WO2022143178A1

WO2022143178A1 - Motion retargeting method and apparatus, electronic device, and storage medium

Info

Publication number: WO2022143178A1
Application number: PCT/CN2021/138407
Authority: WO
Inventors: 朱曈晖; 柴金祥
Original assignee: 魔珐(上海)信息科技有限公司; 上海墨舞科技有限公司
Priority date: 2020-12-31
Filing date: 2021-12-15
Publication date: 2022-07-07
Also published as: CN112634419A; CN112634419B

Abstract

The present disclosure relates to a motion retargeting method and apparatus, an electronic device, and a storage medium. The method comprises: acquiring a first bone posture and a constraint point of a source bone, the constraint point comprising a skin point; constructing a first constraint condition according to the first bone posture and the constraint point of the source bone; obtaining a second constraint condition according to a correspondence between the constraint point of the source bone and a constraint point of a target bone, and the first constraint condition; and obtaining, according to a first position and a second position, a second bone posture of the target bone satisfying the second constraint condition, the first position being a constraint point position of the target bone obtained by mapping the first bone posture to the target bone when there is no second constraint condition, and the second position being a constraint point position of the target bone determined according to the correspondence, the constraint point of the source bone, and the first bone posture. In the present disclosure, skin constraints are added to motion retargeting, so as to avoid phenomena such as motion jitter, stiffness, and unnaturalness, and improve animation production efficiency.

Description

Motion redirection method, device, electronic device and storage medium

This application claims the priority of the Chinese patent application filed on December 31, 2020 with the application number 202011631904.6 and the invention titled "A Motion Redirection Method, Device, Electronic Device and Storage Medium", the entire contents of which are Incorporated herein by reference.

technical field

The present disclosure relates to the field of computer technologies, and in particular, to a motion redirection method, apparatus, electronic device, and storage medium.

Background technique

In the process of 3D content production, animation is very important, but in the process of animation production, the traditional way is to manually produce animation by animators, but this method brings certain problems: hand-made animation is very It is time-consuming and requires a large number of animators to cooperate; high-quality animation requires a high level of animators, and the level of animators with uneven levels will cause uneven levels of animation effects; hand-made animations often only It is difficult to restore the movement details when making general movement actions.

In order to solve the above problems in the manual production method, motion capture is usually used in the current animation production. Motion capture can accurately restore the movements of the actors themselves. What is needed is character animation, and the characters may be of various types, such as real people, or two-dimensional characters, or even animals. Therefore, it is necessary to apply redirection technology to convert the actions of the actors into the corresponding character actions.

In the related art, the redirection technology is relatively simple. Usually, only the bone rotation data of the actor's captured action is migrated. As a result, the redirected action will appear jittery or the overall action will be stiff and unnatural, etc. phenomenon, the action semantics has changed or the relevant motion constraint information is not satisfied, which leads to a lot of time for manual repair in the later stage.

SUMMARY OF THE INVENTION

In view of this, the present disclosure proposes a motion redirection method, apparatus, electronic device, and storage medium.

According to an aspect of the present disclosure, there is provided a motion redirection method, comprising:

obtaining the first bone pose of the source bone and the constraint point of the source bone, the constraint point including the skin point;

constructing a first constraint condition according to the first bone pose and the constraint point of the source bone;

Obtain the second constraint condition according to the correspondence between the constraint point of the source bone and the constraint point of the target bone and the first constraint condition;

According to the first position and the second position, the second bone pose of the target bone that satisfies the second constraint condition is obtained, wherein the first position is to map the first bone pose to the The target bone, the obtained position of the constraint point of the target bone; the second position is based on the correspondence between the constraint point of the source bone and the constraint point of the target bone, the constraint point of the source bone and the posture of the first bone , the determined position of the constraint point of the target bone.

In a possible implementation manner, the method further includes: building one or more virtual bones between the source bone and the target bone;

obtaining the third bone pose of the target bone according to the first bone pose and the virtual bone;

According to the third position and the second position, a second bone pose satisfying the second constraint condition is obtained; wherein the third position is the position of the constraint point of the target bone under the third bone pose.

In a possible implementation manner, the second constraint condition includes: the position of the constraint point of the target bone under the second bone pose is the same as the second position.

In a possible implementation manner, the position of the constraint point of the target bone under the second bone posture is based on the weight information of the constraint point of the target bone, the constraint point of the target bone in the second bone posture is in the target bone The local coordinates on and the transformation matrix from the local coordinate system of the target bone to the world coordinate system are determined.

In a possible implementation manner, the method further includes: processing the second bone pose sequence of the target bone, the motion sequence of the constraint point of the source bone, and the constraint point of the target bone of the current frame by using the prediction model, to obtain The constraint point of the target bone in the next frame; wherein, the motion sequence of the constraint point of the source bone is determined by the positions of different constraint points of the source bone.

In a possible implementation manner, the constraint points further include: mesh points and/or skeleton points.

In a possible implementation manner, the building one or more virtual bones between the source bone and the target bone includes:

According to the source bone, the target bone and one or more preset reference poses, obtain one or more reference poses of the source bone and one or more reference poses of the target bone;

According to the preset corresponding relationship between the source bone and the target bone, one or more virtual bones are built between the same reference pose of the source bone and the target bone.

In a possible implementation manner, the reference posture includes at least: a T-pose.

In a possible implementation manner, the constraint point of the source bone is determined according to the self constraint of the source bone and/or the environment constraint associated with the source bone.

According to another aspect of the present disclosure, there is provided a motion redirection device, comprising:

an acquisition module for acquiring the first bone pose of the source bone and constraint points of the source bone, the constraint points include skin points; a first constraint condition building module for acquiring the first bone pose and the constraints of the source bone point to construct the first constraint condition; the second constraint condition building module is used to obtain the second constraint condition according to the corresponding relationship between the constraint point of the source bone and the constraint point of the target bone and the first constraint condition; motion A redirection module, configured to obtain a second bone pose of the target bone that satisfies the second constraint condition according to the first position and the second position, wherein the first position is to convert the target bone without the second constraint condition The first bone pose is mapped to the target bone, and the obtained position of the constraint point of the target bone; the second position is based on the corresponding relationship between the constraint point of the source bone and the constraint point of the target bone, the constraint point of the source bone and The first bone pose is the determined position of the constraint point of the target bone.

According to another aspect of the present disclosure, there is provided an electronic device comprising: a processor; a memory for storing instructions executable by the processor; wherein the processor is configured to perform the above method.

According to another aspect of the present disclosure, there is provided a non-volatile computer-readable storage medium having computer program instructions stored thereon, wherein the computer program instructions, when executed by a processor, implement the above-described method.

