The reuse of human motion capture data to create new, realistic motions by applying morphing and blending techniques has become an important issue in computer animation. This requires the identification and extraction of logically related motions scattered within some data set. Such content-based retrieval of motion capture data, which is the topic of this paper, constitutes a difficult and timeconsuming problem due to significant spatio-temporal variations between logically related motions. In our approach, we introduce various kinds of qualitative features describing geometric relations between specified body points of a pose and show how these features induce a time segmentation of motion capture data streams. By incorporating spatio-temporal invariance into the geometric features and adaptive segments, we are able to adopt efficient indexing methods allowing for flexible and efficient content-based retrieval and browsing in huge motion capture databases. Furthermore, we obtain an efficient preprocessing method substantially accelerating the cost-intensive classical dynamic time warping techniques for the time alignment of logically similar motion data streams. We present experimental results on a test data set of more than one million frames, corresponding to 180 minutes of motion. The linearity of our indexing algorithms guarantees the scalability of our results to much larger data sets.

The matching of two-dimensional shapes is an important problem with applications in domains as diverse as biometrics, industry, medicine and anthropology. The distance measure used must be invariant to many distortions, including scale, offset, noise, partial occlusion, etc. Most of these distortions are relatively easy to handle, either in the representation of the data or in the similarity measure used. However rotation invariance seems to be uniquely difficult. Current approaches typically try to achieve rotation invariance in the representation of the data, at the expense of discrimination ability, or in the distance measure, at the expense of efficiency. In this work we show that we can take the slow but accurate approaches and dramatically speed them up. On real world problems our technique can take current approaches and make them four orders of magnitude faster, without false dismissals. Moreover, our technique can be used with any of the dozens of existing shape representations and with all the most popular distance measures including Euclidean distance, Dynamic Time Warping and Longest Common Subsequence.

We present a technique to automatically distill a motion-motif graph from an arbitrary collection of motion capture data. Motion motifs represent clusters of similar motions and together with their encompassing motion graph they lend understandable structure to the contents and connectivity of large motion datasets. They can be used in support of motion compression, the removal of redundant motions, and the creation of blend spaces. This paper develops a string-based motif-finding algorithm which allows for a user-controlled compromise between motif length and the number of motions in a motif. It allows for time warps within motifs and assigns the majority of the input data to relevant motifs. Results are demonstrated for large datasets (more than 100,000 frames) with computation times of tens of minutes.

In the past two decades, motion capture (mocap) systems have been developed that allow to track and record human motions at high spatial and temporal resolutions. The resulting motion capture data is used to analyze human motions in fields such as sports sciences and biometrics (person identification), and to synthesize realistic motion sequences in datadriven computer animation. Such applications require efficient methods and tools for the automatic analysis, synthesis and classification of motion capture data, which constitutes an active research area with many yet unsolved problems. Even though there is a rapidly growing corpus of motion capture data, the academic research community still lacks publicly available motion data, as supplied by [4], that can be freely used for systematic research on motion analysis, synthesis, and classification. Furthermore, a common dataset of annotated and well-documented motion capture data would be extremely valuable to the research community in view of an objective comparison and evaluation of the achieved research results. It is the objective of our motion capture database HDM05 1 to supply free motion capture data for research purposes. HDM05 contains more than tree hours of systematically recorded and well-documented motion

This work introduces distance-based criteria for segmentation of object trajectories. Segmentation leads to simplification of the original objects into smaller, less complex primitives that are better suited for storage and retrieval purposes. Previous work on trajectory segmentation attacked the problem locally, segmenting separately each trajectory of the database. Therefore, they did not directly optimize the inter-object separability, which is necessary for mining operations such as searching, clustering, and classification on large databases. In this paper we analyze the trajectory segmentation problem from a global perspective, utilizing data aware distance-based optimization techniques, which optimize pairwise distance estimates hence leading to more efficient object pruning. We first derive exact solutions of the distance-based formulation. Due to the intractable complexity of the exact solution, we present an approximate, greedy solution that exploits forward searching of locally optimal solutions. Since the greedy solution also imposes a prohibitive computational cost, we also put forward more lightweight variance-based segmentation techniques, which intelligently “relax ” the pairwise distance only in the areas that affect the least the mining operations.

