Computer graphics processing and selective visual display system – Computer graphics processing – Animation
Reexamination Certificate
1997-10-09
2001-06-12
Luu, Matthew (Department: 2672)
Computer graphics processing and selective visual display system
Computer graphics processing
Animation
C345S473000
Reexamination Certificate
active
06246420
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to a method for generating animated figures in such fields as computer graphics and computerized animation, and relates particularly to a method for generating connections between a plurality of time series motion datum provided for animating an articulated rigid figure, and to an apparatus employment said motion data connecting method.
DESCRIPTION OF THE PRIOR ART
Expectations of realism in computer graphics and computer-generated animation have made it necessary to create naturally moving figures of humans and other animated subjects that are modelled using articulated rigid figures. To achieve this, human animators have manually generated three-dimensional time series [coordinate] data to express particular movements using key frame interpolation, or using motion capturing, a technique for capturing data representing actual movements by measuring the movements of a real subject performing a particular action.
Motion data is generated in motion units minimized to the smallest amount of information needed to express a particular movement. This is done both to reduce the required data processing time, and to objectify the motion data for reuse. Because these minimized motion data units are used, a plurality of fine motions described by the minimized motion data must be connected to construct a continuous movement that lasts an extended period of time.
The most primitive method of connecting this motion data is for animators to manually generate the connection data defining each link between adjacent motion data segments by estimating the connection between segments based on specialist knowledge. The manual nature of this task, however, means productivity is low. Each animator also estimates motions in three-dimensional space based on individual experience, often resulting in motion data connections causing the animated figure to move in an unnatural manner. To resolve these problems, the following two methods have been proposed for generating connection data controlling the motions of animated figures.
The first method proposes the automatic generation of connection data for cyclical locomotion (see Munetoshi Unuma, et al., “Fourier Principles for Emotion-based Human Figure Animation,” SIGGRAPH95 Proceedings, pp. 91-96, 1995). The method of Unuma, et al., obtains a Fourier expansion of the time series data of before and after motions, extrapolates the motion in the connection period in the frequency domain, and then obtains the inverse Fourier expansion of the extrapolated motion. In short, this method generates motion connection data premised on predictions of periodic movements.
The second method proposes the generation of motion transitions not necessarily premised on periodic movements (see Charles Rose, et al., “Efficient Generation of Motion Transitions using Spacetime Constraints,” SIGGRAPH96 Proceedings, pp. 147-154, 1996). This method generates motion transitions from space-time constraints of the connected before and after motions by estimating motion in a transition space using inverse kinematics, inverse dynamics, optimization.
The problem with the first method above is that it can only be applied to periodic motion, and its range of application is therefore limited. In addition, its dependence on Fourier expansions and inverse Fourier expansions imposes a significant burden on the calculation engine, and requires much processing time.
The latter method above also basically requires much processing time because the motion transition data is generated by extrapolating transitions from the before and after motions. Use of various numeric calculations also requires a long calculation time.
Connecting motions predicted from before and after motions is a problem of appropriate definition. As a result, an appropriate motion is produced only when the objective function of the optimization calculation matches the connection conditions at that point in time. Motion connection parameters, however, are determined individually according to the motions being connected (the before and after motions, e.g., changing from a walk to a run) and various conditions subjectively describing the motion (e.g., whether the walk is at a leisurely or brisk pace), and motion connection data resulting in unnatural movement is often generated because there are no motion connection parameters that are applicable under generalized conditions.
The described extrapolation method considers only torque consumption as an objective function, and applies an optimization algorithm to minimize torque (energy) consumption. Torque minimization, however, is a necessary condition and alone is not a sufficient condition, and therefore cannot be used to achieve smooth motion connections under all conditions. Torque minimization is not sufficient, for example, when a smooth movement or high speed movement is needed. The result of this is that the connected movements frequently appear unnatural despite the relatively long processing time required.
In both of the above described methods, therefore, connection motions are obtained by estimating the motion from the motions before and after the connection, and if the estimation is off, the resulting connection between motions appears unnatural. To compensate for this estimation problem, the first method above must assume that the before and after motions are periodic motions. The second method, however, describes no compensatory means, and simply frequently estimates poorly. The root of the problem leading to unnatural movements is, therefore, that in both methods described above motion connections are generated by some form of estimation.
The object of the present invention is therefore to provide a motion data connecting method and apparatus therefor for generating motion connection data linking motions before and after the connection so that the connection appears natural, the data can be generated in a short period of time, and it is not necessary to impose particular conditions on the motion, including an assumption of periodic movement.
Other objects and attainments together with a fuller understanding of the invention will become apparent and appreciated by referring to the following description and claims taken in conjunction with the accompanying drawings.
SUMMARY OF THE INVENTION
The object of the present invention is therefore to provide an imaging-device which solves these problems.
In order to achieve the aforementioned objective, a motion data connecting method for connecting, in an overlapping period, a first time series motion data indicative of movement of a subject having at least one joint, and a second time series motion data following and overlapping first time series motion data for a particular time; said motion data connecting method comprising the steps of: calculating a first overlapping period motion data of the first time series motion data, and a second overlapping period motion data of the second time series motion data, in the overlapping period, and one of the following: calculating a position of said subject at a current time based on the first and second overlapping period motion data, calculating an attitude angle of said subject at a current time based on the first and second overlapping period motion data, and calculating joint angle of a joint of subject at a current time based on the first and second overlapping period motion data.
REFERENCES:
patent: 5926190 (1999-07-01), Turkowski et al.
Munetoshi Unuma et al., “Fourier Principles for Emotion-based Human Figure Animation”, Computer Graphics Proceedings, Annual Conference Series (SIGGRAPH 95) pp. 91-96.
Charles Rose et al., “Efficient Generation of Motion Transisitions Using Spacetime Constraints”, Computer Graphics Proceedings, Annual Conference Series (SIGGRAPH 96) pp. 147-154.
A. Witkin et al., “Motion Warping”, Computer Graphics Proceedings, Los Angeles, Aug. 6-11, 1995, pp. 105-108.
R. Boulic et al., “Goal-Oriented Design and Corretion of Articulated Figure Motion with the Track System”, Computers and Gr
Mochizuki Yoshiyuki
Naka Toshiya
Good-Johnson Motilewa
Greenblum & Bernstein P.L.C.
Luu Matthew
Matsushita Electric - Industrial Co., Ltd.
LandOfFree
Movement data connecting method and apparatus therefor does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Movement data connecting method and apparatus therefor, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Movement data connecting method and apparatus therefor will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2462593