Markerless tracking of complex human motions from multiple views

We present a method for markerless tracking of complex human motions from multiple camera views. In the absence of markers, the task of recovering the pose of a person during such motions is challenging and requires strong image features and robust tracking. We propose a solution which integrates multiple image cues such as edges, color information and volumetric reconstruction. We show that a combination of multiple image cues helps the tracker to overcome ambiguous situations such as limbs touching or strong occlusions of body parts. Following a model-based approach, we match an articulated body model built from superellipsoids against these image cues. Stochastic Meta Descent (SMD) optimization is used to find the pose which best matches the images. Stochastic sampling makes SMD robust against local minima and lowers the computational costs as a small set of predicted image features is sufficient for optimization. The power of SMD is demonstrated by comparing it to the commonly used Levenberg–Marquardt method. Results are shown for several challenging sequences showing complex motions and full articulation, with tracking of 24 degrees of freedom in ≈1 frame per second.     

A novel approach for recognition of human actions with semi-global features

In this study a new approach is presented for the recognition of human actions of everyday life with a fixed camera. The originality of the presented method consists in characterizing sequences by a temporal succession of semi-global features, which are extracted from “space-time micro-volumes”. The advantage of this approach lies in the use of robust features (estimated on several frames) associated with the ability to manage actions with variable durations and easily segment the sequences with algorithms that are specific to time-varying data. Each action is actually characterized by a temporal sequence that constitutes the input of a Hidden Markov Model system for the recognition. Results presented of 1,614 sequences performed by several persons validate the proposed approach.     

Nonlinearity-aware sub-model combination in trajectory based methods for nonlinear Mor

Trajectory based methods approximate nonlinear dynamical systems by superposition of dimensionally reduced linear systems. The linear systems are obtained by linearisations at multiple points along a state-trajectory. They are combined in a weighted sum and the combinations are switched appropriately to approximate the dynamic behaviour of the nonlinear system. Weights assigned at a specimen point on the trajectory generally depend on the euclidean distance to the linearisation points. In this work, limitations of the conventional weight-assignment scheme are pointed out. It is shown that the procedure is similar across all nonlinearities, and hence ignores the nonlinear vector field curvature for superposition. Additionally, it results in an inadequate assessment of the linear systems when they are equidistant from the specimen point. An improved method for weight-assignment, which uses state-velocities in addition to state-positions is proposed. The method naturally takes into account the system nonlinearity and is hence called Nonlinearity-aware Trajectory Piece-wise Linear (Ntpwl) method. Further, a computationally efficient procedure for estimating the state-velocity is introduced. The new strategy is illustrated and assessed with the help of case studies and it is shown that the Ntpwl model substantially improves the approximation of the nonlinear systems considered. Increased robustness to training and negligible stretching of the computational resources is also obtained.     

Laplacian group sparse modeling of human actions

Recently, many local-feature based methods have been proposed for feature learning to obtain a better high-level representation of human behavior. Most of the previous research ignores the structural information existing among local features in the same video sequences, while it is an important clue to distinguish ambiguous actions. To address this issue, we propose a Laplacian group sparse coding for human behavior representation. Unlike traditional methods such as sparse coding, our approach prefers to encode a group of relevant features simultaneously and meanwhile allow as less atoms as possible to participate in the approximation so that video-level sparsity is guaranteed. By incorporating Laplacian regularization the method is capable to ensure the similar approximation of closely related local features and the structural information is successfully preserved. Thus, a compact but discriminative human behavior representation is achieved. Besides, the objective of our model is solved with a closed-form solution, which reduces the computational cost significantly. Promising results on several popular benchmark datasets prove the efficiency and effectiveness of our approach.     

Matching actions in presence of camera motion

When the camera viewing an action is moving, the motion observed in the video not only contains the motion of the actor but also the motion of the camera. At each time instant, in addition to the camera motion, a different view of the action is observed. In this paper, we propose a novel method to perform action recognition in presence of camera motion. Proposed method is based on the epipolar geometry between any two views. However, instead of relating two static views using the standard fundamental matrix, we model the motions of independently moving cameras in the equations governing the epipolar geometry and derive a new relation which is referred to as the “temporal fundamental matrix.” Using the temporal fundamental matrix, a matching score between two actions is computed by evaluating the quality of the recovered geometry. We demonstrate the versatility of the proposed approach for action recognition in a number of challenging sequences.     