In this embodiment of the present disclosure, the first bone pose of the source bone and the constraint point of the source bone are obtained, where the constraint point includes a skin point; the first constraint condition is constructed according to the first bone pose and the constraint point of the source bone ; According to the corresponding relationship between the constraint point of the source bone and the constraint point of the target bone, and the first constraint condition, the second constraint condition is obtained; According to the first position and the second position, the second constraint condition is obtained. The second bone pose of the target bone, wherein the first position is the constraint point position of the target bone obtained by mapping the first bone pose to the target bone when there is no second constraint condition; The second position is the position of the constraint point of the target bone determined according to the corresponding relationship between the constraint point of the source bone and the constraint point of the target bone, the constraint point of the source bone and the posture of the first bone. In this way, the skin information is added for constraints in the process of motion redirection, which can effectively avoid the possible jitter of motion redirection and the stiffness and unnaturalness of the overall action. Improved animation production efficiency.

Other features and aspects of the present disclosure will become apparent from the following detailed description of exemplary embodiments with reference to the accompanying drawings.

Description of drawings

The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate exemplary embodiments, features, and aspects of the disclosure, and together with the description, serve to explain the principles of the disclosure.

1 shows a flowchart of a motion redirection method according to an embodiment of the present disclosure;

2 shows a flowchart of a motion redirection method according to an embodiment of the present disclosure;

3 shows a schematic diagram of constructing a virtual skeleton according to an embodiment of the present disclosure;

4 shows a structural diagram of a motion redirection device according to an embodiment of the present disclosure;

FIG. 5 shows a block diagram of an electronic device according to an embodiment of the present disclosure.

Detailed ways

Various exemplary embodiments, features and aspects of the present disclosure will be described in detail below with reference to the accompanying drawings. The same reference numbers in the figures denote elements that have the same or similar function. While various aspects of the embodiments are shown in the drawings, the drawings are not necessarily drawn to scale unless otherwise indicated.

The word "exemplary" is used exclusively herein to mean "serving as an example, embodiment, or illustration." Any embodiment described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments.

In addition, in order to better illustrate the present disclosure, numerous specific details are given in the following detailed description. It will be understood by those skilled in the art that the present disclosure may be practiced without certain specific details. In some instances, methods, means, components and circuits well known to those skilled in the art have not been described in detail so as not to obscure the subject matter of the present disclosure.

FIG. 1 shows a flowchart of a motion redirection method according to an embodiment of the present disclosure. As shown in Figure 1, the method may include:

Step 101: Obtain the first bone pose of the source bone and the constraint points of the source bone; wherein, the constraint points include skin points.

In the process of motion redirection, it is necessary to transfer the animation data of the source bone to the animation data of the target bone, where the source bone refers to the bone that already has animation data. The bones drawn by the animator, etc.; the target bones refer to the bones to which the animation data is to be passed, for example, the bones of the virtual characters in the game company; the source bones and the target bones include one or more bones. Source bones and target bones can be in the form of 3D files, such as FBX files (AutoDesk common format), asf format (CMU standard format), mskel (Xmov standard format).

The first bone pose of the source bone can be the action pose of the captured actor, or the action pose of the bone drawn by the animator, etc.; the first bone pose can be determined by the rotation angle ( Euler angles, quaternions, etc.), translation distance and other data are represented, where the animation data of the source bones can be standard animation format files, such as amc (CMU standard format), FBX or sps (Xmov standard format) .

In the actual animation production process, since each bone will have a name, and the naming is not standardized, film and television companies generally name the root bone in the source bone as root, and game companies often name the root bone in the target bone as Bip001. Therefore, in order to In the process of motion redirection, the animation data of the source bone is transferred to the desired target bone, and a semantic association configuration can be made between the source bone and the target bone in advance, so as to correspond to different naming conventions, for example, the source bone The root bone in the target bone is called root, and the root bone in the target bone is called Bip001. Then, if the root corresponds to Bip001, the naming standards of two different companies are corresponding to the root bone. Similarly, the semantics are the same (for example, the bone The function of the same) is associated with other bones, so as to define the corresponding relationship between the source bone and the target bone. At the same time, due to the inconsistency in the number of bones contained in the source bone and the target bone, at this time, based on the principle of the same or similar semantics, the same number of key bones can be selected from the source bone and the target bone, and these key bones can be selected. The bones are mapped to define the corresponding relationship between the source bone and the target bone.

The constraint points may include the constraint points of the target bone and the constraint points of the source bone, which represent the points that constrain the posture of the target bone during the motion redirection process. Exemplarily, the constraint points may include: skin points, bone points, meshes at least one of the points. Among them, the skin point refers to the point attached to the skin surface of the target bone or the source bone, the bone point refers to the point on the target bone or the source bone, and the mesh point refers to the point on other objects in the space where the target bone or the source bone is located. . The data of the skinning points can come from the human database of the motion capture system or the database of other sources; the skinning points of the source bones are associated with each root bone in the source bones, and the skinning points of the target bones are related to each Root bones are associated.

Further, a semantic association configuration may be made in advance between the constraint points of the source bone and the constraint points of the target bone, so as to define the corresponding relationship between the constraint points of the source bone and the constraint points of the target bone. For example, take the bone point of the thumb fingertip in the source bone as the constraint point of the source bone, take the bone point of the thumb fingertip in the target bone as the constraint point of the target bone, and associate the two bone points to define the thumb in the source bone. The correspondence between the bone points of the fingertip and the bone points of the thumb fingertip in the target bone. For another example, the skin point of the toe in the source bone is used as the constraint point of the source bone, the skin point of the toe in the target bone is used as the constraint point of the target bone, and the two skin points are associated to define the toe in the source bone. The correspondence between the skin point of , and the skin point of the toe in the target bone.

In a possible implementation manner, the constraint point of the source bone may be determined according to the self constraint of the source bone and/or the environment constraint associated with the source bone. Among them, the self-constraint of the source skeleton refers to the constraint from the source skeleton itself, and the environment constraint refers to the constraint from the external environment generated by the interaction between the source skeleton and the external environment.

In the process of motion retargeting, since the body proportions, body shapes, and bone topology and joint degrees of freedom of the virtual character and the actor are generally different, the three-dimensional motion of the actor cannot be directly mapped to the virtual character. The various semantic information of the tensor constrains the mapping process; according to whether the semantic information originates from the role itself or from the role and the external environment, it can be divided into self constraints and environmental constraints. For example, in the process of motion redirection, a hand-length People can touch their toes without bending their knees, while people with short hands need to bend their knees when they touch their toes. This is equivalent to two gestures obtained under the same semantic information. When people touch their toes with their hands, they are a kind of self. Constraints, at this time, the skin points and/or bone points on the hands, knees, and toes of the source bone can be taken as the constraint points of the source bone; for another example, a tall person can touch it without jumping The basket, and short people need to jump to touch it. This is equivalent to two gestures obtained under the same semantic information. When a person touches the basket, it is an environmental constraint. At this time, the hand of the source bone can be obtained. , skin and/or bone points on feet, etc., and mesh points on the basket as constraint points for the source bone.

