## Generating Novel, Stylistically Consonant Variations on Human Movement Sequences (2007)

Citations: | 1 - 0 self |

### Citations

3775 | Eigenfaces for recognition
- Turk, Pentland
- 1991
(Show Context)
Citation Context ... a collection of still images—and our goal is to model and use the progressions and correlations in those data. Representations are always key, and several groups (e.g., Mataric et al.[27] or Pentland=-=[56]-=-) have focused on how to construct good primitives for representing movement. Our representations and algorithms, in contrast, are not intended to help humans understand motion or style; they are desi... |

2335 | A Note on Two Problems in Connexion with Graphs
- Dijkstra
- 1959
(Show Context)
Citation Context ...ly sum the scores of the single-keyframe whole-body movements that make up that sequence. To estimate the score for the path from some intermediate state to the goal state, we use Dijkstra’s algorithm=-=[23]-=- to find the shortest-path costs for each joint and then sum those costs. As required for A*, this heuristic will always underestimate the true cost to the goal state; see the appendix for a short pro... |

2039 |
Computational Geometry: An Introduction
- Preparata, Shamos
- 1985
(Show Context)
Citation Context ...mapping to be useful for generating variations, these patches must cover the attractor and they must not intersect. We chose to implement this patchwise division of state space using a Voronoi diagram=-=[46]-=- on the reference trajectory points, as shown in Figure 4(b). A Voronoi cell around one of these points is the region of state space that is closer to that point than to any other. One draws Voronoi c... |

1499 |
Deterministic nonperiodic flow
- Lorenz
- 1963
(Show Context)
Citation Context ...s results. 3.1 Chaos 101 Chaos is a type of complex and yet highly patterned behavior that arises in deterministic dynamical systems 2 . One of the canonical examples in the field is the Lorenz system=-=[43]-=-, a set of three coupled ordinary differential equations that comprise a highly simplified model of flow in a heated fluid: ˙x(t) = 16y(t) − 16x(t) ˙y(t) = 45x(t) − y(t) − x(t)z(t) (1) ˙z(t) = x(t)y(t... |

1207 |
Statistical principles in experimental design
- Winer
- 1971
(Show Context)
Citation Context ...ror Terms (AETs) for each of the three effects tested were established using the methodology described by Scheffe[51]; the QuasiF ratio for Method was constructed using the approach described by Winer=-=[59]-=-. As the data were ordinal in nature, the raw scores generated by each subject were transformed (XT ) using the square root transformation recommended by Dixon and Massey[24] prior to executing the AN... |

1164 |
Classical Mechanics
- Goldstein, Poole, et al.
- 2002
(Show Context)
Citation Context ... as four rigid segments rather than 24 individual vertebrae. The orientation of each joint is specified with a quaternion, a standard representation in rigid-body mechanics that dates back to Hamilton=-=[30]-=-. A quaternion q = (r, �u) consists of an axis of rotation �u and a scalar r that specifies the angle of rotation of the joint about �u. A single body position, in this representation, translates to 2... |

830 |
Nonlinear Dynamics and Chaos
- Strogatz
- 1994
(Show Context)
Citation Context ...attern in the state space remains the same. • they cannot be calculated in closed form, but rather the equations must be solved numerically in order to produce them. See [12, 53] for introductions or =-=[54]-=- for a more-comprehensive treatment of these concepts. The critical features of chaos, for the purposes of this paper, are the first two bullets of the list above. If one starts the Lorenz system from... |

829 |
Experimental Design: Procedures for the Behavioral Sciences. Brooks/Cole
- Kirk
- 1995
(Show Context)
Citation Context ...by Winer[59]. For each of those cases where a random effect was found to be significant, an Intraclass Correlation coefficient (ρI) was calculated as a measure of the importance of the observed effect=-=[37]-=-. This was done only to facilitate the interpretation of the results associated with the Method effect. Table 1 presents the results of the ANOVA for Question 1: ‘Awkwardness.’ As shown by this analys... |

