Continuous Generalized Procrustes analysis

Two-dimensional shape models have been successfully applied to solve many problems in computer vision, such as object tracking, recognition, and segmentation. Typically, 2D shape models are learned from a discrete set of image landmarks (corresponding to projection of 3D points of an object), after applying Generalized Procustes Analysis (GPA) to remove 2D rigid transformations. However, the standard GPA process suffers from three main limitations. Firstly, the 2D training samples do not necessarily cover a uniform sampling of all the 3D transformations of an object. This can bias the estimate of the shape model. Secondly, it can be computationally expensive to learn the shape model by sampling 3D transformations. Thirdly, standard GPA methods use only one reference shape, which can might be insufficient to capture large structural variability of some objects.To address these drawbacks, this paper proposes continuous generalized Procrustes analysis (CGPA). CGPA uses a continuous formulation that avoids the need to generate 2D projections from all the rigid 3D transformations. It builds an efficient (in space and time) non-biased 2D shape model from a set of 3D model of objects. A major challenge in CGPA is the need to integrate over the space of 3D rotations, especially when the rotations are parameterized with Euler angles. To address this problem, we introduce the use of the Haar measure. Finally, we extended CGPA to incorporate several reference shapes. Experimental results on synthetic and real experiments show the benefits of CGPA over GPA.     

Procrustes Analysis with Deformations: A Closed-Form Solution by Eigenvalue Decomposition

Generalized Procrustes Analysis (GPA) is the problem of bringing multiple shapes into a common reference by estimating transformations. GPA has been extensively studied for the Euclidean and affine transformations. We introduce GPA with deformable transformations, which forms a much wider and difficult problem. We specifically study a class of transformations called the Linear Basis Warps, which contains the affine transformation and most of the usual deformation models, such as the Thin-Plate Spline (TPS). GPA with deformations is a nonconvex underconstrained problem. We resolve the fundamental ambiguities of deformable GPA using two shape constraints requiring the eigenvalues of the shape covariance. These eigenvalues can be computed independently as a prior or posterior. We give a closed-form and optimal solution to deformable GPA based on an eigenvalue decomposition. This solution handles regularization, favoring smooth deformation fields. It requires the transformation model to satisfy a fundamental property of free-translations, which asserts that the model can implement any translation. We show that this property fortunately holds true for most common transformation models, including the affine and TPS models. For the other models, we give another closed-form solution to GPA, which agrees exactly with the first solution for models with free-translation. We give pseudo-code for computing our solution, leading to the proposed DefGPA method, which is fast, globally optimal and widely applicable. We validate our method and compare it to previous work on six diverse 2D and 3D datasets, with special care taken to choose the hyperparameters from cross-validation.

     

Detection and classification of topological evolution for linear metamorphosis

The advantage of functional methods for shape metamorphosis is the automatic generation of intermediate shapes possible between the key shapes of different topology types. However, functional methods have a serious problem: shape interpolation is applied without topological information and thereby the time values of topological changes are not known. Thus, it is difficult to identify the time intervals for key frames of shape metamorphosis animation that faithfully visualize the topological evolution. Moreover, information on the types of topological changes is missing. To overcome the problem, we apply topological analysis to functional linear shape metamorphosis and classify the type of topological evolution by using a Hessian matrix. Our method is based on Morse theory and analyzes how the critical points appear. We classify the detected critical points into maximum point, minimum point, and saddle point types. Using the types of critical points, we can define the topological information for shape metamorphosis. We illustrate these methods using shape metamorphosis in 2D and 3D spaces.     

