Synthesizing physically realistic human motion in low-dimensional, behavior-specific spaces

By Alla Safonova, Jessica K. Hodgins, Nancy S. Pollard. ACM Transactions on Graphics, 2004

541 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Control and Systems Engineering

View Paper: https://www.semanticscholar.org/paper/4b777ab9788aeed6954e44e4318bee56d45023a9

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Control and Systems Engineering

Citations in the corpus

Robust Estimation of 3D Human Poses from a Single Image (2014, background): “Although human poses are highly variant, they lie in a low dimensional space [13].”
Terrain-adaptive bipedal locomotion control (2010, methodology): “Offline methods can be used to synthesize animation of bipedal locomotion [Rose et al. 1996; Safonova et al. 2004; Liu et al. 2005].”
Performance animation from low-dimensional control signals (2005, methodology): “Global models might be appropriate for applications such as synthesis of motion without a continuous driving signal (for example, [Safonova et al. 2004]), but given the temporal coherence of the control signals of our performance animation system, they were not required.”
Reconstructing 3D Human Pose from 2D Image Landmarks (2012, background): “It has been successfully applied to constrained deformable objects, such as faces [4] and action-speciﬁc body reconstruction, such as walking, [5]. we add poses from varied actions, the complexity of the distribution of poses increases and, consequently, the dimensionality of the reduced model…”
Dextrous manipulation from a grasping pose (2009, methodology): “Optimization-based approaches, similar to our algorithm can produce physically realistic motion with active control [Witkin and Kass 1988; Safonova et al. 2004; Liu et al. 2005].”
Character controllers using motion VAEs (2020, methodology): “Motion models may also be implicitly defined using Principle Component Analysis (PCA) of motion exemplars, e.g., [Safonova et al. 2004].”
Optimal gait and form for animal locomotion (2009, methodology): “Some techniques have been developed to alleviate the need for an initial motion by either using a database of similar motions as in [Safonova et al. 2004] or by utilizing a specialized dynamics formulation as in [Fang and Pollard 2003], which has produced reasonable flipping and tumbling motions wit”
Mode-adaptive neural networks for quadruped motion control (2018, methodology): “…techniques based on machine learning such as K-Nearest Neigh-bours (KNN), principal component analysis [Chai and Hodgins 2005; Min and Chai 2012; Safonova et al. 2004; Tautges et al. 2011], radial basis functions (RBF) [Kovar and Gleicher 2004; Rose et al. 1998], reinforcement learning…”
Linear Bellman combination for control of character animation (2009, methodology): “[Witkin and Kass 1988; Ngo and Marks 1993; Safonova et al. 2004] to minimize applied force or deviation from desired poses.”
Construction and optimal search of interpolated motion graphs (2007, background): “Continuous optimization has been shown to work well when a good initial guess is provided and for synthesizing relatively short, single behavior motions, such as jumps and runs (see for example [Fang and Pollard 2003; Safonova et al. 2004; Sulejmanpašić and Popović 2005]).”