755 | An algorithm for finding best matches in logarithmic expected time
- Friedman, Bentley, et al.
- 1977
(Show Context)
Citation Context ...e can express a body position � b as a discretized vector �s by setting each of its components sλ equal to the quaternion in Q λ that is closest to bλ. We can do this in log(M λ ) time using K-D trees=-=[28]-=- to represent the Q λ sets. The strategy described at the end of the previous paragraph is analogous to “snapping” objects to a grid in computer-drawing applications. While this quantization is useful... |

654 | Eigentracking, “Robust matching and tracking of articulated objects using a view-based representation
- Black, Jepson
- 1996
(Show Context)
Citation Context ... kinematics problems. There have been hundreds of good papers on recognizing, analyzing, quantifying, and understanding various aspects of motion, especially gait (e.g., [40]) and hand gesture (e.g., =-=[10]-=-). We are interested in how the whole body moves, as well as how it moves differently in different movement genres, and our goal is to generate original movement that follows a given style, not to rec... |

621 |
An inequality and associated maximation technique in statistical estimation for probabilistic functions of markov processes
- Baum
- 1972
(Show Context)
Citation Context ...ends itself neatly to a dynamic programming solution, which would make successor state generation efficient when the set of possible joint orientations is large. A Baum-Welch forward and backward pass=-=[9]-=- can find the top ν most probable body states in time K = 2 ∗ 23 ∗ O(b + ν) instead of the exponentially many in the worst-case example above. Rather than score all possible successors of the current ... |

572 |
The Analysis of Variance
- Scheffe
- 1959
(Show Context)
Citation Context ...d Sequence, as well as their Interaction, were nested within Method. The Appropriate Error Terms (AETs) for each of the three effects tested were established using the methodology described by Scheffe=-=[51]-=-; the QuasiF ratio for Method was constructed using the approach described by Winer[59]. As the data were ordinal in nature, the raw scores generated by each subject were transformed (XT ) using the s... |

538 |
A model for reasoning about persistence and causation
- Dean, Kanazawa
- 1989
(Show Context)
Citation Context ...ture kinesiology and style, and it could even elucidate hidden linkages (e.g., a hip injury that causes unusual shoulder movement). We are currently exploring this idea using dynamic Bayesian networks=-=[22]-=- and a corpus of motion-capture data that is described at more length in the conclusion. Note that none of this analysis and representation of inter-joint coordination would be needed if we had used a... |

455 |
Digital Image Warping
- Wolberg
- 1990
(Show Context)
Citation Context ... interpolate between the discretized, quaternion-valued representations �sA and �sB that correspond to these positions. The morphing techniques that have been so effective in computer graphics (e.g., =-=[61]-=-) do exactly that; as mentioned on page 10, the animation software that we use to produce movies also uses splines to interpolate or ‘tween’ between its keyframes. Simple interpolation techniques, how... |

377 | Motion graphs
- Kovar, Gleicher, et al.
- 2002
(Show Context)
Citation Context ...e joint assumes and to estimate the corresponding transition probabilities between orientations. Note that our graphs are finer grained than the “motion graphs” that are used in the graphics community=-=[39]-=-, whose vertices represent temporal subsequences of the original motion and whose edges capture transitions between those clips. Our graphs capture a jointwise decomposition of the motion. To build th... |

345 | Animating human athletics
- Hodgins, Wooten, et al.
- 1995
(Show Context)
Citation Context ...of algorithm that captures both real physical constraints and behavioral patterns is necessary; building such an algorithm a priori, however—as in the work of Jessica Hodgins and collaborators (e.g., =-=[33, 34]-=-)—is extremely difficult. Another approach is to use machine-learning techniques to build models of movement patterns from corpora of human motion, and that is what we have chosen to do. A critical de... |

308 | Period three implies chaos
- Li, Yorke
- 1975
(Show Context)
Citation Context ... effect upon the evolution of a chaotic system. Edward Lorenz first reported this behavior in a 1963 paper[43] entitled “Deterministic Nonperiodic Flow.” The term “chaos” was coined twelve years later=-=[41]-=-. This fixed attractor structure provides an element of order and predictability in a chaotic system: any trajectory that is started in an attractor’s basin will follow the same overall, time-asymptot... |

