Human activity recognition using multidimensional indexing

In this paper, we develop a novel method for view-based recognition of human action/activity from videos. By observing just a few frames, we can identify the activity that takes place in a video sequence. The basic idea of our method is that activities can be positively identified from a sparsely sampled sequence of a few body poses acquired from videos. In our approach, an activity is represented by a set of pose and velocity vectors for the major body parts (hands, legs, and torso) and stored in a set of multidimensional hash tables. We develop a theoretical foundation that shows that robust recognition of a sequence of body pose vectors can be achieved by a method of indexing and sequencing and it requires only a few pose vectors (i.e., sampled body poses in video frames). We find that the probability of false alarm drops exponentially with the increased number of sampled body poses. So, matching only a few body poses guarantees high probability for correct recognition. Our approach is parallel, i.e., all possible model activities are examined at one indexing operation. In addition, our method is robust to partial occlusion since each body part is indexed separately. We use a sequence-based voting approach to recognize the activity invariant to the activity speed.     

视频中多特征融合人体姿态跟踪

目的 目前已有的人体姿态跟踪算法的跟踪精度仍有待提高,特别是对灵活运动的手臂部位的跟踪。为提高人体姿态的跟踪精度,本文首次提出一种将视觉时空信息与深度学习网络相结合的人体姿态跟踪方法。方法 在人体姿态跟踪过程中,利用视频时间信息计算出人体目标区域的运动信息,使用运动信息对人体部位姿态模型在帧间传递;考虑到基于图像空间特征的方法对形态较为固定的人体部位如躯干和头部能够较好地检测,而对手臂的检测效果较差,构造并训练一种轻量级的深度学习网络,用于生成人体手臂部位的附加候选样本;利用深度学习网络生成手臂特征一致性概率图,与视频空间信息结合计算得到最优部位姿态,并将各部位重组为完整人体姿态跟踪结果。结果 使用两个具有挑战性的人体姿态跟踪数据集VideoPose2.0和YouTubePose对本文算法进行验证,得到的手臂关节点平均跟踪精度分别为81.4%和84.5%,与现有方法相比有明显提高;此外,通过在VideoPose2.0数据集上的实验,验证了本文提出的对下臂附加采样的算法和手臂特征一致性计算的算法能够有效提高人体姿态关节点的跟踪精度。结论 提出的结合时空信息与深度学习网络的人体姿态跟踪方法能够有效提高人体姿态跟踪的精度,特别是对灵活运动的人体姿态下臂关节点的跟踪精度有显著提高。     

多特征融合的人体目标再识别

在非重叠的多摄像机监控系统中,人体目标再识别是需要解决的主要问题之一。针对当前人体目标再识别使用目标的外观统计特征或者通过训练获取目标特征时存在的问题,提出一种无需训练,对视角、光照变化和姿态变化具有较强鲁棒性的基于多特征的人体目标再识别算法。首先根据空间直方图建立目标整体外观表现模型对目标进行粗识别,之后将人体目标分为3部分,忽略头部信息,分别提取躯干和腿部的主色区域的局部颜色和形状特征,并通过EMD(earth movers distance)距离进行目标精识别。实验结果表明,本文算法具有较高的识别率,且不受遮挡和背景粘连的影响。     

Pose analysis using spectral geometry

We propose a novel method to analyze a set of poses of 3D models that are represented with triangle meshes and unregistered. Different shapes of poses are transformed from the 3D spatial domain to a geometry spectrum domain that is defined by Laplace–Beltrami operator. During this space-spectrum transform, all near-isometric deformations, mesh triangulations and Euclidean transformations are filtered away. The different spatial poses from a 3D model are represented with near-isometric deformations; therefore, they have similar behaviors in the spectral domain. Semantic parts of that model are then determined based on the computed geometric properties of all the mapped vertices in the geometry spectrum domain. Semantic skeleton can be automatically built with joints detected as well. The Laplace–Beltrami operator is proved to be invariant to isometric deformations and Euclidean transformations such as translation and rotation. It also can be invariant to scaling with normalization. The discrete implementation also makes the Laplace–Beltrami operator straightforward to be applied on triangle meshes despite triangulations. Our method turns a rather difficult spatial problem into a spectral problem that is much easier to solve. The applications show that our 3D pose analysis method leads to a registration-free pose analysis and a high-level semantic part understanding of 3D shapes.     

