Three-dimensional clothing model binding method, device, electronic device and storage medium

Abstract

The present invention provides a three-dimensional clothing model binding method, device, electronic device and storage medium, which relates to the technical field of three-dimensional modeling. The method comprises: obtaining a first discrete curve set corresponding to a human body model in an initial posture, a second discrete curve set corresponding to a human body model in a current posture and a third discrete curve set corresponding to a clothing model in an initial posture; performing affine transformation on the third discrete curve set according to the first discrete curve set and the second discrete curve set to obtain a fourth discrete curve set; optimizing each topological line of the clothing model in the current posture according to the coordinate system change information of each skeletal node coordinate system in the human body model in the current posture and the fourth discrete curve set; updating the mesh model of the clothing model in the initial posture according to each optimized topological line and the coordinate system change information to generate a clothing form corresponding to the clothing model in the current posture. The present invention improves the binding efficiency of the three-dimensional clothing model while improving the binding effect.

Description

Technical Field

[0001] The present invention relates to the technical field of three-dimensional modeling, and in particular to a three-dimensional clothing model binding method, device, electronic equipment and storage medium. Background Art

[0002] 3D clothing model binding is widely used in virtual fitting, movie special effects, game development and other fields.

[0003] Traditional 3D clothing model binding requires the binding artist to first create a human body model with complete shaping, then wrap the human body model with the clothing model to make the clothing model deform with the movement of the human body model, and finally convert this deformation into the clothing's shaping BS (Blend Shape). However, when the deformation of the human body model is too large or the structure of the clothing model is complex, it is difficult for the shape of the clothing model to keep the same deformation as the human body model, and the binding artist needs to spend a lot of time to repair it, which affects the efficiency and effect of 3D clothing model binding. Summary of the invention

[0004] The present invention provides a 3D clothing model binding method, device, electronic device and storage medium, which are used to solve the defects of low efficiency and poor effect of 3D clothing model binding in the prior art, and improve the efficiency and effect of 3D clothing model binding.

[0005] The present invention provides a three-dimensional clothing model binding method, comprising:

[0006] Acquire a first discrete curve set, a second discrete curve set and a third discrete curve set; the first discrete curve set is extracted from a human body model in an initial posture, the second discrete curve set is extracted from a human body model in a current posture, and the third discrete curve set is extracted from a clothing model in the initial posture;

[0007] According to the first discrete curve set and the second discrete curve set, performing affine transformation on the third discrete curve set to obtain a fourth discrete curve set;

[0008] Optimizing each topological line of the clothing model in the current posture according to the coordinate system change information of each skeletal node coordinate system in the human body model in the current posture and the fourth discrete curve set;

[0009] According to each optimized topological line and the coordinate system change information, the mesh model of the clothing model in the initial posture is updated to generate a clothing form corresponding to the clothing model in the current posture.

[0010] According to a three-dimensional clothing model binding method provided by the present invention, the step of obtaining each topological line of the clothing model in the current posture includes:

[0011] According to each grid coordinate system in the human body model of the current posture, each topological line of the clothing model of the current posture is obtained by affine transformation.

[0012] According to a three-dimensional clothing model binding method provided by the present invention, the topological lines of the clothing model in the current posture are optimized according to the coordinate system change information of each skeletal node coordinate system in the human body model in the current posture and the fourth discrete curve set, including:

[0013] Performing curve optimization on each discrete curve in the fourth discrete curve set, and optimizing each topological line according to each optimized discrete curve; and,

[0014] According to the associated constraints between the topological lines, the coordinate system change information of the coordinate system of each skeletal node is transmitted to each topological line through skinning information interpolation, and each topological line is optimized.

[0015] According to a three-dimensional clothing model binding method provided by the present invention, the curve optimization of each discrete curve in the fourth discrete curve set includes:

[0016] In the fourth discrete curve set, a first curve to be optimized associated with a first skeletal joint point, a second skeletal joint point, and a third skeletal joint point in the human body model of the current posture is obtained;

[0017] Optimizing the distance scaling coefficient corresponding to the first curve to be optimized according to the distance values ​​of each target point pair between the first connection curve between the first skeletal joint point and the second skeletal joint point and the second connection curve between the second skeletal joint point and the third skeletal joint point;

[0018] Performing curve optimization on the first curve to be optimized according to the optimized distance scaling coefficient;

[0019] Among them, the first skeletal joint point and the second skeletal joint point are two adjacent skeletal joint points, the second skeletal joint point and the third skeletal joint point are two adjacent skeletal joint points, and the second skeletal joint point is a skeletal joint point in an action state; the target point pair includes a first target point in the first connecting curve and a second target point in the second connecting curve, and the first target point and the second target point are on the same horizontal plane.

[0020] According to a three-dimensional clothing model binding method provided by the present invention, the curve optimization of each discrete curve in the fourth discrete curve set includes:

[0021] In the fourth discrete curve set, a second curve to be optimized associated with a fourth skeletal joint point and a fifth skeletal joint point in the human body model of the current posture is obtained;

[0022] Optimizing the distance of each line segment in the second curve to be optimized according to the distance value between each third target point in the third connecting curve between the fourth skeletal joint point and the fifth skeletal joint point and the target vertex of the second curve to be optimized;

[0023] Among them, the target vertex is the vertex with the shortest mapping distance to the fifth skeletal joint point; the fourth skeletal joint point and the fifth skeletal joint point are two adjacent skeletal joint points, and there is at least one skeletal joint point in an action state.

[0024] According to a three-dimensional clothing model binding method provided by the present invention, the curve optimization of each discrete curve in the fourth discrete curve set includes:

[0025] According to the average coordinate information of adjacent vertices of the vertices of each discrete curve in the fourth discrete curve set, Laplace smoothing optimization is performed on each discrete curve in the fourth discrete curve set.