Considering that the number of skin points, bone points, and mesh points in the source and target bones is usually large, a part of all skin points, bone points, and mesh points in the source and target bones can be selected as the Key information points, and then select appropriate points from the key information points of the source bone as the constraint points of the source bone according to their own constraints or environmental constraints. Constraint points; among them, key information points may include key skin points, key bone points, and key mesh points. Exemplarily, according to the needs of the animation scene, a number of skin points can be selected from all the skin points of the source bone as the key skin points of the source bone, and at the same time, selected from all the skin points of the target bone and The skin point corresponding to the position of the key skin point of the source bone is used as the key skin point of the target bone. For example, the skin point at the end of the hand, foot, head, knee and other parts can be selected as the key skin point skin point. The number of key skinning points of the source bone selected above is the same as the number of key skinning points of the selected target bone. The correspondence between the key skin points of the source bone and the key skin points of the target bone.

In the animation production process, according to the different first bone poses of the source bone, the key skinning points of the above-mentioned one or more source bones can be used as the constraint points of the source bone. At the same time, according to the key skinning points of the source bone and the target Correspondence between the key skinning points of the bones, correspondingly, the key skinning points of one or more target bones can be used as the constraint points of the target bone; since each constraint point of the source bone and each constraint point of the target bone There is a corresponding relationship, and through the corresponding relationship between the constraint points and different constraint points determined according to the source bone's own constraints and/or environmental constraints associated with the source bone, different effects can be produced in different animations. Exemplarily, when the first bone pose of the source bone is the action pose of hand touching the knee, the key skin points of the hand of the source bone and the key skin points of the knee can be used as the constraint points of the source bone, and correspondingly, the The key skin points of the hand of the target bone and the key skin points of the knee are used as the constraint points of the target bone.

Step 102 , construct a first constraint condition according to the posture of the first bone and the constraint point of the source bone.

In this step, in the first bone pose, a first constraint condition is constructed according to the constraint points of the source bone selected in the above step 101; the first constraint condition represents the relative positions of different constraint points in the source bone under the first bone pose The position constraints for each constraint point obtained by the relationship.

In the embodiment of the present disclosure, the first constraint condition including skin information can be constructed under the condition that the above self-constraint, the above-mentioned environmental constraint exists, or both the above-mentioned self-constraint and the environmental constraint exist.

Exemplarily, when the first bone pose of the source bone is the action pose of hand touching the knee, it can be known from step 101 that at this time, the key skin points of the hand of the source bone and the key skin points of the knee can be used as the key skin points of the source bone. Constraint point, in the gesture of hand touching the knee, the first constraint condition constructed is that the position of the key skin point of the hand of the source bone is the same as the position of the key skin point of the knee of the source bone.

Step 103: Obtain a second constraint condition according to the correspondence between the constraint point of the source bone and the constraint point of the target bone and the first constraint condition.

In this step, on the basis of obtaining the first constraint condition in the above step 102, using the pre-defined correspondence between each constraint point of the source bone and each constraint point of the target bone, obtain the position generation of each constraint point of the target bone. The second constraint of the constraint.

In the embodiment of the present disclosure, the second constraint condition containing the skin information can be constructed under the condition that the above-mentioned self-constraint, the above-mentioned environment constraint exists, or both the above-mentioned self-constraint and the environment-constraint exist.

Exemplarily, when the key skin points of the hand of the source bone and the key skin points of the knee are used as the constraint points of the source bone, correspondingly, the key skin points of the hand and the knee of the target bone are is the constraint point of the target bone; when the first constraint condition is that the position of the key skin point of the hand of the source bone is the same as the position of the key skin point of the knee of the source bone, according to the constraint point of the source bone and the constraint of the target bone The correspondence between the points can be seen, at this time, in the redirection process, the position of the key skin point of the target bone's hand and the position of the key skin point of the target bone's knee should be constrained as follows: the hand of the target bone The position of the key skin point of the target bone is the same as that of the knee of the target bone, that is, the second constraint condition is that the position of the key skin point of the target bone's hand is the same as that of the knee of the target bone. .

Step 104: According to the first position and the second position, obtain the second bone pose of the target bone that satisfies the second constraint condition; wherein, the first position is the position of the first bone when there is no second constraint condition. The pose is mapped to the target bone, and the obtained position of the constraint point of the target bone; the second position is based on the correspondence between the constraint point of the source bone and the constraint point of the target bone, the constraint point of the source bone, and the first A bone pose, the determined position of the constraint point of the target bone.

In this step, when there is no second constraint condition, the first bone posture is mapped to the target bone, and the position of the constraint point under the bone posture of the target bone at this time is the first position; the first position and the target bone constraint point There may be differences in the target position (ie, the second position) to be reached, and the skeleton pose of the target bone at this time is constrained by the second constraint condition, so as to obtain a second skeleton pose that satisfies the second constraint condition.

Considering that after motion redirection, the final animation needs to be added with skinning effects. In the related art, only using bone data for motion redirection will cause jitter and overall motion stiffness and unnaturalness. For example, a A fat person and a thin person may have exactly the same bones, that is, the source bone is exactly the same as the target bone, but the skinning information of the two is different. At this time, if the fat person directly touches the head If the motion is passed to a skinny person, the final output animation may show that the skinny person is not touching his head; therefore, additional information is needed to assist the skeletal animation to handle the above-mentioned skinning encountered in redirection question. In this step, in the process of transferring the animation data of the source bone to the target bone during the motion redirection, according to the first position of the constraint point of the target bone and the second position of the constraint point of the target bone, and using the above-mentioned second constraint The condition constrains the pose of the second bone of the target bone, so as to obtain the animation data of the target bone. Among them, the constraint point of the target bone includes the skin point of the target bone. In this way, in the process of motion redirection, not only the bone data but also the skin information is used for constraint, which can effectively avoid the possible jitter of motion redirection. and the stiffness and unnaturalness of the overall movement; at the same time, there is no need for a lot of art manpower to repair the data in the later stage, which effectively improves the efficiency of animation production. In addition, through motion redirection, the animation data of the existing source bones can be transferred to completely different target bones, which increases the reuse of data and further improves the efficiency of animation production.