基于运动轨迹图形拟合的辅助人体 侧翻机构设计研究

摘 要：仰卧位侧翻是长期卧床患者有效预防压力性损伤病症的有效措施之一,研究人体 运动机理并进行辅助人体仰卧位侧翻机构的拟人化设计具有重要意义。人体肩部和臀部是辅助 仰卧位侧翻较为关键的施力位置,基于人体解剖学分析肩部和臀部骨骼的空间运动,建立运动 标记模型,并进行人体仰卧位侧翻运动捕捉实验,得到肩部和臀部的运动轨迹图形与连杆模型 特征信息。通过分析连杆长度和角度的变化过程进行图形几何分析与连杆机构简化,基于运动 轨迹图形拟合的辅助人体侧翻机构构型,并对机构进行运动学仿真分析,运动仿真结果和实验 图形的拟合误差较小,验证了所提出的辅助人体侧翻机构拟人化设计方法的合理性,相关研究 结果可为人-机运动协同性辅助侧翻康复辅具的设计提供理论依据。     

Real-time control of human actions using inertial sensors

Our study proposes a new local model to accurately control an avatar using six inertial sensors in real-time.Creating such a system to assist interactive control of a full-body avatar is challenging because control signals from our performance interfaces are usually inadequate to completely determine the whole body movement of human actors.We use a pre-captured motion database to construct a group of local regression models,which are used along with the control signals to synthesize whole body human movement.By synthesizing a variety of human movements based on actors’control in real-time,this study verifies the effectiveness of the proposed system.Compared with the previous models,our proposed model can synthesize more accurate results.Our system is suitable for common use because it is much cheaper than commercial motion capture systems.     

Two-dimensional object recognition by matching local properties of contour points

This paper describes a two-stage Hough procedure for two-dimensional object recognition based on matching local properties of a model set of contour points. The procedure makes use of three strategies: subdivision of the parameter space, classification by multiple properties, and non-maximum suppression. The first stage estimates likely values in a coarsely quantized space of position and orientation. The second stage verifies these estimates in a more precise space with a large set of model points. In both cases, matching of model points to image contour points is based on tangent, curvature and contrast. The evaluation of the parameter space is greatly simplified by simultaneous non-maximum suppression over neighborhoods of both parameters. Experiments show that these strategies are both efficient and robust, even in the case of complex and partial data. One example uses overlapping parts, such as might occur in industrial situations.     

模板匹配技术在图像识别中的应用

在图像目标识别技术的研究应用中,模板匹配技术是其中一个重要的研究方向,它具有算法简单、计算量小以及识别率高的特点。介绍了几种改进的模板匹配技术在图像处理、模式识别等领域的应用,包括有条码识别、生物特征识别技术(人脸识别、指纹识别等)、车牌识别、字符识别、飞机识别等。     

一种基于MEM-LBP的动作特征提取及识别方法*

针对动作识别中如何有效地利用人体运动的三维信息的问题,提出一种新的基于深度视频序列的特征提取和识别方法。该方法首先运用运动能量模型（MEM）来表征人体动态特征,即先将整个深度视频序列投影到三个正交的笛卡儿平面上,再把每个投影面的视频系列划分为能量均等的子时间序列,分别计算子序列的深度运动图能量从而得到运动能量模型（MEM）。然后利用局部二值模式（LBP）描述符对运动能量模型编码,进一步提取人体运动的有效信息。最后用 范数协同表示分类器进行动作分类识别。在MSRAction3D、MSRGesture3D数据库上测试所提方法,实验结果表明该方法有较高的识别效果。     

Model-based segmentation and recognition of dynamic gestures in continuous video streams

Segmentation and recognition of continuous gestures are challenging due to spatio-temporal variations and endpoint localization issues. A novel multi-scale Gesture Model is presented here as a set of 3D spatio-temporal surfaces of a time-varying contour. Three approaches, which differ mainly in endpoint localization, are proposed: the first uses a motion detection strategy and multi-scale search to find the endpoints; the second uses Dynamic Time Warping to roughly locate the endpoints before a fine search is carried out; the last approach is based on Dynamic Programming. Experimental results on two arm and single hand gestures show that all three methods achieve high recognition rates, ranging from 88% to 96% for the two arm test, with the last method performing best.     

Invariant trajectory classification of dynamical systems with a case study on ECG

An invariant pattern recognition framework for classification of phase space trajectories of nonlinear dynamical systems is presented. Using statistical shape theory, known external influences can be discriminated from true changes of the system. The external effects are modeled as a transformation group acting on the phase space, and variation of the trajectories not explained by the transformations is accounted for using principal component analysis. The approach suggested is highly adaptable to a wide range of situations and individual differences.The methodology presented is applied to detect abnormalities in electrocardiograms. Results based on measured data indicate that the model developed is resistant to the effects of respiration and body position changes, which are abundant in ambulatory conditions and cause significant morphological artifacts in the signal. The results also show that the detection of an artificially induced acute myocardial infarction is achieved with high performance. Due to its low computational complexity, the method developed can be implemented in real-time. The method developed also adapts to morphological changes caused by various heart conditions.     

特征区域模板匹配法实现汽车牌照的精确识别

模板匹配法在汽车牌照自动识别中已得到广泛应用。在应用该法为某汽车检测中心做牌照自动识别系统时发现，0、D及Q3个字符互相误判率较高。为解决此问题，在模板匹配法的基础上，提出了特征区域模板匹配法，应用该法取得了满意的效果，识别精度大大提高。另外，匹配系数的求法也采用了新的方法，效果比原方法好。