JP2001025464A - Method for analyzing individual facial expression deformation process - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method for simply and automatically analyzing a facial expression deformation process of the face of an optional model with a computer on an individual basis. SOLUTION: By constituting a sample of a finite element model by dividing a representative structure of facial muscles and skin into finite elements, automatically measuring and incorporating shape measurements of a face of an individual person as typical point data by an artificial visual device, and giving the coordinates of a plurality of representative points 31 to 43 thus obtained to the sample of the finite element model, an individual finite element model is generated. By giving time series data of the contraction percentage β of muscles at the time of facial expression deformation to a generated finite element model, and by applying these to a computer system equipped with the function of analyzing achievement of the muscles and skin combining the function of analyzing finite elements based on an active and passive incremental constitutive equation of the facial muscles and the function of analyzing finite elements of a superelastic finite deformed body for computation facial expression deformation is reproduced by an FEM model.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for analyzing facial expression deformation and facial stress processes for each individual.

[0002]

2. Description of the Related Art Human facial expressions appear on the face as contractions of facial muscles due to contraction of facial muscles. Facial muscles are about 3
There are around zero, and the selective combination of muscle contractions transforms the underlying skin-subcutaneous tissue system to form various expressions of emotions. "Minari Lee et al., Proceedings of the 10th Bioengineering Conference of the Japan Society of Mechanical Engineers,
98-1, p160-161 ", a finite element analysis of a hyperelastic finite deformation body of the facial muscle-Hiff soft tissue system, which simulates the facial expression deformation of the Hiff accompanying the contraction of the facial muscles by a computer according to the finite element method System HYEL
-A MUSCL system is described.

[0003]

SUMMARY OF THE INVENTION It is an object of the present invention to provide a method for easily and automatically analyzing a facial expression deformation process of an arbitrary model individually for each individual using the above-mentioned system. .

[0004]

According to the present invention, a representative structure of a finite element model is constructed by dividing a representative structure of facial muscles and hifu into finite elements, and the shape of an individual human face is measured. The coordinates of the plurality of representative points obtained by the above are given to the model of the finite element model to generate an individual face finite element model, and the generated individual finite element model The contraction rate of the muscle from the change in coordinates is determined, and these are applied to a muscle-Hiff coupled finite element analysis system combining the incremental stiffness equation of the muscle and the incremental stiffness equation of the Hiff to transform the facial expression. There is provided an analysis method of an individual facial expression deformation process including a step of calculating displacement, strain and stress of a muscle or a hifu at each point of the face in the process.

[0005]

FIG. 1 shows various muscles around the eye. In FIG. 1, 10 is the orbital muscle, 12 is the frontal muscle,
14 is the wrinkle muscle, 15 is the root muscle, 16 is the upper lip levator ridge, 1
8 is the upper lip muscle, 20 is the levator ani muscle, 22 is the small zygomatic muscle, 24
Is the large cheekbone muscle. Around the eyes, as shown in FIG.
Various muscles are present so as to surround the eyes, and selective contraction of these muscles enables formation of complex facial expressions including opening and closing of the eyes.

A description will now be given of a program for automatically generating a finite element model of the eye marginal system, which is used in the above-described muscle-Hiff coupled hyperelastic finite deformation finite element analysis program HYEL-MUSCL. Until now, for the full-mouth model and the limbus system, a finite element model of the orbicularis 筋 target model and the orbicularis full model have been generated, and facial expression analysis has been performed using these models. With this model, it was difficult to analyze complicated facial expression deformation around the eyes and mouth, and it was not possible to analyze facial expressions for individuals. In addition, in the process of opening and closing the eye, not only the orbicularis muscle but also various radiation muscles surrounding the orbicularis muscle (eight types including frontal muscles) are greatly affected. Various radiation sources must also be considered. Therefore, in one embodiment of the present invention, an FEM automatic mesh generation system for the limbus facial expression muscle-Hiff, in which various radiation muscles are attached to the finite element model of the orbicularis muscle full model, is constructed. The outline of the program is as follows.

(A) Two-dimensional 8-node quadrilateral isoparametric elements are used for muscles and hifu. (B) The origin is taken at the center of the eye, the x-axis is taken in the major axis direction of the eye, the y-axis is taken in the minor axis direction, and the z-axis is taken in the thickness direction of the eye. (C) Combining the three-dimensional coordinate data (FIG. 2) of individual facial representative points captured by the artificial vision device AV-DAVIS with the model mesh data for the eye length, the attachment positions of various muscles, and the calculation area. Generates mesh data of muscles and hifu for each individual.

