Citation Context

... interpolation between existing motions. If the latter is required, multiresolution filtering has to be combined with time-warping techniques (Witkin & Popovic, 1995). Time warps are required to align corresponding signal features in time. Depending on the complexity of the action, time warps are parameterized in terms of simple uniform scaling and translation (e.g., Wiley & Hahn, 1997; Yacoob & Black, 1997) by using nonlinear models, such as B-splines (Ramsay & Li, 1989; Ramsay & Silverman, 1997), or by fitting nonparametric models by means of dynamic programming (Bruderlin & Williams, 1995; Giese & Poggio, 1999; 2000). Our approach to linearize human walking data employs many of the techniques summarized above. Starting with motion capture data from a number of human subjects, we first reduce the dimensionality of each subject’s set of postures using PCA in a way similar to that described by Rosales and Scarloff (2000). This results in a low-dimensional space spanned by the first few eigenpostures. As postures change during walking, the corresponding coefficients change sinusoidally. The temporal behaviour of the sequence is well described by simple sine functions, and the decomposition becomes very...

Citation Context

... the two sequences into alignment. No static-background information was used in this step 1 . This approach is hence referred to in our paper as "trajectory-to-trajectory alignment". Giese a=-=nd Poggio [3]-=- also used trajectory-to-trajectory alignment 1 In a later step [6] refines the spatial alignment using static background information. However, the temporal alignment is already fixed at that point. t...

Citation Context

...ilar appearance properties [4]). Moreover, none of these methods is applicable to alignment of actions performed at different times and places. Methods for aligning actions were also suggested (e.g., =-=[10, 6, 2]-=-). However, these require manual selection of corresponding feature points across the sequences. Some of them provide only temporal synchronization. In [11] an approach was proposed for detecting beha...

Citation Context

... The image sequences were generated using MATLAB. These three motion patterns could be easily distinguished by human observers, and also could be classified correctly with a computer vision algorithm =-=[7]. Figure 2 shows the-=- neural fields that represent the "walking" and the "running" prototype in the form pathway during the presentation of a walking stimulus. The upper part of the figure shows 0 20 4...

Citation Context

... ! method spatio-temporal correspondence at a coarse level is established. 2.2 Morphable Models for modeling and concatenation of movement primitives The technique of spatio-temporal morphable models =-=[8, 10]-=- is based on linearly combining the movement trajectories of prototypical motion patterns in space-time. Linear combinations of movement patterns are defined on the basis of spatio-temporal correspond...

Citation Context

...mouse, the device standardizes and masks many of these differences, but because our vision system directly senses and interprets a user’s movements, it can be sensitive to them. Techniques similar to =-=[17]-=- can allow adaptation to an individual user’s movement patterns. One constraint on any interaction system is that it be able to respond to the user smoothly and in a timely manner. For a vision-based ...

Citation Context

...xperiments are based on a special technique for spatio-temporal morphing. This technique, called ”spatio-temporal morphable models” (STMMs), has been applied before in the context of computer vision (=-=Giese and Poggio, 1999-=-, 2000a,b) and provides a possibility to generate new artificial biological movement patterns by linear combination of prototypical example movements. By changing the weights of the linear combination...

Citation Context

...� of over all matched key features. A more formal description of the algorithm is given in [9]. 2.2 Morphable Models for modeling movement primitives The technique of Spatio-Temporal Morphable Model=-=s [7]-=-,[8] is based on linearly combining the movement trajectories of prototypical motion patterns in space-time. Linear combinations of movement patterns are defined on the basis of spatio-temporal corres...