[0026] According to a three-dimensional clothing model binding method provided by the present invention, the mesh model of the clothing model in the initial posture is updated according to each optimized topological line and the coordinate system change information to generate a clothing form corresponding to the clothing model in the current posture, including:

[0027] According to the coordinate system change information, obtaining change information of each grid corresponding to the clothing model of the current posture;

[0028] Optimizing each of the grids according to the change information of each of the grids, the associated constraints between each of the grids, and the gradient optimization to obtain each optimized grid;

[0029] The mesh model is updated according to each of the optimized meshes and each of the optimized topological lines to generate the clothing form.

[0030] According to a three-dimensional clothing model binding method provided by the present invention, the first discrete curve set is obtained according to the following steps:

[0031] From the human body model in the initial posture, discrete curves perpendicular to the axial direction of the coordinate system of each bone node in the human body model in the initial posture, and / or discrete curves parallel to the axial direction of the coordinate system of each bone node in the human body model in the initial posture are extracted to construct the first discrete curve set.

[0032] According to a three-dimensional clothing model binding method provided by the present invention, the first discrete curve set is obtained according to the following steps:

[0033] From the human body model in the initial posture, extract the discrete curve in the tangential direction of the axial direction of the target bone node coordinate system in the human body model in the initial posture, the discrete curve in the target topological line in the human body model in the initial posture, and at least one of the parallel discrete curves evenly distributed in other areas except the local area where the target bone node coordinate system is located, and construct the first discrete curve set.

[0034] According to a three-dimensional clothing model binding method provided by the present invention, after updating the mesh model of the clothing model in the initial posture according to each optimized topological line and the coordinate system change information and generating the clothing form corresponding to the clothing model in the current posture, the method further includes:

[0035] Determining at least one target deformation algorithm from a plurality of deformation algorithms according to a morphology optimization requirement corresponding to the clothing model of the current posture;

[0036] According to the at least one target deformation algorithm, the garment form is deformed and optimized;

[0037] Among them, the multiple deformation algorithms include multiple ones of deformation gradient algorithm, Laplace gradient, curvature gradient and conformal gradient.

[0038] According to a three-dimensional clothing model binding method provided by the present invention, performing affine transformation on the third discrete curve set according to the first discrete curve set and the second discrete curve set to obtain a fourth discrete curve set includes:

[0039] For each discrete curve in the third discrete curve set, the following steps are performed:

[0040] In the first discrete curve set, a first associated curve of the current discrete curve is obtained; the projection distance from the vertex of the current discrete curve to the first associated curve is the shortest;

[0041] acquiring, from the second discrete curve set, a second association curve associated with the first association curve;

[0042] According to a first direction vector of a tangent direction of a first associated point in the first associated curve, a second direction vector of a tangent direction of a second associated point in the second associated curve, an angle between the first direction vector and the second direction vector, a vertex in the current discrete curve, the first associated point, the second associated point, and a distance scaling coefficient corresponding to the current discrete curve, an affine transformation curve corresponding to the current discrete curve is obtained; the first associated point is a point that matches a vertex of the current discrete curve, and the second associated point is a point that matches the first associated point;

[0043] The fourth discrete curve set is constructed according to the affine transformation curve corresponding to each discrete curve in the third discrete curve set.

[0044] The present invention also provides a three-dimensional clothing model binding device, comprising:

[0045] An acquisition unit, configured to acquire a first discrete curve set, a second discrete curve set, and a third discrete curve set; the first discrete curve set is extracted from a human body model in an initial posture, the second discrete curve set is extracted from a human body model in a current posture, and the third discrete curve set is extracted from a clothing model in the initial posture;

[0046] a transformation unit, configured to perform an affine transformation on the third discrete curve set according to the first discrete curve set and the second discrete curve set, so as to obtain a fourth discrete curve set;

[0047] an optimization unit, configured to optimize each topological line of the clothing model in the current posture according to the coordinate system change information of each skeletal node coordinate system in the human body model in the current posture and the fourth discrete curve set;

[0048] The morphology generation unit is used to update the mesh model of the clothing model in the initial posture according to each optimized topological line and the coordinate system change information, and generate the clothing morphology corresponding to the clothing model in the current posture.

[0049] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, any of the three-dimensional clothing model binding methods described above is implemented.

[0050] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the three-dimensional clothing model binding method as described in any one of the above is implemented.

[0051] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the three-dimensional clothing model binding method as described above is implemented.

[0052] The three-dimensional clothing model binding method, device, electronic device and storage medium provided by the present invention perform affine transformation on a first discrete curve set extracted from a human body model in an initial posture, a second discrete curve set extracted from a human body model in a current posture, and a third discrete curve set extracted from a clothing model in an initial posture, so as to accurately simulate and obtain a fourth discrete curve set corresponding to the clothing model in the current posture, and optimize and generate the clothing form corresponding to the clothing model in the current posture by combining the coordinate system change information of the coordinate system of each skeletal node in the human body model in the current posture and the fourth discrete curve set, so as to guide the clothing form change of the clothing model outside the human body model according to the posture change of the human body model, so that the clothing form corresponding to the clothing model in the current posture correctly keeps synchronous deformation with the human body model, thereby greatly reducing the repair time required by the binder, improving the efficiency of three-dimensional clothing model binding, and improving the binding effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0054] Figure 1 This is one of the flow charts of the three-dimensional clothing model binding method provided by the present invention;

[0055] Figure 2 It is a schematic flow chart of the topology line optimization steps provided by the present invention;

[0056] Figure 3 is a distribution diagram of a discrete curve of a non-exercise clothing model and a discrete curve of a human body model provided by the present invention;

[0057] Figure 4 It is one of the distribution diagrams of the original discrete curve of the clothing model after exercise and the original discrete curve of the human body model provided by the present invention;

[0058] Figure 5 It is one of the distribution schematic diagrams of the optimized discrete curve of the clothing model after exercise and the optimized discrete curve of the human body model provided by the present invention;

[0059] Figure 6 This is the second distribution diagram of the original discrete curve of the clothing model after exercise and the original discrete curve of the human body model provided by the present invention;