292 | Style machines
- Brand, Hertzmann
- 2000
(Show Context)
Citation Context ... other strategies have been proposed specifically for learning movement style from a corpus of examples: using singular-value decomposition[58], principal-components analysis[57], hidden Markov models=-=[16]-=-, probability distributions over the space of all possible poses[31], linear models[35], and even nonlinear optimization techniques[42]. The representation used in our work (joint graphs + Bayesian ne... |

287 |
An Introduction to Statistical Analysis
- Dixon, Massey
- 1951
(Show Context)
Citation Context ...pproach described by Winer[59]. As the data were ordinal in nature, the raw scores generated by each subject were transformed (XT ) using the square root transformation recommended by Dixon and Massey=-=[24]-=- prior to executing the ANOVA. Complete appreciation of the results of the three ANOVA tables that follow requires an understanding that an algebraic combination of the Class and Sequence effects, as ... |

280 | Learning the structure of dynamic probabilistic networks
- Friedman, Murphy, et al.
- 1998
(Show Context)
Citation Context ...ndamental to movement, and one really should learn it from the data, rather than assuming its structure a priori as we did. To this end, we are currently exploring the use of dynamic Bayesian networks=-=[29]-=- to extract these coordination patterns automatically from mocap data. It will be important to learn not only which pairs of joints are related, but also examine combinations of joint movement. In bal... |

207 | Z.: Style-based inverse kinematics
- GROCHOW, MARTIN, et al.
- 2004
(Show Context)
Citation Context ...ent style from a corpus of examples: using singular-value decomposition[58], principal-components analysis[57], hidden Markov models[16], probability distributions over the space of all possible poses=-=[31]-=-, linear models[35], and even nonlinear optimization techniques[42]. The representation used in our work (joint graphs + Bayesian networks) has a significant advantage over these approaches in that it... |

193 | Synthesizing physically realistic human motion in low-dimensional, behavior-specific spaces. SIGGRAPH’04
- Safonova, Hodgins, et al.
(Show Context)
Citation Context ...quation. 22smotion occupies only a subset of the high-dimensional space: “most dynamic human behaviors are intrinsically low dimensional with, for example, legs and arms operating in a coordinated way=-=[50]-=-”. Note that stylistically rooted representations like Feuillet notation avoid all of these issues completely because a huge amount of coordination is implicit in the packets of movement that are capt... |

178 |
Arti Intelligence
- Winston
- 1992
(Show Context)
Citation Context ... 6, generate free-form movement that follow the corpus’s patterns by walking these graphs in different ways. 4.2 Using Joint Movement Patterns to Interpolate We use a memory-bounded A* search strategy=-=[60]-=- to find an interpolation subsequence that moves smoothly between two discretized body postures �sA and �sB. Recall that A* finds a path from an initial state to a goal state by progressively generati... |

147 | Gait analysis for recognition and classification
- Lee, Grimson
(Show Context)
Citation Context ...ts that arise in, say, inverse kinematics problems. There have been hundreds of good papers on recognizing, analyzing, quantifying, and understanding various aspects of motion, especially gait (e.g., =-=[40]-=-) and hand gesture (e.g., [10]). We are interested in how the whole body moves, as well as how it moves differently in different movement genres, and our goal is to generate original movement that fol... |

127 | Learning physics-based motion style with nonlinear inverse optimization
- Liu, Hertzmann, et al.
- 2005
(Show Context)
Citation Context ...tion[58], principal-components analysis[57], hidden Markov models[16], probability distributions over the space of all possible poses[31], linear models[35], and even nonlinear optimization techniques=-=[42]-=-. The representation used in our work (joint graphs + Bayesian networks) has a significant advantage over these approaches in that it allows us to make good use of having multiple instances of a given... |

124 | Automated derivation of primitives for movement classification
- Fod, Matarić, et al.
- 2002
(Show Context)
Citation Context ...dy position from a collection of still images—and our goal is to model and use the progressions and correlations in those data. Representations are always key, and several groups (e.g., Mataric et al.=-=[27]-=- or Pentland[56]) have focused on how to construct good primitives for representing movement. Our representations and algorithms, in contrast, are not intended to help humans understand motion or styl... |

