基于视频的实时自动人体高度测量

怎样从视频出发准确实时地测量场景中运动人体的高度? 针对此问题, 本文提出了一种自动实时的人体高度测量方法. 该方法首先在视频序列中的每帧图像上提取一种新的头部特征点以及一种新的脚部特征点, 然后根据这些特征点建立约束方程求出近似的人体高度, 并同时在视频序列中跟踪双脚. 最后基于获得的双脚跟踪结果, 引入一条关于特征点所对应空间点的几何约束以进一步优化测量结果. 与过去的许多测量方法相比, 本文方法有效地利用了视频序列中包含的运动信息, 有较强的鲁棒性和较高的测量精度, 既能有效地处理透视镜头下的视频又能处理鱼眼镜头下的视频, 而且计算量很低, 可以实现实时测量. 实验结果验证了本文算法的有效性和实时性.     

Foreground extraction via dual-side cameras on a mobile device using long short-term trajectory analysis

This paper presents a foreground extraction method for live-streaming videos using dual-side cameras on mobile devices. Compared to conventional methods, which estimate both foreground and background models from the front camera, the proposed method uses the rear camera to infer the reference background model. To this end, the short-term trajectory analysis is first performed to cluster point trajectories of the front camera, and then the long-term trajectory analysis is performed to compare the paths of the clustered trajectories with the reference path obtained from the rear camera. In particular, clusters having high correlation are classified as background using the Gaussian mixture model. Additionally, a pixel-wise segmentation map is obtained via graph-based segmentation. Experimental results show that the proposed method is robust under a variety of camera motion, outperforming state-of-the-art methods. Code and dataset can be found at https://github.com/YCL92/dualCamSeg.     

Intrinsic dynamics and total energy-shaping control of the ballbot system

Research on bipedal locomotion has shown that a dynamic walking gait is energetically more efficient than a statically stable one. Analogously, even though statically stable multi-wheeled robots are easier to control, they are energetically less efficient and have low accelerations to avoid tipping over. In contrast, the ballbot is an underactuated, nonholonomically constrained mobile robot, whose upward equilibrium point has to be stabilised by active control. In this work, we derive coordinate-invariant, reduced, Euler–Poincaré equations of motion for the ballbot. By means of partial feedback linearisation, we obtain two independent passive outputs with corresponding storage functions and utilise these to come up with energy-shaping control laws which move the system along the trajectories of a new Lagrangian system whose desired equilibrium point is asymptotically stable by construction. The basin of attraction of this controller is shown to be almost global under certain conditions on the design of the mechanism which are reflected directly in the mass matrix of the unforced equations of motion.     

A General Framework for Trajectory Triangulation

The multiple view geometry of static scenes is now well understood. Recently attention was turned to dynamic scenes where scene points may move while the cameras move. The triangulation of linear trajectories is now well handled. The case of quadratic trajectories also received some attention. We present a complete generalization and address the problem of general trajectory triangulation of moving points from non-synchronized cameras. Two cases are considered: (i) the motion is captured in the images by tracking the moving point itself, (ii) the tangents of the motion only are extracted from the images. The first case is based on a new representation (to computer vision) of curves (trajectories) where a curve is represented by a family of hypersurfaces in the projective space ?⁵. The second case is handled by considering the dual curve of the curve generated by the trajectory. In both cases these representations of curves allow: (i) the triangulation of the trajectory of a moving point from non-synchronized sequences, (ii) the recovery of more standard representation of the whole trajectory, (iii) the computations of the set of positions of the moving point at each time instant an image was made. Furthermore, theoretical considerations lead to a general theorem stipulating how many independent constraints a camera provides on the motion of the point. This number of constraint is a function of the camera motion. On the computation front, in both cases the triangulation leads to equations where the unknowns appear linearly. Therefore the problem reduces to estimate a high-dimensional parameter in presence of heteroscedastic noise. Several method are tested.     

Video synchronization and its application to object transfer

In video post-production applications, camera motion analysis and alignment are important in order to ensure the geometric correctness and temporal consistency. In this paper, we trade some generality in estimating and aligning camera motion for reduced computational complexity and increased image-based nature. The main contribution is to use fundamental ratios to synchronize video sequences of distinct scenes captured by cameras undergoing similar motions. We also present a simple method to align 3D camera trajectories when the fundamental ratios are not able to match the noisy trajectories. Experimental results show that our method can accurately synchronize sequences even when the scenes are totally different and have dense depths. An application on 3D object transfer is also demonstrated.     