[0060] Figure 7This is the second distribution schematic diagram of the optimized discrete curve of the clothing model after exercise and the optimized discrete curve of the human body model provided by the present invention;

[0061] Figure 8 This is the second flow chart of the three-dimensional clothing model binding method provided by the present invention;

[0062] Fig. 9 It is a structural schematic diagram of a three-dimensional clothing model binding device provided by the present invention;

[0063] Fig.10 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. According to the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] Traditional 3D clothing model binding requires the binding artist to first create a human body model with complete shaping, then wrap the human body model with the clothing model to allow the clothing model to deform with the movement of the human body model, and finally convert this deformation into the clothing shaping BS. This method is only applicable to the binding of clothing models with simple structures. For example, a single layer of clothing can be well handled by wrapping, and the outer layer of the clothing completely deforms with the movement of the body. However, for scenes where the deformation of the human body model is too large or the structure of the clothing model is complex, it is difficult for the shape of the clothing model to correctly keep in sync with the human body model. At this time, the binding artist needs to spend a lot of time to repair it, which affects the efficiency and effect of the 3D clothing model binding.

[0066] For example, for the elbows and knees of a loose clothing model, when the deformation of the human body model exceeds 120°, the method of wrapping the human body model with the clothing model cannot accurately keep in sync with the deformation of the human body model; for another example, for a clothing model with shoulders and multiple accessories, when the shoulders of the human body model are at different angles, it is difficult for the clothing model to obtain a more reasonable shape by wrapping the human body model with the clothing model, resulting in jumps in the dynamic change process of the clothing model, and it is difficult to keep in sync with the deformation of the human body model. For another example, for a clothing model containing multiple layers of clothing, the clothing model of the multiple layers of clothing produces differentiated deformations during the deformation process of the human body model, and the method of wrapping the human body model with the clothing model cannot capture these deformations well, resulting in the interlacing of multiple layers of clothing; for another example, for a clothing model containing a collar, the lines of the collar are complex, and if the human body model is wrapped with the clothing model, the deformation of the lines of the collar cannot be captured well. For example, for clothing models with many wrinkles on the surface, such as ancient costume models, when the posture of the human body model changes, the deformation of the clothing model needs more detailed adjustment, especially when the elbows and knees have large angles. If the human body model is wrapped by the clothing model, these deformations cannot be captured well. The dynamic shaping processing by the binding artist takes a lot of time, which affects the efficiency and effect of the 3D clothing model binding.

[0067] In view of the above problems, the present invention provides a three-dimensional clothing model binding method. Figure 1 It is one of the flow charts of the three-dimensional clothing model binding method provided by the present invention, such as Figure 1 As shown, the method includes:

[0068] Step 110, obtaining a first discrete curve set, a second discrete curve set and a third discrete curve set; the first discrete curve set is extracted from a human body model in an initial posture, the second discrete curve set is extracted from a human body model in a current posture, and the third discrete curve set is extracted from a clothing model in the initial posture;

[0069] Here, the clothing model in the initial posture is a clothing model bound to the human body model in the initial posture. The clothing structure of the clothing model can be configured according to actual needs, such as a single-layer clothing model, a clothing model with more wrinkles in appearance, such as ancient costumes, clothing models with multiple layers of clothing stacked on top of each other, a clothing model including arm guards, a clothing model including hand guards, a clothing model including armor, etc. This embodiment does not make specific limitations on this.

[0070] Optionally, before binding the three-dimensional clothing model, it is necessary to extract discrete curves from the human body model in the initial posture to obtain a first discrete curve set; extract discrete curves from the human body model in the current posture to obtain a second discrete curve set; and extract discrete curves from the clothing model in the initial posture to obtain a third discrete curve set.

[0071] Here, the extraction of discrete curves in the first discrete curve set and the second discrete curve can be realized by extracting discrete curves in one or more orientations for all skeletal joint point coordinate systems, such as the discrete curve of the upper arm can be extracted along the positive direction of the X-axis of the shoulder joint coordinate system, and the discrete curve of the forearm can be extracted along the tangential direction of the axial direction of the forearm coordinate system, etc.; it can also be realized by extracting discrete curves in one or more orientations for some key skeletal node coordinate systems, and it can also be realized by extracting discrete curves for target topological lines, which is not specifically limited in this embodiment. The so-called target topological line is a curve pre-configured in the human body model.

[0072] Here, the step of extracting the discrete curves in the third discrete curve set can be to obtain a clothing edge image from the clothing model in the initial posture according to an edge algorithm, so as to extract the clothing geometric edge discrete curves from the clothing edge image, so that after the human body model moves and deforms, the local details of the clothing can be better determined based on the extracted discrete curves of the clothing edges; or, the edge seams in the clothing model in the initial posture can be first obtained, and then the sparse edge seams are interpolated by seam line interpolation to achieve the acquisition of discrete curves.

[0073] Step 120, performing affine transformation on the third discrete curve set according to the first discrete curve set and the second discrete curve set to obtain a fourth discrete curve set;

[0074] Optionally, after obtaining the first discrete curve set and the second discrete curve set before and after the posture change of the human body model and the third discrete curve set before the posture change of the clothing model, the dynamic changes between the discrete curves in the first discrete curve set and the second discrete curve set before and after the posture change of the human body model can be used to drive the dynamic changes of the discrete curves in the third discrete curve set before the posture change of the clothing model, and the discrete curves after the posture change of the clothing model are obtained by affine transformation, thereby constructing a fourth discrete curve set.

[0075] The so-called affine transformation is achieved by mapping at least one of the translation, rotation and scaling changes of the discrete curves in the first discrete curve set and the discrete curves in the second discrete curve set to translate, rotate and scale the third discrete curve set.