Human skeleton tracking from depth data using geodesic distances and optical flow

In this paper, we present a method for human full-body pose estimation from depth data that can be obtained using Time of Flight (ToF) cameras or the Kinect device. Our approach consists of robustly detecting anatomical landmarks in the 3D data and fitting a skeleton body model using constrained inverse kinematics. Instead of relying on appearance-based features for interest point detection that can vary strongly with illumination and pose changes, we build upon a graph-based representation of the depth data that allows us to measure geodesic distances between body parts. As these distances do not change with body movement, we are able to localize anatomical landmarks independent of pose. For differentiation of body parts that occlude each other, we employ motion information, obtained from the optical flow between subsequent intensity images. We provide a qualitative and quantitative evaluation of our pose tracking method on ToF and Kinect sequences containing movements of varying complexity.     

A Statistical Model of Human Pose and Body Shape

Generation and animation of realistic humans is an essential part of many projects in today's media industry. Especially, the games and special effects industry heavily depend on realistic human animation. In this work a unified model that describes both, human pose and body shape is introduced which allows us to accurately model muscle deformations not only as a function of pose but also dependent on the physique of the subject. Coupled with the model's ability to generate arbitrary human body shapes, it severely simplifies the generation of highly realistic character animations. A learning based approach is trained on approximately 550 full body 3D laser scans taken of 114 subjects. Scan registration is performed using a non-rigid deformation technique. Then, a rotation invariant encoding of the acquired exemplars permits the computation of a statistical model that simultaneously encodes pose and body shape. Finally, morphing or generating meshes according to several constraints simultaneously can be achieved by training semantically meaningful regressors.     

Visual‐inertial curve simultaneous localization and mapping: Creating a sparse structured world without feature points

We present a simultaneous localization and mapping (SLAM) algorithm that uses Bézier curves as static landmark primitives rather than feature points. Our approach allows us to estimate the full six degrees of freedom pose of a robot while providing a structured map that can be used to assist a robot in motion planning and control. We demonstrate how to reconstruct the three‐dimensional (3D) location of curve landmarks from a stereo pair and how to compare the 3D shape of curve landmarks between chronologically sequential stereo frames to solve the data association problem. We also present a method to combine curve landmarks for mapping purposes, resulting in a map with a continuous set of curves that contain fewer landmark states than conventional point‐based SLAM algorithms. We demonstrate our algorithm's effectiveness with numerous experiments, including comparisons to existing state‐of‐the‐art SLAM algorithms.     

Stable Region Correspondences Between Non‐Isometric Shapes

We consider the problem of finding meaningful correspondences between 3D models that are related but not necessarily very similar. When the shapes are quite different, a point‐to‐point map is not always appropriate, so our focus in this paper is a method to build a set of correspondences between shape regions or parts. The proposed approach exploits a variety of feature functions on the shapes and makes use of the key observation that points in matching parts have similar ranks in the sorting of the corresponding feature values. Our algorithm proceeds in two steps. We first build an affinity matrix between points on the two shapes, based on feature rank similarity over many feature functions. We then define a notion of stability of a pair of regions, with respect to this affinity matrix, obtained as a fixed point of a nonlinear operator. Our method yields a family of corresponding maximally stable regions between the two shapes that can be used to define shape parts. We observe that this is an instance of the biclustering problem and that it is related to solving a constrained maximal eigenvalue problem. We provide an algorithm to solve this problem that mimics the power method. We show the robustness of its output to noisy input features as well its convergence properties. The obtained part correspondences are shown to be almost perfect matches in the isometric case, and also semantically appropriate even in non‐isometric cases. We provide numerous examples and applications of this technique, for example to sharpening correspondences in traditional shape matching algorithms.     

Human body segmentation based on shape constraint

Human body segmentation is essential for many practical applications, e.g., video surveillance analysis in intelligent urban. However, existing methods mainly suffer from various human poses. In this paper, we try to address this issue by introducing human shape constraint. First, human pose estimation is performed, and locations of human body parts are determined. Contrast to the previous work, we just use the human body parts with high precision. Then we combines the star convexity and the human body parts’ locations as shape constraint. The final segmentation results are acquired through the optimization step. Comprehensive and comparative experimental results demonstrate that the proposed method achieves promising performance and outperforms many state-of-the-art methods over publicly available challenging datasets.     