(D) The number of element divisions of streaks and hifu can be freely set. (E) It is possible to freely set the number of divisions of a line / hiff and the size of an element near the outer or outer corner of the eye by input data. (F) In this model, in the region where the orbicularis muscle, the frontalis muscle, the creased eyebrows muscle, and the rhinopsalis muscle exist, the muscle and the hifu constitute independent elements, but share a node.

(G) In a region where a plurality of muscle bundles overlap each other, they constitute independent elements but share a node. (H) The direction of the muscle bundle is determined for each muscle element. (I) The FEM automatic mesh generation system is aimed at analyzing facial expressions for each individual. Therefore, the size of the eye, the position of the muscle attachment, and the like that vary depending on the individual are determined by the artificial vision device A
Individual input data is created from V-DAVIS, US ultrasonic measurement, etc., and individual FEM (finite element method) mesh data can be created based on the input data.

For example, referring to FIG.
The simulation area is determined by the coordinates of 4, 42 and 43. The right eye position is determined based on the coordinates of the representative points 35, 36, 37, and 38. The attachment position of the rhinopsalis muscle 15 (FIG. 1) is determined based on the coordinates of the representative points 32 and 33. Representative point 3
The attachment position of the wrinkle line 14 is determined by the coordinates of 9, 40, and 41.

(J) Regarding the orbicularis muscle, each of the first to fourth quadrants as viewed from the center of the eye opening is approximated by an ellipse from the representative point input data, and has a node common to the hyph, and is divided into elements. I do. (K) The root of the nose and the line of the wrinkles are approximately attached to the Hiff element based on the representative point input data, and have a node common to the Hiff element. (L) For the frontal muscle, the position of attachment to the outermost element of the orbicularis muscle is determined from the template mesh data, and is approximately attached to the Hiff element to have a common node with the Hiff element.

(M) For the large zygomatic muscle, the small zygomatic muscle, the levator ani muscle, the upper lip muscle, and the upper lip pterygoid muscle, the position of attachment to the outermost element of the orbicularis muscle and the direction of the muscle bundle are determined from the template mesh data. And have a common node with the outermost element of the orbicularis muscle. FIG. 3 shows an individual FE automatically generated by the above program.
An example of M data is shown. Also, FIG. 4 shows the frontal muscle 12, wrinkle muscle 14, nasal root muscle 15, upper lip pterygoid muscle 16, orbital levator muscle 20, upper lip muscle 18 and small cheekbone muscle 2 in the FEM model.
2 and the area of the large zygomatic muscle 24 are indicated by thick lines.

Next, an example of the incremental stiffness equation used in the above-described hyperelastic finite deformation finite element program HYEL-MUSCL of the muscle-Hiff coupling system will be described (for details, see Norio Tsuta et al., Hiroshima University). Research Report, 4
6 , p59-71). The incremental stiffness equation of the muscle applied to the finite element of the muscle portion is given by, for example, the following equation (1).

Further, an incremental stiffness equation applied to the finite element of the part of the stiff part is given by the following equation (3), for example. Equations (1) and (3) are obtained for the muscles and the hiff of the face target part, and these are coupled according to the actual connection state. The displacement is determined. When the results are used in Equations (2) -1 and (2) -2, the incremental strain of the muscle, the incremental active stress, and the incremental passive stress are obtained, and the incremental displacement of the Hiff is given to Equation (2) -1. Using the equation (4), the incremental strain and the incremental stress of the Hiff can be obtained.

(Equation 1)

(Equation 2)

The incremental stiffness equation for the part of the hiff is given by the following equation (3).

[0016]

(Equation 3)