Bayes Reconstruction of Missing Teeth

We propose a method for restoring the surface of tooth crowns in a 3D model of a human denture, so that the pose and anatomical features of the tooth will work well for chewing. This is achieved by including information about the position and anatomy of the other teeth in the mouth. Our system contains two major parts: A statistical model of a selection of tooth shapes and a reconstruction of missing data. We use a training set consisting of 3D scans of dental cast models obtained with a laser scanner, and we have build a model of the shape variability of the teeth, their neighbors, and their antagonists, using the eigenstructure of the covariance matrix, also known as Principle Component Analysis (PCA). PCA is equivalent to fitting a multivariate Gaussian distribution to the data and the principle directions constitute a linear model for stochastic data and is used both for a data reduction or equivalently noise elimination and for data analysis. However for small sets of high dimensional data, the log-likelihood estimator for the covariance matrix is often far from convergence, and therefore reliable models must be obtained by use of prior information. We propose a natural and intrinsic regularization of the log-likelihood estimate based on differential geometrical properties of teeth surfaces, and we show general conditions under which this may be considered a Bayes prior. Finally we use Bayes method to propose the reconstruction of missing data, for e.g. finding the most probable shape of a missing tooth based on the best match with our shape model on the known data, and we superior improved reconstructions of our full system.     

2D‐D Lifting for Shape Reconstruction

We present an algorithm for shape reconstruction from incomplete 3D scans by fusing together two acquisition modes: 2D photographs and 3D scans. The two modes exhibit complementary characteristics: scans have depth information, but are often sparse and incomplete; photographs, on the other hand, are dense and have high resolution, but lack important depth information. In this work we fuse the two modes, taking advantage of their complementary information, to enhance 3D shape reconstruction from an incomplete scan with a 2D photograph. We compute geometrical and topological shape properties in 2D photographs and use them to reconstruct a shape from an incomplete 3D scan in a principled manner. Our key observation is that shape properties such as boundaries, smooth patches and local connectivity, can be inferred with high confidence from 2D photographs. Thus, we register the 3D scan with the 2D photograph and use scanned points as 3D depth cues for lifting 2D shape structures into 3D. Our contribution is an algorithm which significantly regularizes and enhances the problem of 3D reconstruction from partial scans by lifting 2D shape structures into 3D. We evaluate our algorithm on various shapes which are loosely scanned and photographed from different views, and compare them with state‐of‐the‐art reconstruction methods.     

自监督深度残差函数映射网络的3维模型对应关系计算

目的 针对传统非刚性3维模型的对应关系计算方法需要模型间真实对应关系监督的缺点,提出一种自监督深度残差函数映射网络（self-supervised deep residual functional maps network,SSDRFMN）。方法 首先将局部坐标系与直方图结合以计算3维模型的特征描述符,即方向直方图签名（signature of histograms of orientations,SHOT）描述符;其次将源模型与目标模型的SHOT描述符输入SSDRFMN,利用深度函数映射（deep functional maps,DFM）层计算两个模型间的函数映射矩阵,并通过模糊对应层将函数映射关系转换为点到点的对应关系;最后利用自监督损失函数计算模型间的测地距离误差,对计算出的对应关系进行评估。结果 实验结果表明,在MPI-FAUST数据集上,本文算法相比于有监督的深度函数映射（supervised deep functional maps,SDFM）算法,人体模型对应关系的测地误差减小了1.45;相比于频谱上采样（spectral upsampling,SU）算法减小了1.67。在TOSCA数据集上,本文算法相比于SDFM算法,狗、猫和狼等模型的对应关系的测地误差分别减小了3.13、0.98和1.89;相比于SU算法分别减小了2.81、2.22和1.11,并有效克服了已有深度函数映射方法需要模型间的真实对应关系来监督的缺点,使得该方法可以适用于不同的数据集,可扩展性大幅增强。结论 本文通过自监督深度残差函数映射网络训练模型的方向直方图签名描述符,提升了模型对应关系的准确率。本文方法可以适应于不同的数据集,相比传统方法,普适性较好。     

Estimating 3D human shapes from measurements

Recent advances in 3D imaging technologies give rise to databases of human shapes, from which statistical shape models can be built. These statistical models represent prior knowledge of the human shape and enable us to solve shape reconstruction problems from partial information. Generating human shape from traditional anthropometric measurements is such a problem, since these 1D measurements encode 3D shape information. Combined with a statistical shape model, these easy-to-obtain measurements can be leveraged to create 3D human shapes. However, existing methods limit the creation of the shapes to the space spanned by the database and thus require a large amount of training data. In this paper, we introduce a technique that extrapolates the statistically inferred shape to fit the measurement data using non-linear optimization. This method ensures that the generated shape is both human-like and satisfies the measurement conditions. We demonstrate the effectiveness of the method and compare it to existing approaches through extensive experiments, using both synthetic data and real human measurements.     