[0076] Step 130, optimizing each topological line of the clothing model in the current posture according to the coordinate system change information of each skeletal node coordinate system in the human body model in the current posture and the fourth discrete curve set;

[0077] Here, each skeletal node in the human body model has a local coordinate system pre-built; when each skeletal node undergoes a posture change, that is, after the initial posture is transformed to the current posture, its corresponding local coordinate system also changes accordingly. Therefore, the so-called coordinate system change information is the change information generated by the coordinate system of each skeletal node in the human body model as the posture changes from the initial posture to the current posture.

[0078] Optionally, the topological lines after the movement are obtained first, that is, the topological lines of the clothing model in the current posture.

[0079] In some embodiments, the step of acquiring each topological line of the clothing model in the current posture includes: obtaining each topological line of the clothing model in the current posture by affine transformation according to each grid coordinate system in the human body model in the current posture.

[0080] The so-called mesh may be a triangular face or a quadrilateral face, etc. The following takes the mesh being a triangular face as an example to describe the method provided in the embodiment of the present application.

[0081] Optionally, based on the coordinate system of each triangle surface in the human body model of the current posture, an affine transformation is performed on the topological line to obtain each topological line of the clothing model of the current posture.

[0082] The affine transformation here can be to calculate the difference between the triangular surface coordinate system in the human body model of the current posture and the triangular surface coordinate system in the human body model of the initial posture, so as to obtain the translation, rotation and scaling parameters required for the affine transformation based on the difference, and perform an affine transformation on the topological lines of the clothing model of the initial posture according to the translation, rotation and scaling parameters to obtain the topological lines of the clothing model of the current posture.

[0083] After obtaining the topological lines of the clothing model in the current posture, the topological lines of the clothing model in the current posture are optimized according to the discrete curves in the fourth discrete curve set, as well as the coordinate system transformation matrix between the coordinate system of each bone node in the human body model in the current posture and the coordinate system of each bone node in the human body model in the initial posture; according to the coordinate system transformation matrix, the displacement of each topological line of the clothing model in the current posture is optimized to ensure smoothness during movement, to ensure the continuity and rationality of the topological structure, so that each optimized topological line can effectively follow the change of the coordinate system, so as to subsequently obtain a clothing form that is more matched with the human body model in the current posture, that is, the clothing form corresponding to the clothing model in the current posture formed by the optimized topological lines can be consistent with the deformation of the human body model in the current posture, thereby ensuring that the form of the clothing model can correctly keep in synchronization with the deformation of the human body model, avoiding the binding engineer spending a lot of time on repair, thereby improving the efficiency and effect of three-dimensional clothing model binding.

[0084] It should be noted that the execution order of the optimization steps based on discrete curves and the optimization steps based on coordinate system change information can be adaptively adjusted. For example, the optimization steps based on discrete curves can be executed first, or the optimization steps based on coordinate system change information can be executed first. No specific limitation is made here.

[0085] The so-called optimization based on the coordinate system change information can be to adaptively adjust the displacement of the topological line according to the change information of the coordinate system of the bone node, so that it can better adapt to the posture and movement of the human body model. It should be noted that during the binding process, the arm guards in the clothing model can be treated as soft clothing and treated in the same way, so that the topological lines on the arm guards can also be deformed to fit the human body model. For the topological lines on the hand guards in the clothing model, they will deform along the back of the hand when the hand is turned back, and keep their own shape unchanged in other directions of movement. For the topological lines on the armor in the clothing model, the degree of freedom of change is limited, and they can be configured to deform along with the clavicle joints.

[0086] Step 140 , updating the mesh model of the clothing model in the initial posture according to each optimized topological line and the coordinate system change information, and generating a clothing form corresponding to the clothing model in the current posture.

[0087] Optionally, after obtaining each optimized topological line, the mesh model of the clothing model in the initial pose can be updated according to each optimized topological line and the coordinate system change information; the update referred to here can be to apply the coordinate system change information to each mesh surface corresponding to the clothing model in the current pose to obtain an optimized mesh surface; thereby, the shape of each optimized deformed mesh surface and each optimized topological line are applied to the mesh model of the clothing model in the initial pose, and by updating the vertices or topological structure of the mesh of the mesh model of the clothing model in the initial pose, a clothing shape corresponding to the clothing model in the current pose can be obtained, which can correctly keep synchronous deformation with the human body model and retain the local details of the clothing model as much as possible.

[0088] The method provided in this embodiment performs affine transformation on the first discrete curve set extracted from the human body model in the initial posture, the second discrete curve set extracted from the human body model in the current posture, and the third discrete curve set extracted from the clothing model in the initial posture, so as to accurately simulate and obtain the fourth discrete curve set corresponding to the clothing model in the current posture, and optimizes and generates the clothing form corresponding to the clothing model in the current posture in combination with the coordinate system change information of each bone node coordinate system in the human body model in the current posture and the fourth discrete curve set, so as to guide the clothing form change of the clothing model outside the human body model according to the posture change of the human body model, so that the clothing form corresponding to the clothing model in the current posture correctly keeps synchronous deformation with the human body model, thereby greatly reducing the repair time required by the binder, improving the binding efficiency of the three-dimensional clothing model, and improving the binding effect.

[0089] In some embodiments, virtual digital humans have serious homogeneity of character settings, low quality and efficiency, etc. Therefore, using this application for the generation of virtual digital humans can improve the generation efficiency of virtual digital humans compared to manual modeling by modelers. In addition, the virtual digital humans have high-quality facial modeling. The hyper-realistic virtual humans thus obtained can be used for user interaction, virtual digital human videos, and virtual digital human services, so that users can visually have the same experience as real-life interactions, real-life videos, and real-life services.

[0090] In some embodiments, Figure 2 A schematic diagram of the process of topology line optimization steps provided in the embodiment of the present application; Figure 2 As shown, in step 130, each topological line of the clothing model of the current posture is optimized according to the coordinate system change information of each skeletal node coordinate system in the human body model of the current posture and the fourth discrete curve set, including:

[0091] Step 210, performing curve optimization on each discrete curve in the fourth discrete curve set;

[0092] Step 220, optimizing each of the topological lines according to each optimized discrete curve; and,

[0093] Step 230: According to the associated constraints between the topological lines, the coordinate system change information of the coordinate system of each skeleton node is transmitted to each topological line through skinning information interpolation, so as to optimize each topological line.