Distribution-Based Dimensionality Reduction Applied to Articulated Motion Recognition

Some articulated motion representations rely on frame-wise abstractions of the statistical distribution of low-level features such as orientation, color, or relational distributions. As configuration among parts changes with articulated motion, the distribution changes, tracing a trajectory in the latent space of distributions, which we call the configuration space. These trajectories can then be used for recognition using standard techniques such as dynamic time warping. The core theory in this paper concerns embedding the frame-wise distributions, which can be looked upon as probability functions, into a low-dimensional space so that we can estimate various meaningful probabilistic distances such as the Chernoff, Bhattacharya, Matusita, Kullback-Leibler (KL) or symmetric-KL distances based on dot products between points in this space. Apart from computational advantages, this representation also affords speed-normalized matching of motion signatures. Speed normalized representations can be formed by interpolating the configuration trajectories along their arc lengths, without using any knowledge of the temporal scale variations between the sequences. We experiment with five different probabilistic distance measures and show the usefulness of the representation in three different contexts—sign recognition (with large number of possible classes), gesture recognition (with person variations), and classification of human-human interaction sequences (with segmentation problems). We find the importance of using the right distance measure for each situation. The low-dimensional embedding makes matching two to three times faster, while achieving recognition accuracies that are close to those obtained without using a low-dimensional embedding. We also empirically establish the robustness of the representation with respect to low-level parameters, embedding parameters, and temporal-scale parameters.     

Dynamic balance preservation and prevention of sliding for humanoid robots in the presence of multiple spatial contacts

The main indicator of dynamic balance is the \(\mathit{ZMP}\). Its original notion assumes that both feet of the robot are in contact with the flat horizontal surface (all contacts are in the same plane) and that the friction is high enough so that sliding does not occur. With increasing capabilities of humanoid robots and the higher complexity of the motion that needs to be performed, these assumptions might not hold. Having in mind that the system is dynamically balanced if there is no rotation about the edges of the feet and if the feet do not slide, we propose a novel approach for testing the dynamic balance of bipedal robots, by using linear contact wrench conditions compiled in a single matrix (Dynamic Balance Matrix). The proposed approach has wide applicability since it can be used to check the stability of different kinds of contacts (including point, line, and surface) with arbitrary perimeter shapes. Motion feasibility conditions are derived on the basis of the conditions which the wrench of each contact has to satisfy. The approach was tested by simulation in two scenarios: biped climbing up and walking sideways on the inclined flat surface which is too steep for a regular walk without additional support. The whole-body motion was synthesized and performed using a generalized task prioritization framework.     

Less restrictive camera odometry estimation from monocular camera

This paper addresses the problem of estimating a camera motion from a non-calibrated monocular camera. Compared to existing methods that rely on restrictive assumptions, we propose a method which can estimate camera motion with much less restrictions by adopting new example-based techniques compensating the lack of information. Specifically, we estimate the focal length of the camera by referring to visually similar training images with which focal lengths are associated. For one step camera estimation, we refer to stationary points (landmark points) whose depths are estimated based on RGB-D candidates. In addition to landmark points, moving objects can be also used as an information source to estimate the camera motion. Therefore, our method simultaneously estimates the camera motion for a video, and the 3D trajectories of objects in this video by using Reversible Jump Markov Chain Monte Carlo (RJ-MCMC) particle filtering. Our method is evaluated on challenging datasets demonstrating its effectiveness and efficiency.     

Multilinear Factorizations for Multi-Camera Rigid Structure from Motion Problems

Camera networks have gained increased importance in recent years. Existing approaches mostly use point correspondences between different camera views to calibrate such systems. However, it is often difficult or even impossible to establish such correspondences. But even without feature point correspondences between different camera views, if the cameras are temporally synchronized then the data from the cameras are strongly linked together by the motion correspondence: all the cameras observe the same motion. The present article therefore develops the necessary theory to use this motion correspondence for general rigid as well as planar rigid motions. Given multiple static affine cameras which observe a rigidly moving object and track feature points located on this object, what can be said about the resulting point trajectories? Are there any useful algebraic constraints hidden in the data? Is a 3D reconstruction of the scene possible even if there are no point correspondences between the different cameras? And if so, how many points are sufficient? Is there an algorithm which warrants finding the correct solution to this highly non-convex problem? This article addresses these questions and thereby introduces the concept of low-dimensional motion subspaces. The constraints provided by these motion subspaces enable an algorithm which ensures finding the correct solution to this non-convex reconstruction problem. The algorithm is based on multilinear analysis, matrix and tensor factorizations. Our new approach can handle extreme configurations, e.g. a camera in a camera network tracking only one single point. Results on synthetic as well as on real data sequences act as a proof of concept for the presented insights.