The second bone pose of the target bone is the action pose of the virtual character that satisfies the second constraint condition obtained by reorienting the first bone pose of the source bone, and the second bone pose can be included in the animation data of the target bone. The rotation angle, translation distance and other data are represented. The animation data of the target bone is the animation data of the target bone obtained by the animation data of the source bone through motion redirection, and the style of the animation data of the target bone can be the same as the animation data of the source bone. style is consistent or similar. The animation data of the target bone can be a standard format file of animation, such as amc (CMU standard format), FBX or sps (Xmov standard format).

In a possible implementation manner, the second constraint condition may include: the position of the constraint point of the target bone under the second bone pose is the same as the second position.

Wherein, the position of the constraint point of the target bone in the second bone posture can be based on the weight information of the constraint point of the target bone, the local coordinates of the constraint point of the target bone on the target bone in the second bone posture, and the local coordinate system of the target bone to The transformation matrix of the world coordinate system is determined. Wherein, when the constraint point of the target bone includes the key skin point of the target bone, the position of the key skin point in the second bone pose can be determined, and the following formula (1) can be used to determine the source bone or a certain point of the target bone. The position of a constraint point under a certain bone pose, for example, the following formula (1) can be used to determine the position of a certain constraint point of the target bone under the second bone pose;

In the formula, S(q) represents the position of the constraint point (key skin point) of the target bone when the second bone pose of the target bone is q, and n represents the constraint point (key skin point) contained in the target bone and the target bone. The number of bones associated with the skin point), h is the serial number of the associated bone; W _h represents the weight of the constraint point (key skin point) of the target bone controlled by the bone h; L _h represents the pose of the second bone When q is the local coordinate of the constraint point (key skinning point) of the target bone on the hth bone, M _h represents the transformation matrix from the local coordinate system of the hth bone to the world coordinate system.

Exemplarily, when the first bone pose of the source bone is an action pose of hand touching the knee, the constraint points of the target bone include the key skin points of the hand and the key skin points of the knee of the target bone, and the key skin points of the hand. The key skin points of the point and knee are controlled by multiple bones in the target bone. According to the control weights of these root bones, the key skin points of the hand, and the local coordinates of the key skin points of the knee on each bone and each The transformation matrix from the local coordinate system of the root bone to the world coordinate system can obtain the key skin points of the hand and knee of the target bone when the second bone pose of the target bone is q in the world coordinate system. Location.

The second position of the constraint point of the target bone is the constraint point of the target bone determined according to the preset correspondence between the constraint point of the source bone and the constraint point of the target bone, the constraint point of the source bone, and the posture of the first bone Location. Since there is a correspondence between the constraint point of the target bone and the constraint point of the source bone, in the process of transferring the animation data of the source bone to the target bone, the position of the constraint point of the source bone under the first bone pose is determined by the corresponding relationship. The relationship can be mapped to obtain the second position of the constraint point of the target bone, and the second position is the target position expected to be reached by the constraint point of the target bone in the process of motion redirection.

Exemplarily, when the first bone pose of the source bone is the action pose of hand touching the knee, the constraint points of the target bone include the key skin points of the hand and the key skin points of the knee of the target bone. The position of the key skin point of the hand and the key skin point of the knee, combined with the corresponding relationship between the constraint point of the target bone and the constraint point of the source bone, can know the second position of the key skin point of the target bone, that is The expected target position of the key skin points of the knee of the target bone and the expected target position of the key skin points of the hand.

The constraint function can be determined according to the position and the second position of the constraint point of the target bone in the second bone pose, and the second constraint condition is expressed as the minimum value of the constraint function. It can be understood that the value of the constraint function is The smallest can indicate that the position of the constraint point of the target bone under the second bone pose is the same as the second position; through the constraint condition, the bone posture of the target bone in the process of motion redirection is constrained, so that the second constraint condition is satisfied. The second bone pose of the target bone. Exemplarily, as shown in the following formula (2),

In the formula, S(q) represents the position of the jth constraint point of the target bone when the second bone pose of the target bone is q, where the constraint point includes the key skin point, and m represents the constraint point of the target bone (the key mask). The number of skin points), c _j represents the second position of the jth constraint point (key skin point) of the target bone. According to the formula (2), based on the position of each constraint point of the target bone and the second position of the constraint point of the target bone, the difference between the position of each constraint point of the target bone and the second position of the constraint point of the target bone is as small as possible when the second bone posture is q, and the second bone posture of the final target bone is obtained by optimization. .

In the embodiment of the present disclosure, not only the skeleton data but also the skin information is used effectively to establish the constraint conditions, so as to carry out more perfect constraints on the motion redirection process. In the related art, there is only bone data in the process of motion retargeting. When constraining body parts in human motion, a lot of constraint information is missing. For example, when the hand touches the knee, there will be a constraint relationship between the hand and the knee. This kind of constraint relationship is difficult to reflect in the skeleton data. For example, in the movement of the foot, the skeleton data has only two bones to control the foot, but the skeleton data cannot know the effect of the skinning, and the feet are often uneven. Or the switching of the constraints of the soles of the feet and the toes during walking is unnatural; that is, it is difficult to establish an effective constraint relationship in the process of motion redirection only through bone data, and a satisfactory effect cannot be achieved. In this embodiment, key skin points such as hands and feet can be used as constraint points, and a first constraint condition and a second constraint condition can be constructed, and then the first position and the second position of the constraint point of the target bone can be used to satisfy the The second bone pose of the target bone of the second constraint condition, the positions of key skin points such as hands and feet under the second bone pose of the target bone and the corresponding key skin points under the first bone pose of the source bone The positions of the points correspond to each other, so that the action of the virtual character can be expressed more realistically and naturally through the second skeletal pose.

Further, the constraint problem in motion redirection can also be described as the optimization form shown in the following formula (3):

In the formula, q represents the second bone pose of the target bone, q` represents the bone pose of the target bone obtained without constraints (that is, directly transfer the animation data of the source bone without constraints to obtain the target bone), FK ( q) Represents the coordinate information of the constraint point (bone point) when the second bone pose of the target bone is q, wherein the constraint point includes the bone point, and c _i represents the ith joint point (bone point) of the target bone corresponding to the need to reach The specified position of , k represents the number of joint points (bone points) in the target bone.

Exemplarily, the above formula (2) can be combined with the above formula (3) to obtain a more comprehensive optimization formula, so as to optimize the skeletal posture of the target bone during the motion redirection process, and the optimization formula is shown in the following formula (4) :

In the formula, q represents the second bone pose of the target bone, q` represents the bone pose of the target bone obtained without constraints, and FK(q) represents the constraint point when the second bone pose of the target bone is q (the key Bone point) information, c _i represents the specified position corresponding to the ith joint point (key bone point) of the target bone that needs to be reached, k represents the number of joint points (key bone points) in the target bone; m represents the target bone The number of constraint points (key _skinning points) of the The second position of the jth constraint point (key skin point) of the target bone, w ₁ , w ₂ , and w ₃ are all weight coefficients.