[0094] Optionally, after the fourth discrete curve set is obtained through step 120, in order to further improve the clothing model binding quality, before optimizing each topological line, each discrete curve in the fourth discrete curve set needs to be optimized to obtain the correct shape of each discrete curve in the fourth discrete curve set. The curve optimization here includes but is not limited to at least one of scaling distance optimization, line segment distance optimization and smoothing optimization.

[0095] Next, the optimization information of the optimized discrete curve is transmitted to each topological line to optimize each topological line, so that a clothing form that can correctly keep synchronous deformation with the human body model can be constructed according to each optimized topological line.

[0096] In addition, since the topological lines of the clothing model have a corresponding relationship with the entity inside it (i.e., the human body model), for example, the clothes on the arm have a corresponding relationship with the arm inside it. And each bone node, such as after the arm moves, the coordinate system of each bone node will also move. Therefore, each topological line can also be optimized according to the coordinate system transformation matrix of each bone node coordinate system. The specific steps are as follows:

[0097] Since the coordinate systems of each bone node are discontinuous, the curves obtained directly through the mesh surfaces under the coordinate systems of each bone node are curved. Therefore, it is necessary to add constraints to the association conditions between each topological line so that the topological line meets both the position requirements and the smoothness requirements. It should be noted that the association constraints between each topological line are used to limit the changes between lines to be gradual, that is, the curves of the topological lines between adjacent nodes are smooth and the position and direction changes of the topological lines between adjacent nodes are gradual.

[0098] Therefore, for each bone node, the coordinate system transformation matrix is ​​calculated according to its displacement change in the current posture; for each topological line, the displacement change information of the topological line in the current posture in the global coordinate system compared with the position in the initial posture is calculated through skinning information interpolation according to the coordinate system transformation matrix of the bone node to which it belongs and the associated constraints between the topological lines;

[0099] Next, based on the displacement change information of each topological line, each topological line is optimized, thereby obtaining an optimized topological line that satisfies both the change information of each topological line and the associated constraint conditions.

[0100] The method provided in this embodiment optimizes the displacement of each topological line by combining the coordinate system transformation matrix of each skeletal node coordinate system with the associated constraints between each topological line, thereby achieving a more accurate topological line deformation effect, thereby improving the authenticity and realism of the clothing model. At the same time, the continuity and smoothness between the topological lines can also be retained, making the clothing form constructed based on the optimized topological lines more natural and smooth.

[0101] In some embodiments, performing curve optimization on each discrete curve in the fourth discrete curve set in step 210 includes:

[0102] In the fourth discrete curve set, a first curve to be optimized associated with a first skeletal joint point, a second skeletal joint point, and a third skeletal joint point in the human body model of the current posture is obtained;

[0103] Optimizing the distance scaling coefficient corresponding to the first curve to be optimized according to the distance values ​​of each target point pair between the first connection curve between the first skeletal joint point and the second skeletal joint point and the second connection curve between the second skeletal joint point and the third skeletal joint point;

[0104] Performing curve optimization on the first curve to be optimized according to the optimized distance scaling coefficient;

[0105] Among them, the first skeletal joint point and the second skeletal joint point are two adjacent skeletal joint points, the second skeletal joint point and the third skeletal joint point are two adjacent skeletal joint points, and the second skeletal joint point is a skeletal joint point in an action state; the target point pair includes a first target point in the first connecting curve and a second target point in the second connecting curve, and the first target point and the second target point are on the same horizontal plane.

[0106] Optionally, when any three adjacent skeletal joints are bent (for example, the arm bends, which drives the clothes to bend; or the knee bends, which drives the clothes to bend), it is necessary to adjust the distance scaling coefficient of the discrete curve associated with the three adjacent skeletal joints, otherwise the discrete curves on the clothing model may be interlaced, thereby affecting the binding quality of the clothing model. Therefore, when any three adjacent skeletal joints are bent, on the basis of rotation and translation, the distance scaling is also performed so that the scaling coefficient is associated with the bending motion angle between any three adjacent skeletal joints, thereby avoiding the interlacing of curves.

[0107] Here, the steps of curve optimization include:

[0108] First, from the fourth discrete curve set, three adjacent skeletal joint points in the human body model of the current posture that are in an action state, that is, the first skeletal joint point and the second skeletal joint point and the third skeletal joint point are extracted. The so-called association means that the mapping distance between one vertex of the first curve to be optimized and the first skeletal joint point is the shortest, and the mapping distance between another vertex and the third skeletal point is the shortest.

[0109] Next, based on the distance value between the first connecting curve between the first skeletal joint and the second skeletal joint and the second connecting curve between the second skeletal joint and the third skeletal joint, the distance scaling coefficient corresponding to the first curve to be optimized is optimized so that the optimized distance scaling coefficient can achieve the following functions: when the distance value between any target point pair is smaller, the distance value between the point pairs associated with the target point pair in the first curve to be optimized can be controlled to be smaller; when the distance value between any target point pair is larger, the distance value between the point pairs associated with the target point pair in the first curve to be optimized can be controlled to be larger, thereby achieving more accurate synchronous deformation with the discrete curve maintained by the human body model and avoiding curve interlacing.

[0110] Figure 3 A schematic diagram of the distribution of the discrete curve of the non-exercise clothing model and the discrete curve of the human body model provided in an embodiment of the present application; Figure 4 One of the distribution diagrams of the original discrete curve of the clothing model after exercise and the original discrete curve of the human body model provided in the embodiment of the present application; Figure 5 One of the distribution diagrams of the optimized discrete curve of the clothing model after exercise and the optimized discrete curve of the human body model provided in the embodiment of the present application; Figure 3-Figure 5 In the figure, S1 is the discrete curve in the clothing model; S2 is the discrete curve in the human body model.