Shape Indexing and Recognition Based on Regional Analysis

Shape indexing and recognition have received great attention in multimedia processing communities due to the wide range of utilities. In achieving shape representation, most influential methods treat shapes as intrinsic curves, which is not in agreement with the way human vision systems achieve the same task. In this paper, a new framework is developed where a shape is treated as a 2-D region. We first perform an eigen analysis to align P in the standard orientation. Three numbers are generated to indicate the global geometrical nature. Next, according to the two eigen vectors, we partition P into four halves and eight quadrants. Five numbers are then produced for each region to signify its geometrical properties and relation with P. The aggregate of these numbers, 63 in total, is the actual index for P. Recognition is effected by weighted LI distances between shapes. This indexing scheme captures the global geometry of shapes and is resilient to rotations and scales, which are of crucial importance in the perceptive process of human vision systems. It can tolerate occlusions present in most standard shape datasets but is not robust against severe occlusions. Empirical studies conducted on standard synthetic and real-world datasets demonstrate encouraging performances.     

Bayesian 3D shape from silhouettes

This paper introduces a smooth posterior density function for inferring shapes from silhouettes. Both the likelihood and the prior are modelled using kernel density functions and optimisation is performed using gradient ascent algorithms. Adding a prior allows for the recovery of concave areas of the shape that are usually lost when estimating the visual hull. This framework is also extended to use colour information when it is available in addition to the silhouettes. In these cases, the modelling not only allows for the shape to be recovered but also its colour information. Our new algorithms are assessed by reconstructing 2D shapes from 1D silhouettes and 3D faces from 2D silhouettes. Experimental results show that using the prior can assist in reconstructing concave areas and also illustrate the benefits of using colour information even when only small numbers of silhouettes are available.     

Structure from Motion: Beyond the Epipolar Constraint

The classic approach to structure from motion entails a clear separation between motion estimation and structure estimation and between two-dimensional (2D) and three-dimensional (3D) information. For the recovery of the rigid transformation between different views only 2D image measurements are used. To have available enough information, most existing techniques are based on the intermediate computation of optical flow which, however, poses a problem at the locations of depth discontinuities. If we knew where depth discontinuities were, we could (using a multitude of approaches based on smoothness constraints) accurately estimate flow values for image patches corresponding to smooth scene patches; but to know the discontinuities requires solving the structure from motion problem first. This paper introduces a novel approach to structure from motion which addresses the processes of smoothing, 3D motion and structure estimation in a synergistic manner. It provides an algorithm for estimating the transformation between two views obtained by either a calibrated or uncalibrated camera. The results of the estimation are then utilized to perform a reconstruction of the scene from a short sequence of images.The technique is based on constraints on image derivatives which involve the 3D motion and shape of the scene, leading to a geometric and statistical estimation problem. The interaction between 3D motion and shape allows us to estimate the 3D motion while at the same time segmenting the scene. If we use a wrong 3D motion estimate to compute depth, we obtain a distorted version of the depth function. The distortion, however, is such that the worse the motion estimate, the more likely we are to obtain depth estimates that vary locally more than the correct ones. Since local variability of depth is due either to the existence of a discontinuity or to a wrong 3D motion estimate, being able to differentiate between these two cases provides the correct motion, which yields the least varying estimated depth as well as the image locations of scene discontinuities. We analyze the new constraints, show their relationship to the minimization of the epipolar constraint, and present experimental results using real image sequences that indicate the robustness of the method.     