3D shape estimation in video sequences provides high precision evaluation of facial expressions

Person independent and pose invariant estimations of facial expressions and action unit (AU) intensity estimation are important for situation analysis and for automated video annotation. We evaluated raw 2D shape data of the CK+ database, used Procrustes transformation and the multi-class SVM leave-one-out method for classification. We found close to 100% performance demonstrating the relevance and the strength of details of the shape. Precise 3D shape information was computed by means of constrained local models (CLM) on video sequences. Such sequences offer the opportunity to compute a time-averaged ‘3D personal mean shape’ (PMS) from the estimated CLM shapes, which – upon subtraction – gives rise to person independent emotion estimation. On CK+ data PMS showed significant improvements over AU0 normalization; performance reached and sometimes surpassed state-of-the-art results on emotion classification and on AU intensity estimation. 3D PMS from 3D CLM offers pose invariant emotion estimation that we studied by rendering a 3D emotional database for different poses and different subjects from the BU 4DFE database. Frontal shapes derived from CLM fits of the 3D shape were evaluated. Results demonstrate that shape estimation alone can be used for robust, high quality pose invariant emotion classification and AU intensity estimation.     

Automatic body landmark identification for various body figures

In this paper, we presented automatic body landmark identification algorithms that deals flexibly with the difference in body shapes and reduces the inconsistency resulting from the differences in body shapes. First, the landmark search range was defined using the statistical analysis. Next, body scan direction was identified and it was segmented. Next, automatic landmark identification algorithms were developed for each of the six landmarks and the accuracy was examined for each body shape. The scans were extracted from 5th Size Korea database. This algorithms were successfully tested on various body shapes and improved the robustness.

Relevance to industry

In automatic body measurement systems, the landmark location error occurring at nonstandard body shapes nullifies the advantage of saving time. It also makes the 3D scan measurements unreliable. The improvement of reliability and accuracy of the automatic 3D body measurement algorithm for various human body shapes will reduce the time for performing measurements and be practical for use in human-size-related production processes.     

Human action recognition using an ensemble of body‐part detectors

This paper describes an approach to human action recognition based on a probabilistic optimization model of body parts using hidden Markov model (HMM). Our method is able to distinguish between similar actions by only considering the body parts having major contribution to the actions, for example, legs for walking, jogging and running; arms for boxing, waving and clapping. We apply HMMs to model the stochastic movement of the body parts for action recognition. The HMM construction uses an ensemble of body‐part detectors, followed by grouping of part detections, to perform human identification. Three example‐based body‐part detectors are trained to detect three components of the human body: the head, legs and arms. These detectors cope with viewpoint changes and self‐occlusions through the use of ten sub‐classifiers that detect body parts over a specific range of viewpoints. Each sub‐classifier is a support vector machine trained on features selected for the discriminative power for each particular part/viewpoint combination. Grouping of these detections is performed using a simple geometric constraint model that yields a viewpoint‐invariant human detector. We test our approach on three publicly available action datasets: the KTH dataset, Weizmann dataset and HumanEva dataset. Our results illustrate that with a simple and compact representation we can achieve robust recognition of human actions comparable to the most complex, state‐of‐the‐art methods.     

3D Body Shapes Estimation from Dressed‐Human Silhouettes

Estimation of 3D body shapes from dressed‐human photos is an important but challenging problem in virtual fitting. We propose a novel automatic framework to efficiently estimate 3D body shapes under clothes. We construct a database of 3D naked and dressed body pairs, based on which we learn how to predict 3D positions of body landmarks (which further constrain a parametric human body model) automatically according to dressed‐human silhouettes. Critical vertices are selected on 3D registered human bodies as landmarks to represent body shapes, so as to avoid the time‐consuming vertices correspondences finding process for parametric body reconstruction. Our method can estimate 3D body shapes from dressed‐human silhouettes within 4 seconds, while the fastest method reported previously need 1 minute. In addition, our estimation error is within the size tolerance for clothing industry. We dress 6042 naked bodies with 3 sets of common clothes by physically based cloth simulation technique. To the best of our knowledge, We are the first to construct such a database containing 3D naked and dressed body pairs and our database may contribute to the areas of human body shapes estimation and cloth simulation.     

Deformable model for estimating clothed and naked human shapes from a single image

Estimation of human shape from images has numerous applications ranging from graphics to surveillance. A single image provides insufficient constraints (e.g. clothing), making human shape estimation more challenging. We propose a method to simultaneously estimate a person’s clothed and naked shapes from a single image of that person wearing clothing. The key component of our method is a deformable model of clothed human shape. We learn our deformable model, which spans variations in pose, body, and clothes, from a training dataset. These variations are derived by the non-rigid surface deformation, and encoded in various low-dimension parameters. Our deformable model can be used to produce clothed 3D meshes for different people in different poses, which neither appears in the training dataset. Afterward, given an input image, our deformable model is initialized with a few user-specified 2D joints and contours of the person. We optimize the parameters of the deformable model by pose fitting and body fitting in an iterative way. Then the clothed and naked 3D shapes of the person can be obtained simultaneously. We illustrate our method for texture mapping and animation. The experimental results on real images demonstrate the effectiveness of our method.     