[0111] like Figure 3 As shown, the relative distance between the discrete curve of the non-moving clothing model and the discrete curve of the human body model is kept fixed; Figure 4 As shown in , if the relative distance between the original discrete curve of the clothing model after exercise and the original discrete curve of the human body model is kept fixed, the original discrete curves of the clothing model after exercise will be interlaced, seriously affecting the quality of clothing shape. Figure 5As shown, by optimizing the distance scaling coefficient corresponding to the discrete curve, the discrete curve of the clothing model after exercise can be synchronously optimized and deformed according to the curve deformation of the clothing model after exercise, so that the optimized discrete curve of the clothing model after exercise and the optimized discrete curve of the human body model maintain a certain relative distance. Based on the relationship between the discrete curve of the human body model and the discrete curve of the clothing model, the clothes can be more finely controlled in the elbows, armpits, knees and other places by adjusting parameters, so as to better ensure that the multiple layers of clothing under the clothing model will not be intertwined with each other or penetrate into the human body model within the influence range of the discrete curve of the human body model.

[0112] In some embodiments, performing curve optimization on each discrete curve in the fourth discrete curve set in step 210 further includes:

[0113] In the fourth discrete curve set, a second curve to be optimized associated with the third and fourth skeletal joint points and the fifth skeletal joint point in the action state in the human body model of the current posture is obtained;

[0114] Optimizing the distance of each line segment in the second curve to be optimized according to the distance value between each third target point in the third skeletal joint point of the third connecting curve between the fourth skeletal joint point and the fifth skeletal joint point and the target vertex of the second curve to be optimized;

[0115] Among them, the target vertex is the vertex with the closest mapping distance to the fourth and fifth skeletal joint points; the fourth skeletal joint point and the fifth skeletal joint point are two adjacent skeletal joint points, and there is at least one skeletal joint point in action, the fourth skeletal joint point, which is adjacent to the third skeletal joint point and is in a static state.

[0116] Optionally, in order to ensure that the dynamic changes of the external curves of the human body model are correct and the dynamic changes of the clothing model driven by it are also correct, the discrete curve can be re-divided by line segment distance to optimize the second curve to be optimized associated with at least one arbitrary adjacent fourth skeletal joint point and fifth skeletal joint point in action, such as the discrete curve associated with the upper limb node adjacent to the torso node, so that the line segment close to the fifth skeletal joint point that is in a static state or has a relatively small movement amplitude is driven to a smaller amplitude by the fourth skeletal joint point that has a larger movement amplitude relative to the fifth skeletal joint point, and the line segment away from the fifth skeletal joint point is driven to a larger amplitude by the fourth skeletal joint point.

[0117] Here, the steps of curve optimization include:

[0118] First, from the fourth discrete curve set, extract the second curve to be optimized that is associated with the third fourth skeletal joint point and the fifth skeletal joint point; the so-called association means that the mapping distance between the fourth skeletal point at one end of the second curve to be optimized is the shortest, and the mapping distance between the other end to the fifth skeletal point is the shortest.

[0119] Next, based on the distance values ​​between each third target point in the third connecting curve between the fourth skeletal joint point and the fifth skeletal joint point and the target vertex of the second curve to be optimized, the distances of each line segment in the second curve to be optimized are adaptively optimized so that in the optimized second curve to be optimized, the smaller the relative distance to the fifth skeletal joint point, the larger the distance between the line segment and the fifth skeletal joint point, and the smaller the distance between the line segment and the fifth skeletal joint point, thereby ensuring that the dynamic changes of clothing driven by the discrete curve of the human body model are correct.

[0120] Figure 6 The second distribution diagram of the original discrete curve of the clothing model after exercise and the original discrete curve of the human body model provided in the embodiment of the present application; Figure 7 The second schematic diagram of the distribution of the optimized discrete curve of the clothing model after exercise and the optimized discrete curve of the human body model provided in the embodiment of the present application; Figure 6-Figure 7 In , S1 is the discrete curve in the clothing model; S2 is the discrete curve in the human body model, and the top of the discrete curve is the target vertex, that is, the point with the smallest relative distance to the fifth skeletal joint point.

[0121] like Figure 6 As shown in , if the original discrete curve of the clothing model after exercise is kept relatively consistent with the line segments of the original discrete curve of the human body model, each part of the discrete curve in the clothing model will be driven equally, which will seriously affect the quality of the clothing shape. Figure 7 As shown, by optimizing the distance between curve segments, it can be achieved that in the optimized discrete curve, the smaller the relative distance between the segment and the fifth skeletal joint point, the larger the distance between the segment and the fifth skeletal joint point, and the larger the relative distance between the segment and the fifth skeletal joint point, the smaller the distance between the segment and the fifth skeletal joint point, thereby ensuring that the dynamic changes of clothing driven by the discrete curve of the human body model are correct.

[0122] In some embodiments, performing curve optimization on each discrete curve in the fourth discrete curve set in step 210 further includes:

[0123] According to the average coordinate information of adjacent vertices of the vertices of each discrete curve in the fourth discrete curve set, Laplace smoothing optimization is performed on each discrete curve in the fourth discrete curve set.

[0124] Optionally, in order to make each discrete curve in the fourth discrete curve set smoother, Laplace smoothing optimization can be performed on each discrete curve; the steps of Laplace smoothing optimization here can be, first, obtaining the average coordinate information of adjacent vertices of the vertices of each discrete curve; then, iteratively updating the vertices of each discrete curve according to the average coordinate information of the adjacent vertices until the clothing details of the clothing model, such as seams and wrinkles, are retained while achieving the desired smoothing effect, then stopping the Laplace smoothing optimization, thereby retaining the shape characteristics of the curve, while eliminating irregularities and noise on the curve, making the curve smoother and more natural, thereby improving the binding effect of the clothing model.