Automatic 3D Shape Co-Segmentation Using Spectral Graph Method

Co-analyzing a set of 3D shapes is a challenging task considering a large geometrical variability of the shapes. To address this challenge, this paper proposes a new automatic 3D shape co-segmentation algorithm by using spectral graph method.Our method firstly represents input shapes as a set of weighted graphs and extracts multiple geometric features to measure the similarities of faces in each individual shape.Secondly all graphs are embedded into the spectral domain to find meaningful correspondences across the set.After that we build a joint weighted matrix for the graph set and then apply normalized cut criterion to find optimal co-segmentation of the input shapes.Finally we evaluate our approach on different categories of 3D shapes, and the experimental results demonstrate that our method can accurately co-segment a wide variety of shapes, which may have different poses and significant topology changes.     

Nonrigid structure-from-motion: estimating shape and motion with hierarchical Priors

This paper describes methods for recovering time-varying shape and motion of non-rigid 3D objects from uncalibrated 2D point tracks. For example, given a video recording of a talking person, we would like to estimate the 3D shape of the face at each instant, and learn a model of facial deformation. Time-varying shape is modeled as a rigid transformation combined with a non-rigid deformation. Reconstruction is ill-posed if arbitrary deformations are allowed, and thus additional assumptions about deformations are required. We first suggest restricting shapes to lie within a low-dimensional subspace, and describe estimation algorithms. However, this restriction alone is insufficient to constrain reconstruction. To address these problems, we propose a reconstruction method using a Probabilistic Principal Components Analysis (PPCA) shape model, and an estimation algorithm that simultaneously estimates 3D shape and motion for each instant, learns the PPCA model parameters, and robustly fills-in missing data points. We then extend the model to model temporal dynamics in object shape, allowing the algorithm to robustly handle severe cases of missing data.     

Subspace Procrustes Analysis

Procrustes analysis (PA) has been a popular technique to align and build 2-D statistical models of shapes. Given a set of 2-D shapes PA is applied to remove rigid transformations. Later, a non-rigid 2-D model is computed by modeling the residual (e.g., PCA). Although PA has been widely used, it has several limitations for modeling 2-D shapes: occluded landmarks and missing data can result in local minima solutions, and there is no guarantee that the 2-D shapes provide a uniform sampling of the 3-D space of rotations for the object. To address previous issues, this paper proposes subspace PA (SPA). Given several instances of a 3-D object, SPA computes the mean and a 2-D subspace that can model rigid and non-rigid deformations of the 3-D object. We propose a discrete (DSPA) and continuous (CSPA) formulation for SPA, assuming that 3-D samples of an object are provided. DSPA extends the traditional PA, and produces unbiased 2-D models by uniformly sampling different views of the 3-D object. CSPA provides a continuous approach to uniformly sample the space of 3-D rotations, being more efficient in space and time. We illustrate the benefits of SPA in two different applications. First, SPA is used to learn 2-D face and body models from 3-D datasets. Experiments on the FaceWarehouse and CMU motion capture (MoCap) datasets show the benefits of our 2-D models against the state-of-the-art PA approaches and conventional 3-D models. Second, SPA learns an unbiased 2-D model from CMU MoCap dataset and it is used to estimate the human pose on the Leeds Sports dataset.     

图像驱动的三维模型外观细节增强

随着计算机游戏与电影视觉特效的普及应用,仅通过纹理贴图已不能满足用户需求。然而,为三维模型制作视觉真实的表面外观,即在包含几何细节变化(如金属腐蚀、石料风化、木质裂纹等)同时辅之以与几何变化合理匹配的材质颜色却是一件困难且耗时的工作。文章提出了一种能够从单幅图像中提取物体的外观细节(材质与几何信息)并增强至三维模型表面的方法。该方法是一种新型的几何约束的非刚性2D-3D融合配准算法,可以通过将配准后的模型提供的几何信息作为先验知识实现更好的本征图像分解结果。利用图像中同步分解出的互相关的多尺度几何信息与材质纹理信息组成的非参数外观模型,文中提出了一种法向量约束的几何变形算法将外观纹理恢复到代理模型上。通过上述步骤,文章提出的方法能够帮助建模师制作具有多尺度外观细节的三维模型。     