Among them, FK(q) can express the coordinate information of the constraint point (key bone point) of the target bone, S(q) can express the constraint point (key skin point) information of the target bone, and c _i and c _j respectively represent the pair of bone information The position constraints and the position constraints of the skin information, and the specific values of the weight coefficients w ₁ , w ₂ , and w ₃ can be adjusted according to actual needs, so that in the process of motion redirection, the valid information is no longer only the rotation of the bones. Angle and other information, the position of the skin point becomes an important additional constraint; at the same time, the use of bone motion information and skin information for constraints can effectively avoid the jitter that may occur in motion redirection and the stiffness and unnaturalness of the overall action; At the same time, the animation production efficiency is improved.

Further, in the process of motion redirection, there are also the following environmental constraints. For example, when walking, the feet and the ground are an environmental constraint. When doing pull-ups, the hands and the horizontal bar are an environmental constraint. Embracing each other between individuals is also an environmental constraint (in this case, the environment refers to the object of the hug). Therefore, in order to achieve a more ideal effect, it is necessary to consider the switching of various constraints generated by the environment, as well as the spatial difference between the real environment where the source bones are located and the virtual environment where the target bones are located.

Illustratively, the motion sequence of the constraint points of the source bone consists of the position of the key skin point of the source bone and the position of the key mesh point, or the position of the key skin point of the source bone and the position of the key bone point, or the position of the key skin point of the source bone and the position of the key bone point. The position of the key skin point is determined with the position of other key skin points. For example, key mesh points may include: the ground on which people walk, tables and chairs that people touch, horizontal bars, etc., and other key skin points may include: key skin points of bones of other characters associated with the source bone skin point.

The prediction model uses the second pose sequence of the target bone, the motion sequence of the constraint point of the source bone, and the position of the constraint point of the target bone in the current frame as input, and uses the constraint point of the target bone in the next frame as the prediction result. Train, and train an algorithmic model that reaches convergence. The second bone pose sequence of the target bone is a sequence composed of the second bone poses of the target bone obtained through motion redirection within a certain time window; the motion sequence of the constraint points of the source bone is the source corresponding to each first bone pose of the source bone. A sequence of different constraint points for a bone. The constraint points of the source bones are related to the positions of key grid points, key bone points or other key skin points in the above motion sequence. The position of the point or other key skin point can determine the constraint point of the source bone. For example, according to the position of the key information point and the key grid point, key bone point, or other key skin point, the key information point is determined as The constraint points of the source bone, thereby determining the motion sequence of the constraint points of the source bone. Exemplarily, a piece of animation data of the source bone includes 50 frames, and the second bone pose sequence of the target bone may include the second bone pose of the target bone obtained after the animation data of the source bone in the first 30 frames of motion redirection, and the constraints of the source bone. The motion sequence of points can include the constraint points of the source bone in the first bone pose of the source bone in 50 frames, through the second bone pose of the target bone in the 30 frames, the constraint points of the source bone in the 50 frames, and the target bone in the 30th frame. to determine the constraint point of the target bone in the current 31st frame of data; for example, as shown in the following formula (5):

In the formula, F is the trained neural network prediction model, c` represents the constraint point of the target bone in the current frame, c ^* represents the constraint point of the target bone in the previous frame,

represents the second bone pose sequence of the target bone,

A motion sequence representing the constraint points of the source bone.

In the embodiment of the present disclosure, on the basis of constructing the constraints by using the skin information of the target bone itself, other key information points in the environment (that is, the key grid points of objects in the space where the source bones are located, the The key skinning points of other bones, where the key mesh points of the object in the space where the source bone is located have a corresponding relationship with the key mesh points of the object in the space where the target bone is located, and the key skinning points of other bones in the space where the source bone is located are the same as The key skin points of other bones in the space where the target bone is located have a corresponding relationship), and the distance judgment in the redirection space is added. The distance judgment is: look at the source and target bones in the space and space objects (objects with interaction) As a whole, the distance between them will become the basis for constraint determination, that is, not only the target bone of a single virtual character is considered, but the redirection space where the entire target bone is located is regarded as a whole. For example, for the entire motion redirection space multiple characters and objects in multiple environments, mark these characters and objects in the redirection space in advance (that is, mark key information points), and update the constraints of the target bones in real time as a whole during the motion redirection process In this way, the influence of environmental factors is fully considered, and the constraint point of the target bone and the second constraint determined by the constraint point change in real time with changes in the environment and motion Updated to dynamically constrain the transfer of animation data from the source bone to the target bone.

Exemplarily, a section of the source bone that runs first, then grabs the horizontal bar to do pull-ups and then falls is transferred to the target bone. In this action, the running is the constraint between the footsteps and the ground. During the upward movement of the body, the hand and the horizontal bar are constrained, and finally, when the body falls, it returns to the constraint between the footsteps and the ground. During this entire movement, the ground and the horizontal bar are used as environmental factors, and the ground is used in the process of motion redirection. , the horizontal bar and the source bone (target bone) as a whole, the key information points include the key skin points of the hands and feet, and the key mesh points on the ground and the horizontal bar that are in contact with the key skin points on the source bone , the constraint points of the source bones are related to these key information points; according to this whole, that is, the key mesh points on the ground, the key mesh points of the horizontal bar, the key skin points of the hands, and the key skin points of the feet, Therefore, the motion sequence of the constraint points of the source bones can be known, that is, the key skin points of the feet first contact the key mesh points on the ground, then the key skin points of the hand contact the key mesh points of the horizontal bar, and finally return. The key skin points to the feet are in contact with the key grid points on the ground; assuming that a part of the motion redirection has been completed for the entire action, the second bone pose sequence of the target bone is obtained, according to the constraint point of the source bone. The motion sequence, the second skeleton pose sequence of the target skeleton, and the constraint point of the target skeleton of the current frame, the constraint point of the target skeleton of the next frame can be obtained through the above formula (5); After that, you can refer to the above formula (2) or formula (4), and use the real-time update second constraint condition of the constraint point of the target bone, the second constraint condition includes skin information and environment information, through the second constraint condition The second bone pose of the target bone is constrained in real time during the motion redirection process, so that the position of the constraint point of the target bone under the optimized second bone pose of the target bone is the same as the second position.