116 | Efficient synthesis of physically valid human motion
- Fang, Pollard
- 2003
(Show Context)
Citation Context ...l ways—e.g, to introduce a limp into a walk. Other approaches that have proved to be useful in creating and/or adapting “natural-looking” motion include optimization of appropriate objective functions=-=[25]-=-, combinations of IK and optimization[45], dynamical modeling & control theory[34], and detailed musculoskeletal modeling[38]. All of these approaches devote significant analytical effort to the model... |

102 | Synthesizing animations of human manipulation tasks
- Yamane, Kuffner, et al.
- 2004
(Show Context)
Citation Context ... motion—let alone understand or decompose it in any detail. Several other groups have produced interesting results in this area. Some of this work, not surprisingly, has used inverse kinematics (IK). =-=[62]-=-, for instance, uses planning and a data-driven constraint-based IK to attain naturalness. [44] also uses IK, tuning a couple of its parameters (joint stiffness, hip swing) to alter the style of a giv... |

88 | Style translation for human motion
- Hsu, Pulli, et al.
- 2005
(Show Context)
Citation Context ...rpus of examples: using singular-value decomposition[58], principal-components analysis[57], hidden Markov models[16], probability distributions over the space of all possible poses[31], linear models=-=[35]-=-, and even nonlinear optimization techniques[42]. The representation used in our work (joint graphs + Bayesian networks) has a significant advantage over these approaches in that it allows us to make ... |

77 |
Bayesian networks for knowledge discovery,” Advances in knowledge discovery and data mining
- Heckerman
- 1996
(Show Context)
Citation Context ...n of the wrist, for instance, strongly affects the orientation and movement of the fingers but has little effect on the toes. We put this simplifying assumption into effect by using a Bayesian network=-=[32]-=- to explicitly represent the relationships between how different joints move. This network reflects the structure and physics of the human body: a tree with the pelvis at the root. Three branches lead... |

67 | A data-driven approach to quantify natural human motion
- Ren, Patrick, et al.
- 2005
(Show Context)
Citation Context ... thus easier to apply to a new body—one that may violate a priori modeling assumptions in subtle ways. The work of the Graphics Lab at Carnegie Mellon is probably the most closely related to ours. In =-=[49]-=-, for instance, motion is decomposed by body part, modeled individually using different machine-learning techniques, and combined via an ensemble method that strives for naturalness. Our techniques ar... |

58 |
Does God play dice?: the mathematics of chaos
- Stewart
- 1989
(Show Context)
Citation Context ... the overall time-asymptotic pattern in the state space remains the same. • they cannot be calculated in closed form, but rather the equations must be solved numerically in order to produce them. See =-=[12, 53]-=- for introductions or [54] for a more-comprehensive treatment of these concepts. The critical features of chaos, for the purposes of this paper, are the first two bullets of the list above. If one sta... |

50 |
editor): The Harvard Dictionary of Music
- Randel
- 1986
(Show Context)
Citation Context ...ic, the term commonly means elaboration of the melody or accompaniment; other kinds of modifications, such as development or transformation, are often considered to be outside the scope of variation.”=-=[48]-=-. 4 Recall that frames in this sequence are evenly spaced in time. 10soccurred in Cunningham’s phrase-shuffled work as well, and caused it to be met by substantial resistance from audiences, dancers, ... |

46 | Human motion signatures: Analysis, synthesis, recognition
- Vasilescu
(Show Context)
Citation Context ...ethods require a bit less tuning. In the past few years, some other strategies have been proposed specifically for learning movement style from a corpus of examples: using singular-value decomposition=-=[58]-=-, principal-components analysis[57], hidden Markov models[16], probability distributions over the space of all possible poses[31], linear models[35], and even nonlinear optimization techniques[42]. Th... |

24 |
Style-based motion synthesis
- Urtasun, Glardon, et al.
(Show Context)
Citation Context ...n the past few years, some other strategies have been proposed specifically for learning movement style from a corpus of examples: using singular-value decomposition[58], principal-components analysis=-=[57]-=-, hidden Markov models[16], probability distributions over the space of all possible poses[31], linear models[35], and even nonlinear optimization techniques[42]. The representation used in our work (... |