[0125] In some embodiments, in step 140, the mesh model of the clothing model in the initial posture is updated according to each optimized topological line and the coordinate system change information to generate a clothing form corresponding to the clothing model in the current posture, including:

[0126] According to the coordinate system change information, obtaining change information of each grid corresponding to the clothing model of the current posture;

[0127] Optimizing each of the grids according to the change information of each of the grids, the associated constraints between each of the grids, and the gradient optimization to obtain each optimized grid;

[0128] The mesh model is updated according to each of the optimized meshes and each of the optimized topological lines to generate the clothing form.

[0129] The mesh here can be a triangular face or a quadrilateral face, etc. The following takes the mesh being a triangular face as an example to describe the method provided in the embodiment of the present application.

[0130] Optionally, after obtaining the coordinate system change information, since the changes in the triangular surfaces can be obtained through the changes in the skeletal joints in the human body model, the coordinate system changes in the skeletal joints in the human body model can be transferred to the triangular surfaces to obtain the change information of each triangular surface corresponding to the clothing model in the current posture.

[0131] Here, each triangular face corresponding to the clothing model of the current posture can be constructed based on the vertices of each optimized topological line, or it can be constructed based on the reference vertices of the clothing model of the current posture obtained by affine transformation of other vertices in the initial posture model except the discrete curves in the third discrete curve set. This embodiment does not make any specific limitation on this.

[0132] Here, the change information of each triangular face can be calculated by calculating the coordinate system transformation matrix for each bone node according to its rotation, translation, scaling and other information in the current posture; for each triangular face, the displacement of the position of the triangular face in the current posture in the global coordinate system compared with the position in the initial posture is calculated according to the coordinate system transformation matrix of the bone node to which it belongs, as well as the angle flipping of the triangular face in the current posture compared with the previous frame, thereby obtaining the change information of each triangular face.

[0133] Furthermore, since each triangular face is discontinuous, it is necessary to add constraints to the association conditions between the triangular faces so that the triangular faces meet both the position requirements and the continuous smoothness requirements. It should be noted that the association constraint conditions between the triangular faces are used to limit the continuous smoothness between the faces, such as the constraint condition that the square of the matrix of two adjacent faces is equal to zero.

[0134] Next, the change information of each triangle and the associated constraints between the triangles are combined to build a triangle optimization model, and each triangle is optimized according to the triangle optimization model, and each triangle is optimized based on gradient optimization, thereby obtaining an optimized triangle that satisfies both the change information of each triangle and the associated constraints;

[0135] Finally, all optimized triangular faces and optimized topological lines are applied to the mesh model of the clothing model in the initial posture, so as to obtain the mesh clothing form of the clothes, thereby realizing the control of the clothing form corresponding to the clothing model in the current posture by the deformation of the discrete curve in the human body model, or the control of the clothing form corresponding to the clothing model in the current posture by the deformation of the triangular faces in the human body model, thereby retaining the local details of the clothing model as much as possible, and then realizing that the shape of the clothing model can correctly keep synchronous deformation with the human body model, and at the same time, it can also improve the binding quality of the three-dimensional clothing model, thereby reducing the repair time of the binder and saving the cost and production threshold of clothing binding.

[0136] In some embodiments, the first discrete curve set in step 110 is obtained according to the following steps:

[0137] From the human body model in the initial posture, discrete curves perpendicular to the axial direction of the coordinate system of each bone node in the human body model in the initial posture, and / or discrete curves parallel to the axial direction of the coordinate system of each bone node in the human body model in the initial posture are extracted to construct the first discrete curve set.

[0138] Optionally, the first discrete curve set may be obtained according to the following steps:

[0139] For each skeletal node in the human body model of the initial posture, a discrete curve is extracted in a manner perpendicular to the axial direction of its coordinate system, or in a manner parallel to the axial direction of its coordinate system; thereby, a first discrete curve set is constructed based on the extracted discrete curves perpendicular to the axial direction and / or the discrete curves parallel to the axial direction.

[0140] Similarly, discrete curves perpendicular to the axial direction and / or discrete curves parallel to the axial direction may be extracted from the human body model of the current posture in the above manner to construct a second discrete curve set.

[0141] In some embodiments, the first discrete curve set in step 110 may also be obtained according to the following steps:

[0142] From the human body model in the initial posture, extract the discrete curve in the tangential direction of the axial direction of the target bone node coordinate system in the human body model in the initial posture, the discrete curve in the target topological line in the human body model in the initial posture, and at least one of the parallel discrete curves evenly distributed in other areas except the local area where the target bone node coordinate system is located, and construct the first discrete curve set.

[0143] Optionally, the step of constructing the first discrete curve set may further include:

[0144] First, determine the target bone nodes from the human body model in the initial pose. The so-called target bone nodes are all the bone nodes in the human body model in the initial pose or the key movable bone nodes selected from all the bone nodes, such as the forearm nodes, upper arm nodes and other key bone nodes. Because the key points and lines are limited, it is convenient for later optimization and adjustment.

[0145] Next, discrete curves are extracted in the tangential direction of the axis of the target bone node, and parallel lines are further divided in the remaining area except the local area where the target bone node is located to extract evenly divided parallel discrete curves, and discrete curves are extracted from the target topological lines in the human body model of the initial posture. At least one of the discrete curve extraction methods is used to extract multiple discrete curves to construct a first discrete curve set.

[0146] Here, when the equally divided parallel discrete curves are extracted, the density of the equally divided curves can be configured according to the actual modeling quality requirements. The higher the modeling quality requirements, the denser the equally divided density can be configured, and the lower the modeling quality requirements, the sparser the equally divided density can be configured.

[0147] Similarly, at least one of the discrete curves in the axial tangential direction, the averaged parallel discrete curves, and the discrete curves in the target topological line can be extracted from the human body model in the current posture in the above manner to construct a second discrete curve set.