In this way, in the process of motion redirection, the valid information is no longer only the motion information such as the rotation angle of the bones, and the key skin points and key grid points become an important additional constraint; at the same time, the bone motion information and skin information are used to constrain , effectively avoid the jitter that may occur in motion redirection and the stiffness and unnaturalness of the overall action; at the same time, it improves the efficiency of animation production.

FIG. 2 shows a flowchart of a motion redirection method according to an embodiment of the present disclosure. As shown in Figure 2, the method may include:

Step 200, between the source bone and the target bone, build one or more virtual bones;

In the process of motion redirection, there are generally three essential differences between the source bone and the target bone, namely: the difference in the number of bones, the difference in the axial direction, and the difference in length. Among them, the difference in the number of bones makes it impossible to transfer the spatial rotation data of the joints one-to-one between the source bone and the target bone. The axis of the bone refers to the local coordinate system of each bone. The difference in the axial direction causes the joint between the source bone and the target bone to directly transmit spatial rotation data. The difference in length leads to errors in transferring animation data from the source bone to the target bone (eg foot sliding, etc.).

In this step, one or more virtual bones are built by comparing the source bone and the target bone, the source bone and the target bone are bridged through the virtual bone, and the difference between the source bone and the target bone is transmitted through the virtual bone, thereby eliminating the need for Axial difference between two kinds of bones, exemplarily, FIG. 3 shows a schematic diagram of constructing a virtual bone according to an embodiment of the present disclosure. As shown in FIG. 3, the length of each bone in the virtual bone is related to the corresponding source bone The length of each bone in the virtual bone is the same, and the axial direction of each bone in the virtual bone is the same as the axial direction of each bone in the corresponding target bone. At the same time, the number of virtual bones can be multiple, and multiple virtual bones are used to bridge the source bone and the target bone, so as to achieve a smooth transition from the source bone to the target bone, and further optimize the effect of eliminating axial differences. In addition, the number of virtual bones may be different from the source bones and the target bones, and the number of virtual bones may be the same as the number of bones selected in the above-mentioned predefined correspondence between the source bones and the target bones, for example, the source bones and the target bones Each of the bones has 100 bones. In the above association configuration, 10 bones are selected as key bones for correspondence, and 10 virtual bones are calculated through 100 root bones and 100 target bones and their corresponding relationships; for example, the source bones have 100 The root bone and the target bone have 10 bones. In the above semantic association, 10 bones are selected as key bones for correspondence, and 10 virtual bones are calculated through 100 root bones and 10 target bones and their corresponding relationships. Therefore, the above-mentioned influence of the difference in the number of bones on the motion redirection is effectively solved, and the operation efficiency is improved.

In a possible implementation manner, this step may include: obtaining one or more reference poses of the source skeleton and one of the target skeletons according to the source skeleton, the target skeleton and one or more preset reference poses or multiple reference poses; according to the preset correspondence between the source bones and the target bones, one or more virtual bones are built between the same reference poses of the source bones and the target bones.

In the related art, only one reference pose is used to define the style association between the source bone and the target bone, which makes it difficult to control the style of the character action. For example, in the state of T-pose, the styles of the source bone and the target bone are the same, However, when bending over or squatting, it is difficult to guarantee the style similarity between the source bone and the target bone. Therefore, in the embodiment of the present disclosure, in order to meet the requirements for different animation styles in the actual animation production process, the virtual bone is constructed by adding Multiple reference poses are used, and the multi-frame reference pose is used to define the style relationship between the source bone and the target bone, so as to control the final animation style of the character. For example, through one or more preset reference poses, and combining the source bone, the target bone, and the above-mentioned correspondence between the source bone and the target bone, one or more virtual bones corresponding to the reference poses are obtained, so as to use the virtual bones to The style transitions from the source bone to the target bone. For example, if the reference pose is standing, the source bone is a normal standing style, but the target bone is expected to be in a hunchbacked standing style, the corresponding relationship between the source bone and the target bone is used to make the normal standing source A virtual skeleton of the standing posture is built between the skeleton and the target bone of hunchbacked standing. The style of the virtual skeleton's standing posture is the transition from the normal standing style to the hunchbacked standing style, so that the virtual bone is used to transition the normal standing style of the source bone to the hunchbacked standing style. The hunchback standing style of the target skeleton.

In a possible implementation manner, the reference pose includes: a T-pose. Among the preset one or more reference poses, the basic reference pose is a T-pose (T-pose), that is, a standing posture with both hands raised. For this T-pose, according to the corresponding relationship between the source bone and the target bone, and the T-pose presented by the source bone and the T-pose presented by the target bone. Between the source bone and the target bone, a T-pose virtual bone is built, and the T-pose of the virtual bone bridges the source bone and the target bone. The stylistic form when standing between hands raised.

In a possible implementation manner, the reference posture further includes at least one posture of squatting, taking a step, akimbo, bending over, and raising both hands.

The above-mentioned T-pose virtual bone is constructed by using the source bone and target bone, as well as the corresponding relationship between the source bone and the target bone. The T-pose virtual bone is used as the basic virtual bone, which basically bridges the axial difference between the source bone and the target bone. . In order to convert the different styles of the source bone and the target bone through the virtual bone, in this embodiment of the present disclosure, on the basis of the T-pose virtual bone, other reference poses are added to meet different style requirements. Exemplarily, the reference postures may also include other 5 different reference postures: squatting posture, stepping posture, akimbo posture, bending posture, and raising hands posture. For any one of the five reference poses, a virtual skeleton of the reference pose is built, so as to use the virtual skeleton to transition the style of the source skeleton under the reference pose to the style of the target skeleton under the reference pose. It should be noted that the preset reference pose has a T- pose, and other reference poses can be added according to different animation production requirements, and the number of reference poses is not limited in this embodiment of the present disclosure.

Step 201: Acquire the first bone pose of the source bone and the constraint points of the source bone, wherein the constraint points include skin points.

This step is the same as step 101 in the above-mentioned FIG. 1 , and will not be repeated here.

Step 202 , constructing a first constraint condition according to the posture of the first bone and the constraint points of the source bone.

This step is the same as step 102 in the above-mentioned FIG. 1 , and will not be repeated here.

Step 203: Obtain a second constraint condition according to the correspondence between the constraint point of the source bone and the constraint point of the target bone and the first constraint condition.

This step is the same as step 103 in the above-mentioned FIG. 1 , and will not be repeated here.

Step 204: Obtain a third bone pose of the target bone according to the first bone pose and the virtual bone.

In this step, on the basis of the first bone pose of the source bone, one or more virtual bones are obtained, and then one or more virtual bones are fused to obtain the third bone pose of the target bone, so that the style of the source bone can be combined Mapped to the target bone; the pose of the third bone is the pose of the target bone obtained without constraints.