An Elasticity-Based Covariance Analysis of Shapes

We introduce the covariance of a number of given shapes if they are interpreted as boundary contours of elastic objects. Based on the notion of nonlinear elastic deformations from one shape to another, a suitable linearization of geometric shape variations is introduced. Once such a linearization is available, a principal component analysis can be investigated. This requires the definition of a covariance metric—an inner product on linearized shape variations. The resulting covariance operator robustly captures strongly nonlinear geometric variations in a physically meaningful way and allows to extract the dominant modes of shape variation. The underlying elasticity concept represents an alternative to Riemannian shape statistics. In this paper we compare a standard L ²-type covariance metric with a metric based on the Hessian of the nonlinear elastic energy. Furthermore, we explore the dependence of the principal component analysis on the type of the underlying nonlinear elasticity. For the built-in pairwise elastic registration, a relaxed model formulation is employed which allows for a non-exact matching. Shape contours are approximated by single well phase fields, which enables an extension of the method to a covariance analysis of image morphologies. The model is implemented with multilinear finite elements embedded in a multi-scale approach. The characteristics of the approach are demonstrated on a number of illustrative and real world examples in 2D and 3D.     

Shape registration in implicit spaces using information theory and free form deformations

We present a novel, variational and statistical approach for shape registration. Shapes of interest are implicitly embedded in a higher-dimensional space of distance transforms. In this implicit embedding space, registration is formulated in a hierarchical manner: the mutual information criterion supports various transformation models and is optimized to perform global registration; then, a B-spline-based incremental free form deformations (IFFD) model is used to minimize a sum-of-squared-differences (SSD) measure and further recover a dense local nonrigid registration field. The key advantage of such framework is twofold: 1) it naturally deals with shapes of arbitrary dimension (2D, 3D, or higher) and arbitrary topology (multiple parts, closed/open) and 2) it preserves shape topology during local deformation and produces local registration fields that are smooth, continuous, and establish one-to-one correspondences. Its invariance to initial conditions is evaluated through empirical validation, and various hard 2D/3D geometric shape registration examples are used to show its robustness to noise, severe occlusion, and missing parts. We demonstrate the power of the proposed framework using two applications: one for statistical modeling of anatomical structures, another for 3D face scan registration and expression tracking. We also compare the performance of our algorithm with that of several other well-known shape registration algorithms.     

基于相关型图像传感器3维人脸成像的3维AAMs人脸识别方法的研究

主动外观模型是基于统计分析建立物体2维模型的有效方法,它融合了目标的形状和纹理信息。在基于相关型图像传感器3维人脸成像的基础上,提出了一种建立3维人脸模型的方法,该方法利用由相关型图像传感器得到的深度信息和与之对应的亮度信息将2维AAMs扩展为3维AAMs,融合人脸的形状,纹理和深度信息来构建3维人脸模型。人脸识别实验结果表明,该方法在不同人脸姿态,表情和光照条件下识别效果要优于Eigenface和2维AAMs。     

Three-dimensional face recognition using shapes of facial curves

We study shapes of facial surfaces for the purpose of face recognition. The main idea is to 1) represent surfaces by unions of level curves, called facial curves, of the depth function and 2) compare shapes of surfaces implicitly using shapes of facial curves. The latter is performed using a differential geometric approach that computes geodesic lengths between closed curves on a shape manifold. These ideas are demonstrated using a nearest-neighbor classifier on two 3D face databases: Florida State University and Notre Dame, highlighting a good recognition performance     

Hierarchical vibrations for part-based recognition of complex objects

We propose a technique for the recognition and segmentation of complex shapes in 2D images using a hierarchy of finite element vibration modes in an evolutionary shape search. The different levels of the shape hierarchy can influence each other, which can be exploited in top-down part-based image analysis. Our method overcomes drawbacks of existing structural approaches, which cannot uniformly encode shape variation and co-variation, or rely on training. We present results demonstrating that by utilizing a quality-of-fit function the model explicitly recognizes missing parts of a complex shape, thus allowing for categorization between shape classes.