20 | Autonomous exploration and control of chaotic systems
- Bradley
- 1995
(Show Context)
Citation Context ...scale “mixing” is not only ubiquitous in science and engineering, but also highly intriguing, and it has a variety of practical applications, ranging from spacecraft control to heart-attack prevention=-=[11, 52]-=-. The following section explains how to exploit these properties to create chaotic variations on movement sequences. 7s(a) (b) Figure 4: Generating a tiling of a chaotic attractor: A Voronoi diagram i... |

20 | Methods for Exploring Expressive Stance
- Neff, Fiume
- 2006
(Show Context)
Citation Context ...lk. Other approaches that have proved to be useful in creating and/or adapting “natural-looking” motion include optimization of appropriate objective functions[25], combinations of IK and optimization=-=[45]-=-, dynamical modeling & control theory[34], and detailed musculoskeletal modeling[38]. All of these approaches devote significant analytical effort to the modeling process, whereas we learn the models ... |

16 |
Adapting motion capture data using weighted real-time inverse kinematics
- Meredith, Maddock
- 2005
(Show Context)
Citation Context ...interesting results in this area. Some of this work, not surprisingly, has used inverse kinematics (IK). [62], for instance, uses planning and a data-driven constraint-based IK to attain naturalness. =-=[44]-=- also uses IK, tuning a couple of its parameters (joint stiffness, hip swing) to alter the style of a given movement sequence in physically meaningful ways—e.g, to introduce a limp into a walk. Other ... |

11 | Using chaos to generate variations on movement sequences
- Bradley, Stuart
- 1998
(Show Context)
Citation Context ...orward, but capturing the state of the human body is far more complicated, as described in Section 2. The mathematics of our mapping is different from Dabby’s in some formal ways that are detailed in =-=[14, 15]-=-. And while musical instruments can play arbitrary pitch sequences, subject to instrument range and performer ability, kinesiology and style impose a variety of constraints on consecutive body posture... |

10 |
Nonlinear maps as generators of musical design
- Pressing
- 1988
(Show Context)
Citation Context ... Section 1, chaos has been used to create musical variations, and that was the catalyst for the project described in this paper. Chaos has also been used to generate music from scratch as well (e.g., =-=[47]-=-), but the results are not at all consonant with any established musical style. There are a few other groups working at the intersection of computer science and dance, including NYU, Arizona State, an... |

10 |
Progress in the Control of Chaos
- Shinbrot
- 1995
(Show Context)
Citation Context ...scale “mixing” is not only ubiquitous in science and engineering, but also highly intriguing, and it has a variety of practical applications, ranging from spacecraft control to heart-attack prevention=-=[11, 52]-=-. The following section explains how to exploit these properties to create chaotic variations on movement sequences. 7s(a) (b) Figure 4: Generating a tiling of a chaotic attractor: A Voronoi diagram i... |

9 |
Merce Cunningham: The Modernizing of Modern Dance
- Copeland
- 2004
(Show Context)
Citation Context ...ced the idea that movement could be decomposed and manipulated by means other than the kinesthetic logic that is rooted in the body’s neuromuscular system, thereby removing much of its thematic nature=-=[18]-=-. One of Cunningham’s strategies was to compose motion sequences for different quadrants of the body and then combine them in arbitrary ways; in another, he used coin tosses or other randomization tec... |

9 |
Musical variations from a chaotic mapping
- Dabby
- 1996
(Show Context)
Citation Context ...sequence and chaos’s sensitive dependence on initial conditions can be used to produce variations. This intriguing notion was proposed by Diana Dabby in the mid-1990s in the context of music and image=-=[20, 21]-=-. Inspired by Dabby’s work, we set out to apply some similar ideas to dance. The result was the pair of tools that are described in this paper: Chaographer, which produces chaotic variations of keyfra... |

