate the advantage of the proposed method.

Keyframe Extraction for Human Motion Capture Data Based on Joint Kernel Sparse Representation

Human motion capture data, which are used to animate animation characters, have been widely used in many areas. To satisfy the high-precision requirement, human motion data are captured with a high frequency (120 frames/s) by a high-precision capture system. However, the high frequency and nonlinear structure make the storage, retrieval, and browsing of motion data challenging problems, which can be solved by keyframe extraction. Current keyframe extraction methods do not properly model two important characteristics of motion data, i.e., sparseness and Riemannian manifold structure. Therefore, we propose a new model called joint kernel sparse representation (SR), which is in marked contrast to all current keyframe extraction methods for motion data and can simultaneously model the sparseness and the Riemannian manifold structure. The proposed model completes the SR in a kernel-induced space with a geodesic exponential kernel, whereas the traditional SR cannot model the nonlinear structure of motion data in the Euclidean space. Meanwhile, because of several important modifications to traditional SR, our model can also exploit the relations between joints and solve two problems, i.e., the unreasonable distribution and redundancy of extracted keyframes, which current methods do not solve. Extensive experiments demonstrate the effectiveness of the proposed method.

Isoperimetric structure of asymptotically conical manifolds

We study the isoperimetric structure of Riemannian manifolds that are asymptotic to cones with non-negative Ricci curvature. Specifically, we generalize to this setting the seminal results of G. Huisken and S.-T. Yau on the existence of a canonical foliation by volume preserving stable constant mean curvature surfaces at infinity of asymptotically flat manifolds as well as the results of the second-named author with S. Brendle and J. Metzger on the isoperimetric structure of asymptotically flat manifolds. We also include an observation on the isoperimetric cone angle of such manifolds. This result is a natural analogue of the positive mass theorem in this setting.

Duct heights inferred from radar sea clutter using proper orthogonal bases

Maritime electromagnetic (EM) based communication and detection systems are strongly influenced by meteorological conditions, as they can cause anomalous electromagnetic propagation within the surface layer. To predict the performance of such systems, detailed knowledge of the refractivity profile is required. In recent years, refractivity from clutter (RFC) methods have been developed to estimate this refractivity profile by measuring radar clutter return from the rough ocean surface. The current work proposes an RFC framework that utilizes a novel surrogate model for EM propagation. The surrogate model is based on an offline created library of sparsely sampled field data of clutter returns, compressed into proper orthogonal bases, and indexed on specific surface layer refractive parameters. By exploiting the Riemannian manifold structure of the space that proper orthogonal bases occur in, we are able to interpolate among them. This, then, enables us to use the surrogate model in an inverse problem setting, whose goal is to uncover in situ maritime EM propagation conditions efficiently. We demonstrate the feasibility of our proposed surrogate model-based RFC approach for evaporation duct by testing it with field data obtained from an experimental campaign.

Manifold learning on brain functional networks in aging.

We propose a new analysis framework to utilize the full information of brain functional networks for computing the mean of a set of brain functional networks and embedding brain functional networks into a low-dimensional space in which traditional regression and classification analyses can be easily employed. For this, we first represent the brain functional network by a symmetric positive matrix computed using sparse inverse covariance estimation. We then impose a Log-Euclidean Riemannian manifold structure on brain functional networks whose norm gives a convenient and practical way to define a mean. Finally, based on the fact that the computation of linear operations can be done in the tangent space of this Riemannian manifold, we adopt Locally Linear Embedding (LLE) to the Log-Euclidean Riemannian manifold space in order to embed the brain functional networks into a low-dimensional space. We show that the integration of the Log-Euclidean manifold with LLE provides more efficient and succinct representation of the functional network and facilitates regression analysis, such as ridge regression, on the brain functional network to more accurately predict age when compared to that of the Euclidean space of functional networks with LLE. Interestingly, using the Log-Euclidean analysis framework, we demonstrate the integration and segregation of cortical-subcortical networks as well as among the salience, executive, and emotional networks across lifespan.

The $$k$$ k -means algorithm for 3D shapes with an application to apparel design

Clustering of objects according to shapes is of key importance in many scientific fields. In this paper we focus on the case where the shape of an object is represented by a configuration matrix of landmarks. It is well known that this shape space has a finite-dimensional Riemannian manifold structure (non-Euclidean) which makes it difficult to work with. Papers about clustering on this space are scarce in the literature. The basic foundation of the $$k$$k-means algorithm is the fact that the sample mean is the value that minimizes the Euclidean distance from each point to the centroid of the cluster to which it belongs, so, our idea is integrating the Procrustes type distances and Procrustes mean into the $$k$$k-means algorithm to adapt it to the shape analysis context. As far as we know, there have been just two attempts in that way. In this paper we propose to adapt the classical $$k$$k-means Lloyd algorithm to the context of Shape Analysis, focusing on the three dimensional case. We present a study comparing its performance with the Hartigan-Wong $$k$$k-means algorithm, one that was previously adapted to the field of Statistical Shape Analysis. We demonstrate the better performance of the Lloyd version and, finally, we propose to add a trimmed procedure. We apply both to a 3D database obtained from an anthropometric survey of the Spanish female population conducted in this country in 2006. The algorithms presented in this paper are available in the Anthropometry R package, whose most current version is always available from the Comprehensive R Archive Network.

A global approach for image orientation using Lie algebraic rotation averaging and convex L∞ minimisation

Abstract. In this paper we present a new global image orientation approach for a set of multiple overlapping images with given homologous point tuples which is based on a two-step procedure. The approach is independent on initial values, robust with respect to outliers and yields the global minimum solution under relatively mild constraints. The first step of the approach consists of the estimation of global rotation parameters by averaging relative rotation estimates for image pairs (these are determined from the homologous points via the essential matrix in a pre-processing step). For the averaging we make use of algebraic group theory in which rotations, as part of the special orthogonal group SO(3), form a Lie group with a Riemannian manifold structure. This allows for a mapping to the local Euclidean tangent space of SO(3), the Lie algebra. In this space the redundancy of relative orientations is used to compute an average of the absolute rotation for each image and furthermore to detect and eliminate outliers. In the second step translation parameters and the object coordinates of the homologous points are estimated within a convex L∞ optimisation, in which the rotation parameters are kept fixed. As an optional third step the results can be used as initial values for a final bundle adjustment that does not suffer from bad initialisation and quickly converges to a globally optimal solution. We investigate our approach for global image orientation based on synthetic data. The results are compared to a robust least squares bundle adjustment. In this way we show that our approach is independent of initial values and more robust against outliers than a conventional bundle adjustment.