We demonstrate a data-driven approach for representing, compressing, and indexing human-motion databases. Our modeling approach is based on piecewise-linear components that are determined via a divisive clustering method. Selection of the appropriate linear model is determined automatically via a classifier using a subspace of the most significant, or principle features (markers). We show that, after offline training, our model can accurately estimate and classify human motions. We can also construct indexing structures for motion sequences according to their transition trajectories through these linear components. Our method not only provides indices for whole and/or partial motion sequences, but also serves as a compressed representation for the entire motion database. Our method also tends to be immune to temporal variations, and thus avoids the expense of time-warping.

Abstract. This study proposes a roadmap for the creation and specification of a virtual humanoid capable of performing expressive gestures in real time. We present a gesture motion data acquisition protocol capable of handling the main articulators involved in human expressive gesture (whole body, fingers and face). The focus is then shifted to the postprocessing of captured data leading to a motion database complying with our motion specification language and capable of feeding data driven animation techniques. Issues. Embodying a virtual humanoid with expressive gestures raises many problems such as computation-cost efficiency, realism and level of expressiveness, or high level specification of expressive gesture [1]. Here, we focus on the acquisition of motion capture data from the main articulators involved in communicative gesture (whole body, face mimics and finger motion). We then show how acquired data are postprocessed in order to build a database compatible with high level gesture specification and capable of feeding real time data-driven

Fast searching of content in large motion databases is essential for efficient motion analysis and synthesis. In this work we demonstrate that identifying locally similar regions in human motion data can be practical even for huge databases, if medium-dimensional (15–90 dimensional) feature sets are used for kd-tree-based nearest-neighborsearches. On the basis of kd-tree-based local neighborhood searches we devise a novel fast method for global similarity searches. We show that knn-searches can be used efficiently within the problems of (a) “numerical and logical similarity searches”, (b) reconstruction of motions from sparse marker sets, and (c) building so called “fat graphs”, tasks for which previously algorithms with preprocessing time quadratic in the size of the database and thus only applicable to small collections of motions had been presented. We test our techniques on the two largest freely available motion capture databases, the CMU and HDM05 motion databases comprising more than 750 min of motion capture data proving that our approach is not only theoretically applicable but also solves the problem of fast similarity searches in huge motion databases in practice.

An interesting issue in moving objects databases is to find similar trajectories of moving objects. Previous work on this topic focuses on movement patterns (trajectories with time dimension) of moving objects, rather than spatial shapes (trajectories without time dimension) of their trajectories. In this paper we propose a simple and effective way to compare spatial shapes of moving object trajectories. We introduce a new distance function based on “one way distance” (OWD). Algorithms for evaluating OWD in both continuous (piece wise linear) and discrete (grid representation) cases are developed. An index structure for OWD in grid representation, which guarantees no false dismissals, is also given to improve the efficiency of similarity search. Empirical studies show that OWD out-performs existent methods not only in precision, but also in efficiency. And the results of OWD in continuous case can be approximated by discrete case efficiently.

Time series motifs are approximately repeated patterns found within the data. Such motifs have utility for many data mining algorithms, including rule-discovery, novelty-detection, summarization and clustering. Since the formalization of the problem and the introduction of efficient linear time algorithms, motif discovery has been successfully applied to many domains, including medicine, motion capture, robotics and meteorology. In this work we show that most previous applications of time series motifs have been severely limited by the definition’s brittleness to even slight changes of uniform scaling, the speed at which the patterns develop. We introduce a new algorithm that allows discovery of time series motifs with invariance to uniform scaling, and show that it produces objectively superior results in several important domains. Apart from being more general than all other motif discovery algorithms, a further contribution of our work is that it is simpler than previous approaches, in particular we have drastically reduced the number of parameters that need to be specified.