9 | Learning the grammar of dance
- Stuart, Bradley
- 1998
(Show Context)
Citation Context ...entations observed in different keyframes to obtain body-posture sequences that are novel, kinesiologically valid, and stylistically consonant. A preliminary version of this strategy was described in =-=[55]-=-; the rest of this section describes our final version and assesses the results using a simple Turing Test. 4.1 Capturing Joint Movement Patterns We capture the patterns in a joint’s motion using a tr... |

7 | Using chaos to generate choreographic variations
- Bradley, Stuart
- 1997
(Show Context)
Citation Context ...orward, but capturing the state of the human body is far more complicated, as described in Section 2. The mathematics of our mapping is different from Dabby’s in some formal ways that are detailed in =-=[14, 15]-=-. And while musical instruments can play arbitrary pitch sequences, subject to instrument range and performer ability, kinesiology and style impose a variety of constraints on consecutive body posture... |

6 |
Attaching physiological effects to motion-capture data
- Komura, Shinagawa
(Show Context)
Citation Context ...ral-looking” motion include optimization of appropriate objective functions[25], combinations of IK and optimization[45], dynamical modeling & control theory[34], and detailed musculoskeletal modeling=-=[38]-=-. All of these approaches devote significant analytical effort to the modeling process, whereas we learn the models from the motion-capture data. This makes our approach better at capturing the vagari... |

5 |
A chaotic mapping for musical and image variation
- Dabby
- 1997
(Show Context)
Citation Context ...sequence and chaos’s sensitive dependence on initial conditions can be used to produce variations. This intriguing notion was proposed by Diana Dabby in the mid-1990s in the context of music and image=-=[20, 21]-=-. Inspired by Dabby’s work, we set out to apply some similar ideas to dance. The result was the pair of tools that are described in this paper: Chaographer, which produces chaotic variations of keyfra... |

4 | Causes and effects of chaos
- Bradley
- 1990
(Show Context)
Citation Context ... the overall time-asymptotic pattern in the state space remains the same. • they cannot be calculated in closed form, but rather the equations must be solved numerically in order to produce them. See =-=[12, 53]-=- for introductions or [54] for a more-comprehensive treatment of these concepts. The critical features of chaos, for the purposes of this paper, are the first two bullets of the list above. If one sta... |

3 |
Computers and choreography
- Bradley, Capps, et al.
- 1999
(Show Context)
Citation Context .... In our experience, the dance community has been quite receptive to the notion of mathematically generated movement; not only has our work been warmly received at Dance/Technology conferences (e.g., =-=[13]-=-), but dancers have even adopted moves created by our algorithms. The music community, in contrast, was initially quite resistant to the notion of applying chaos to respected composers like Bach[19]. ... |

2 |
An Introduction to Benesh Movement Notation
- Causley
- 1968
(Show Context)
Citation Context ...emantics (e.g., names of specific steps, such as pliés). These notational systems are harder for humans to learn because they are more expressive, but they are much more widely useful. Benesh notation=-=[17]-=-, which is currently utilized particularly by ballet dancers, graphs positions on a staff similar to musical notation using symbols to indicate posture, body part by body part. See Figure 2(b) for an ... |

2 |
The Dance Notation
- Labanotation
- 1970
(Show Context)
Citation Context ...on. Spatial coordinates and facings are added below the staff, and expressive qualities like force or strength are indicated via symbols common in musical notation, like fff (fortissimo.) Labanotation=-=[36]-=-, widely used in a variety of movement disciplines and shown in Figure 2(c), also indicates time via a linear staff. Unlike Benesh and music notation, the staff is vertical, moving from bottom to top.... |

1 |
ou L’art de décrire la dance, par caracteres, figures et signes desmonstratifs, avec lesquels on apprend facilement de soy même toutes sortes de dances ... Chez l’auteur...et chez Micel Brunet
- Chorégraphie
(Show Context)
Citation Context ...ions used in the dance community: (a) the Feuillet notation for Baroque-era court dance (b) Benesh notation (c) labanotation. These images are reproduced, with permission where copyright exists, from =-=[26]-=-, [1], and [2], respectively. interpolation to smooth or “tween” between the keyframes, which can be spaced coarsely and unevenly in time when generated by human animators. The computer science commun... |