基于角点聚类的移动机器人自然路标检测与识别

针对未知环境中机器人视觉导航的自然路标检测,提出了一种基于角点聚类的自然路标局部特征提取、不变性表示及其匹配算法.用SUSAN算子提取左右视图中的角点,在极线约束下对左右视图的角点进行匹配,消除遮挡或噪声引起的角点;同时应用立体视觉计算角点视差,进一步筛选角点.根据角点聚类策略提取自然路标局部特征,并提出不随距离、角度变化的局部特征不变性表示及匹配方法.理论分析和实验结果表明,该算法具有较好的鲁棒性,在一定距离和角度变换下能够对路标进行正确识别.     

姿态和表情变化下的三维人脸标志点定位

三维人脸标志点定位在人脸识别、人脸跟踪、人脸建模、表情分析等方面具有非常重要的作用. 然而, 在姿态和表情变化很大的条件下进行标志点定位, 这仍然是一个很具挑战性的课题. 本文提出一种对姿态和表情不敏感的三维人脸标志点定位方法, 利用HK曲率分析检测出候选标志点, 根据对面部形状的先验知识, 提出一种基于人脸几何结构的分类策略对候选标志点进一步细分, 通过把候选标志点与人脸标志点模型进行匹配, 实现标志点的精确定位. 首先在CASIA数据集对该方法的标志点定位精度进行测试, 然后在UND/FRGC v2.0数据集对该方法与其他方法进行比较. 实验结果表明该方法在姿态和表情变化很大的条件下具有高精度和高鲁棒性.     

Robust Structure‐Based Shape Correspondence

We present a robust method to find region‐level correspondences between shapes, which are invariant to changes in geometry and applicable across multiple shape representations. We generate simplified shape graphs by jointly decomposing the shapes, and devise an adapted graph‐matching technique, from which we infer correspondences between shape regions. The simplified shape graphs are designed to primarily capture the overall structure of the shapes, without reflecting precise information about the geometry of each region, which enables us to find correspondences between shapes that might have significant geometric differences. Moreover, due to the special care we take to ensure the robustness of each part of our pipeline, our method can find correspondences between shapes with different representations, such as triangular meshes and point clouds. We demonstrate that the region‐wise matching that we obtain can be used to find correspondences between feature points, reveal the intrinsic self‐similarities of each shape and even construct point‐to‐point maps across shapes. Our method is both time and space efficient, leading to a pipeline that is significantly faster than comparable approaches. We demonstrate the performance of our approach through an extensive quantitative and qualitative evaluation on several benchmarks where we achieve comparable or superior performance to existing methods.     

Environment‐Adaptive Contact Poses for Virtual Characters

We present a novel method to generate a virtual character's multi‐contact poses adaptive to the various shapes of the environment. Given the user‐specified center of mass (CoM) position and direction as inputs, our method finds the potential contacts for the character in the surrounding geometry of the environment and generates a set of stable poses that are contact‐rich. Major contributions of the work are in efficiently finding admissible support points for the target environment by precomputing candidate support points from a human pose database, and in automatically generating interactive poses that can maintain stable equilibrium. We develop the concept of support complexity to scale the set of precomputed support points by the geometric complexity of the environment. We demonstrate the effectiveness of our method by creating contact poses for various test cases of environments.     

Combining Generative and Discriminative Models in a Framework for Articulated Pose Estimation

We develop a method for the estimation of articulated pose, such as that of the human body or the human hand, from a single (monocular) image. Pose estimation is formulated as a statistical inference problem, where the goal is to find a posterior probability distribution over poses as well as a maximum a posteriori (MAP) estimate. The method combines two modeling approaches, one discriminative and the other generative. The discriminative model consists of a set of mapping functions that are constructed automatically from a labeled training set of body poses and their respective image features. The discriminative formulation allows for modeling ambiguous, one-to-many mappings (through the use of multi-modal distributions) that may yield multiple valid articulated pose hypotheses from a single image. The generative model is defined in terms of a computer graphics rendering of poses. While the generative model offers an accurate way to relate observed (image features) and hidden (body pose) random variables, it is difficult to use it directly in pose estimation, since inference is computationally intractable. In contrast, inference with the discriminative model is tractable, but considerably less accurate for the problem of interest. A combined discriminative/generative formulation is derived that leverages the complimentary strengths of both models in a principled framework for articulated pose inference. Two efficient MAP pose estimation algorithms are derived from this formulation; the first is deterministic and the second non-deterministic. Performance of the framework is quantitatively evaluated in estimating articulated pose of both the human hand and human body. Most of this work was done while the first author was with Boston University.