The finite element analysis of the coupled muscle-Hiff system is expressed by the following equation (1).
By solving the incremental finite element equations of the whole system of the muscle and the Hiff given by Equation (3) and the equation, the incremental displacement of each node of the muscle and the Hiff can be obtained, and the muscle and the Hiff can be obtained from this. The incremental strain and incremental stress of are calculated.
In addition, as for the entire rigidity matrix of the analysis model, when the components of the muscle element and the skin element overlap, an element rigidity matrix relating to a common node is assembled. The physical properties of muscles are described in “Takeshi Iwamoto et al., Proceedings of Computational Engineering Sciences, 3-1 .
The values shown in Fig. 5 obtained by the method of "p259-262" can be used. As the physical property values of Hiff, "Tomohiro Fujimura et al., Proceedings of Computational Engineering Lectures, 3-1" , p287-29
The value shown in FIG. 6 obtained by the method of “0” can be used. Β in the equation (2) is a muscle activity rate or a contraction rate, and has a certain relationship with the integrated myoelectric potential (EMG) (Iwamoto) Tsuyoshi et al., Proceedings of the Computer Engineering Lectures, 3-1, p259-26
2). Therefore, the integral EMG of each muscle when an expression such as “happy” or “angry” was created was measured, and the integral EM measured separately was measured.
By converting the contraction rate into a contraction rate based on the relationship between G and the contraction rate, time-series data of the contraction rate β of each muscle when a certain expression is created can be obtained. By inputting this into the HYEL-MUSCL system, changes in facial expressions such as "happy" and "angry" can be reproduced by the FEM model. Alternatively, the above representative point data can be used instead.

FIG. 7 shows the calculation results when β values are given to the orbicularis oculi muscle and the large zygomatic muscle to reproduce the facial expression change of “pleasure” shown in FIG. FIG. 9 shows a calculation result when β values are given to the root of the nose and the line of the wrinkles of the eyebrows in order to reproduce the facial expression change of “anger” shown in FIG.

[0019]

As described above, according to the present invention, the facial expression deformation process of an arbitrary model is subjected to complicated procedures of the finite element method by inputting representative point data taken in from an artificial vision device. It becomes possible to analyze the expression of each individual almost automatically without any.

[Brief description of the drawings]

FIG. 1 is a diagram showing various muscles around an eye.

FIG. 2 is a diagram illustrating representative points.

FIG. 3 is a diagram showing an example of an individual FEM model.

FIG. 4 is a diagram showing regions of each muscle in the FEM model by thick lines.

FIG. 5 is a diagram showing an example of a physical property value of a muscle.

FIG. 6 is a diagram illustrating an example of a physical property value of a hifu.

FIG. 7 is a diagram showing a calculation result of an FEM model that reproduces the expression of “pleasure”;

FIG. 8 is a photograph showing an example of an expression of “pleasure”.

FIG. 9 is a diagram illustrating a calculation result of an FEM model that reproduces an expression of “anger”.

FIG. 10 is a photograph showing an example of an expression of “anger”.

[Explanation of symbols]

Reference Signs List 10: orbicularis muscle 12 ... frontal muscle 14 ... wrinkle muscle 15 ... nose ridge muscle 16 ... upper lip levator ani muscle 18 ... upper lip levator muscle 20 ... levator ani muscle 22 ... small cheekbone muscle 24 ... large zygomatic muscle 31, 34, 42, 43 ... Representative points that determine the simulation area 35,36,37,38 ... Representative points that determine the right eye position 32,33 ... Representative points that determine the attachment position of the nose ridge muscle 39,40,41 ... Representative points that determine the attachment position of the wrinkle muscle

 ──────────────────────────────────────────────────続 き Continuing from the front page (72) Inventor Tomohiro Fujimura 1-4-1 Kagamiyama, Higashihiroshima-shi, Hiroshima Hiroshima University Faculty of Engineering (72) Inventor Nobuyuki Tozaki 1-4-1-1 Kagamiyama, Higashihiroshima-shi, Hiroshima Hiroshima Within the Faculty of Engineering (72) Inventor Osamu Kaneko 3-9-1 Nishi Gotanda, Shinagawa-ku, Tokyo Shiseido Beauty Science Research Institute

Claims (1)

[Claims] 1. A representative structure of facial muscles and hifu is divided into finite elements to form a model of a finite element model, and coordinates of a plurality of representative points obtained by measuring the shape of each human face And the change in their coordinate values during the transformation process is given to the model of the finite element model to generate a finite element model of the face for each individual, and the instantaneous line at the time of the facial expression deformation from the time history data of the representative points. By providing these to the generated individual finite element model after giving the contraction rate of the physician, the function with finite element analysis based on the active / passive incremental constitutive equation of the facial muscles, An individual who has a step of calculating the deformation, strain, and stress in the facial expression deformation process by applying this to a computer system having a muscle and Hiff achievement analysis function combined with the finite element analysis function of the hyperelastic finite deformation body Analysis method of the muscle and skin of facial expression variations and expression stress process.