Exemplarily, according to the first bone pose and one or more virtual bones constructed above and corresponding to the first bone pose, the third bone pose of the target bone is obtained through fusion calculation.

Step 205: According to the third position and the second position, obtain a second bone pose that satisfies the second constraint condition; wherein, the third position is the constraint point of the target bone under the third bone pose. position; the second position is the position of the constraint point of the target bone determined according to the correspondence between the constraint point of the source bone and the constraint point of the target bone, the position of the constraint point of the source bone, and the posture of the first bone.

In this step, the position of the constraint point of the target bone under the posture of the third bone, that is, the third position, may be different from the target position that the constraint point of the target bone needs to reach (that is, the second position). On the basis of obtaining the third bone pose of the target bone from the virtual bone, a second constraint is added to constrain the bone pose of the target bone, so as to obtain a more vivid and realistic second bone pose of the target bone, making the motion redirection action more accurate.

In this step, the animation data of the source bone is transferred to the target bone through one or more virtual bones constructed above, that is, the first bone pose of the source bone is mapped to the third bone pose of the target bone through the virtual bone; By bridging the source bone and the target bone, the virtual bone eliminates the quantitative and axial differences between the source bone and the target bone, thus ensuring that the animation data can be effectively transferred from the source bone to the target bone; The reference pose of the source skeleton and the target skeleton's different pose styles are also converted through the virtual skeleton. At the same time, the third bone pose of the target bone obtained during the motion redirection process is constrained by the second constraint condition in FIG. 1, so as to obtain the second bone pose of the target bone. In this way, the skin information is effectively used to carry out Constraints can more effectively avoid the jitter that may occur in motion redirection and the stiffness and unnaturalness of the overall movement, which further improves the efficiency of animation production and avoids a lot of manual work.

It should be noted that, although the motion redirection method is described above by taking the above embodiment as an example, those skilled in the art can understand that the present disclosure should not be limited thereto. In fact, the user can flexibly set each implementation manner according to personal preferences and/or actual application scenarios, as long as it conforms to the technical solutions of the present disclosure.

In this way, in this embodiment of the present disclosure, by acquiring the first bone pose of the source bone and the constraint points of the source bone, the constraint points include skin points; according to the first bone pose and the constraint points of the source bone, a first Constraint condition; according to the corresponding relationship between the constraint point of the source bone and the constraint point of the target bone, and the first constraint condition, the second constraint condition is obtained; according to the first position and the second position, the second constraint condition is obtained. The second bone pose of the target bone with the constraint condition, wherein the first position is the position of the constraint point of the target bone obtained by mapping the first bone pose to the target bone without the second constraint condition; The second position is the position of the constraint point of the target bone determined according to the corresponding relationship between the constraint point of the source bone and the constraint point of the target bone, the constraint point of the source bone and the posture of the first bone. In this way, the skin information is added for constraints in the process of motion redirection, which can effectively avoid the possible jitter of motion redirection and the stiffness and unnaturalness of the overall action. Improved animation production efficiency.

FIG. 4 shows a structural diagram of a motion redirection apparatus according to an embodiment of the present disclosure. As shown in FIG. 4 , the apparatus includes: an acquisition module 301 for acquiring a first bone pose of a source bone and constraints of the source bone point, the constraint points include skin points; the first constraint construction module 302 is used to construct the first constraint according to the first bone pose and the constraint points of the source bone; the second constraint construction module 303 is used to construct the first constraint to obtain the second constraint condition according to the corresponding relationship between the constraint point of the source bone and the constraint point of the target bone and the first constraint condition; the motion redirection module 304 is used for, according to the first position and the second position, obtaining the second bone pose of the target bone that satisfies the second constraint condition, wherein the first position is the target bone obtained by mapping the first bone pose to the target bone when there is no second constraint condition The second position is the constraint point of the target bone determined according to the corresponding relationship between the constraint point of the source bone and the constraint point of the target bone, the constraint point of the source bone and the posture of the first bone point location.

In a possible implementation manner, the apparatus further includes: a virtual bone building module, configured to: build one or more virtual bones between the source bone and the target bone; according to the first bone posture and the virtual bone to obtain the third bone posture of the target bone; according to the third position and the second position, obtain the second bone posture that satisfies the second constraint condition; wherein, the third position is the target The position of the constraint point of the bone in the third bone pose.

In a possible implementation manner, the apparatus further includes: a constraint point position prediction module, configured to predict the second bone pose sequence of the target bone, the motion sequence of the constraint point of the source bone, and the target of the current frame by using the prediction model The constraint points of the bones are processed to obtain the constraint points of the target bone in the next frame; wherein, the motion sequence of the constraint points of the source bone is determined by the positions of different constraint points of the source bone.

In a possible implementation manner, the virtual skeleton building module is further configured to: obtain one or more references of the source skeleton according to the source skeleton, the target skeleton and one or more preset reference poses The pose and one or more reference poses of the target bone; according to the preset correspondence between the source bone and the target bone, one or more virtual bones are built between the same reference pose of the source bone and the target bone.

It should be noted that, although the motion redirection device is described above by taking the above embodiment as an example, those skilled in the art can understand that the present disclosure should not be limited thereto. In fact, the user can flexibly set each implementation manner according to personal preferences and/or actual application scenarios, as long as it conforms to the technical solutions of the present disclosure.

FIG. 5 shows a block diagram of an electronic device 1900 according to an embodiment of the present disclosure. For example, the electronic device 1900 may be provided as a server. 5, electronic device 1900 includes processing component 1922, which further includes one or more processors, and a memory resource represented by memory 1932 for storing instructions executable by processing component 1922, such as applications. An application program stored in memory 1932 may include one or more modules, each corresponding to a set of instructions. Additionally, the processing component 1922 is configured to execute instructions to perform the above-described methods.

The electronic device 1900 may also include a power supply assembly 1926 configured to perform power management of the electronic device 1900, a wired or wireless network interface 1950 configured to connect the electronic device 1900 to a network, and an input output (I/O) interface 1958 . Electronic device 1900 may operate based on an operating system stored in memory 1932, such as Windows Server™, Mac OS X™, Unix™, Linux™, FreeBSD™ or the like.

In an exemplary embodiment, a non-volatile computer-readable storage medium is also provided, such as memory 1932 comprising computer program instructions executable by processing component 1922 of electronic device 1900 to perform the above-described method.

The present disclosure may be a system, method and/or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of the present disclosure.

A computer-readable storage medium may be a tangible device that can hold and store instructions for use by the instruction execution device. The computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples (non-exhaustive list) of computer readable storage media include: portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM) or flash memory), static random access memory (SRAM), portable compact disk read only memory (CD-ROM), digital versatile disk (DVD), memory sticks, floppy disks, mechanically coded devices, such as printers with instructions stored thereon Hole cards or raised structures in grooves, and any suitable combination of the above. Computer-readable storage media, as used herein, are not to be construed as transient signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (eg, light pulses through fiber optic cables), or through electrical wires transmitted electrical signals.

The computer readable program instructions described herein may be downloaded to various computing/processing devices from a computer readable storage medium, or to an external computer or external storage device over a network such as the Internet, a local area network, a wide area network, and/or a wireless network. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and/or edge servers. A network adapter card or network interface in each computing/processing device receives computer-readable program instructions from a network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in each computing/processing device .

Computer program instructions for carrying out operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or instructions in one or more programming languages. Source or object code, written in any combination, including object-oriented programming languages, such as Smalltalk, C++, etc., and conventional procedural programming languages, such as the "C" language or similar programming languages. The computer readable program instructions may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer, or entirely on the remote computer or server implement. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (eg, using an Internet service provider through the Internet connect). In some embodiments, custom electronic circuits, such as programmable logic circuits, field programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), can be personalized by utilizing state information of computer readable program instructions. Computer readable program instructions are executed to implement various aspects of the present disclosure.

Aspects of the present disclosure are described herein with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the disclosure. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer readable program instructions.

These computer readable program instructions may be provided to a processor of a general purpose computer, special purpose computer or other programmable data processing apparatus to produce a machine that causes the instructions when executed by the processor of the computer or other programmable data processing apparatus , resulting in means for implementing the functions/acts specified in one or more blocks of the flowchart and/or block diagrams. These computer readable program instructions can also be stored in a computer readable storage medium, these instructions cause a computer, programmable data processing apparatus and/or other equipment to operate in a specific manner, so that the computer readable medium on which the instructions are stored includes An article of manufacture comprising instructions for implementing various aspects of the functions/acts specified in one or more blocks of the flowchart and/or block diagrams.

Computer readable program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other equipment to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other equipment to produce a computer-implemented process , thereby causing instructions executing on a computer, other programmable data processing apparatus, or other device to implement the functions/acts specified in one or more blocks of the flowcharts and/or block diagrams.

The flowchart and block diagrams in the Figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of instructions, which comprises one or more functions for implementing the specified logical function(s) executable instructions. In some alternative implementations, the functions noted in the blocks may occur out of the order noted in the figures. For example, two blocks in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It is also noted that each block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or flowchart illustrations, can be implemented in dedicated hardware-based systems that perform the specified functions or actions , or can be implemented in a combination of dedicated hardware and computer instructions.

Various embodiments of the present disclosure have been described above, and the foregoing descriptions are exemplary, not exhaustive, and not limiting of the disclosed embodiments. Numerous modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terminology used herein was chosen to best explain the principles of the embodiments, the practical application or technical improvement in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

A motion redirection method, comprising:

obtaining the first bone pose of the source bone and the constraint point of the source bone, the constraint point including the skin point;

constructing a first constraint condition according to the first bone pose and the constraint point of the source bone;

Obtain the second constraint condition according to the correspondence between the constraint point of the source bone and the constraint point of the target bone and the first constraint condition;

According to the first position and the second position, the second bone pose of the target bone that satisfies the second constraint condition is obtained, wherein the first position is to map the first bone pose to the The target bone, the obtained position of the constraint point of the target bone; the second position is based on the correspondence between the constraint point of the source bone and the constraint point of the target bone, the constraint point of the source bone and the posture of the first bone , the determined position of the constraint point of the target bone.
The method according to claim 1, wherein the method further comprises: building one or more virtual bones between the source bone and the target bone;

obtaining the third bone pose of the target bone according to the first bone pose and the virtual bone;

According to the third position and the second position, a second bone pose satisfying the second constraint condition is obtained; wherein the third position is the position of the constraint point of the target bone under the third bone pose.
The method according to claim 1, wherein the second constraint condition comprises:

The position of the constraint point of the target bone in the second bone pose is the same as the second position.
The method according to claim 3, wherein the position of the constraint point of the target bone in the second bone pose is based on the weight information of the constraint point of the target bone and the constraint of the target bone in the second bone pose The local coordinates of the point on the target bone and the transformation matrix from the local coordinate system of the target bone to the world coordinate system are determined.
The method according to claim 1, wherein the method further comprises: using the prediction model to the second bone pose sequence of the target bone, the motion sequence of the constraint point of the source bone, and the constraint point of the target bone of the current frame Perform processing to obtain the constraint points of the target bone in the next frame; wherein, the motion sequence of the constraint points of the source bone is determined by the positions of different constraint points of the source bone.
The method according to claim 1, wherein the constraint points further comprise: grid points and/or bone points.
The method according to claim 2, wherein the building one or more virtual bones between the source bone and the target bone comprises:

According to the source bone, the target bone and one or more preset reference poses, obtain one or more reference poses of the source bone and one or more reference poses of the target bone;

According to the preset corresponding relationship between the source bone and the target bone, one or more virtual bones are built between the same reference pose of the source bone and the target bone.
The method according to claim 6, wherein the reference posture at least comprises: a T-pose.
The method according to claim 1, wherein the constraint point of the source bone is determined according to the self constraint of the source bone and/or the environment constraint associated with the source bone.
A motion redirection device, comprising:

an acquisition module, configured to acquire the first bone pose of the source bone and the constraint point of the source bone, the constraint point including the skin point;

a first constraint condition building module for constructing a first constraint condition according to the first skeleton pose and the constraint point of the source skeleton;

The second constraint condition building module is configured to obtain the second constraint condition according to the corresponding relationship between the constraint point of the source bone and the constraint point of the target bone and the first constraint condition;

The motion redirection module is used to obtain the second bone pose of the target bone that satisfies the second constraint condition according to the first position and the second position, wherein the first position is the position where the target bone will be placed without the second constraint condition. The first bone pose is mapped to the target bone, and the obtained position of the constraint point of the target bone; the second position is based on the corresponding relationship between the constraint point of the source bone and the constraint point of the target bone, the constraint point of the source and the first bone pose, and the determined position of the constraint point of the target bone.
An electronic device, comprising:

processor;

memory for storing processor-executable instructions;

Wherein, the processor is configured to implement the method of any one of claims 1 to 9 when executing executable instructions stored in the memory.
A non-volatile computer-readable storage medium on which computer program instructions are stored, characterized in that, when the computer program instructions are executed by a processor, the method described in any one of claims 1 to 9 is implemented.