[0148] In some embodiments, after updating the mesh model of the clothing model in the initial posture according to each optimized topological line and the coordinate system change information to generate the clothing form corresponding to the clothing model in the current posture, the method further includes:

[0149] Determining at least one target deformation algorithm from a plurality of deformation algorithms according to a morphology optimization requirement corresponding to the clothing model of the current posture;

[0150] According to the at least one target deformation algorithm, the garment form is deformed and optimized;

[0151] Among them, the multiple deformation algorithms include multiple ones of deformation gradient algorithm, Laplace gradient, curvature gradient and conformal gradient.

[0152] Optionally, after obtaining the clothing morphology corresponding to the clothing model of the current posture, at least one target deformation algorithm can be selected from multiple deformation algorithms including deformation gradient algorithm, Laplace gradient, curvature gradient and conformal gradient according to the morphology optimization requirements corresponding to the clothing model of the current posture, so as to perform deformation optimization on the clothing morphology according to the at least one selected target deformation algorithm, so that the optimized deformed clothing morphology better matches the human body model of the current posture, thereby further improving the binding quality.

[0153] In some embodiments, in step 120, performing an affine transformation on the third discrete curve set according to the first discrete curve set and the second discrete curve set to obtain a fourth discrete curve set includes:

[0154] For each discrete curve in the third discrete curve set, the following steps are performed:

[0155] In the first discrete curve set, a first associated curve of the current discrete curve is obtained; the projection distance from the vertex of the current discrete curve to the first associated curve is the shortest;

[0156] acquiring, from the second discrete curve set, a second association curve associated with the first association curve;

[0157] According to a first direction vector of a tangent direction of a first associated point in the first associated curve, a second direction vector of a tangent direction of a second associated point in the second associated curve, an angle between the first direction vector and the second direction vector, a vertex in the current discrete curve, the first associated point, the second associated point, and a distance scaling coefficient corresponding to the current discrete curve, an affine transformation curve corresponding to the current discrete curve is obtained; the first associated point is a point that matches a vertex of the current discrete curve, and the second associated point is a point that matches the first associated point;

[0158] The fourth discrete curve set is constructed according to the affine transformation curve corresponding to each discrete curve in the third discrete curve set.

[0159] Optionally, the affine transformation step of the fourth discrete curve set includes:

[0160] For each discrete curve in the third discrete curve set corresponding to the clothing model of the initial pose, perform the following steps:

[0161] First, starting from the vertex Vc of the current discrete curve, find the first associated point Vb matching the vertex Vc in the first discrete curve set corresponding to the human body model of the initial posture; the projection distance from the vertex Vc to the first associated curve to which the first associated point Vb belongs is the shortest.

[0162] Next, the affine transformation of the first associated point Vb on the first associated curve from the initial posture to the current posture is calculated, and the affine transformation is applied to the vertex Vc of the corresponding current discrete curve, thereby obtaining the affine transformation curve corresponding to the current discrete curve.

[0163] Here, the step of obtaining the affine transformation curve corresponding to the current discrete curve further includes obtaining a second associated point Vb matching the first associated point Vb on the first associated curve in the second discrete curve set corresponding to the human body model of the current posture. b '; Here, the second associated point V b ′ is the point obtained after the first associated point Vb is transformed by the human body model posture. Next, the first direction vector C′ (V b ), the second associated point V b The second direction vector C′(V b ′), and the first direction vector C′(V b ) and the second direction vector C′(V b ′) between the angle θ; According to the first direction vector C′(V b ), the second direction vector C′(V b ') and the angle θ to calculate the first associated point Vb and the second associated point V b The rotation transformation parameter R(V b ,V b '); Then, obtain the coordinate position distance between the vertex Vc of the current discrete curve and the first associated point Vb, and the coordinate position distance between the first associated point Vb and the second associated point V b ′; according to the rotation transformation parameter R(V b ,V b′), the coordinate position of the vertex Vc of the current discrete curve, the distance scaling factor S, the coordinate position distance between the vertex Vc of the current discrete curve and the first associated point Vb, and the coordinate position distance between the first associated point Vb and the second associated point V b The coordinate position distance between ′ is obtained by affine transformation to obtain the affine transformation curve corresponding to the current discrete curve, and a fourth discrete curve set is constructed according to the affine transformation curves corresponding to each discrete curve.

[0164] Through the above-mentioned affine transformation, it can be ensured that the affine transformation of the first associated point Vb on the first associated curve from the initial posture to the current posture can be accurately mapped to the vertex Vc of each discrete curve in the corresponding third discrete curve set, thereby ensuring that each discrete curve in the fourth discrete curve set obtained by the affine transformation can correctly maintain synchronous deformation with the discrete curve of the human body model, thereby improving the binding effect and avoiding the problem that the binding engineer needs to spend a lot of time to repair the problem that the clothing form corresponding to the clothing model in the current posture cannot correctly maintain synchronous deformation with the human body model, thereby improving the binding efficiency of the three-dimensional clothing model.

[0165] Here, the distance scaling factor S is used to scale the distance from the vertex Vc of the current discrete curve to the first associated point Vb.

[0166] Figure 8 The second flow chart of the three-dimensional clothing model binding method provided in the embodiment of the present application is as follows: Figure 8 As shown, the method includes:

[0167] First, a human body model of an initial posture is obtained, and discrete curves are extracted from the human body model of the initial posture to form a first discrete curve set; and a human body model of a current posture is obtained, and discrete curves are extracted from the human body model of the current posture to form a second discrete curve set; and a clothing model of an initial posture is obtained, and discrete curves are extracted from the clothing model of the initial posture to form a third discrete curve set;

[0168] Next, a local coordinate system of each bone node coordinate system in the human body model of the current posture is formed on the second discrete curve set;

[0169] Next, an affine transformation is performed according to the first discrete curve set, the second discrete curve set and the third discrete curve set to obtain a fourth discrete curve set corresponding to the clothing model of the current posture; and each discrete curve in the fourth discrete curve set is optimized, and an affine transformation is first performed on the topological line according to the triangular surface coordinate system of the human body model of the current posture to obtain each topological line of the clothing model of the current posture; according to the coordinate system change information of each bone node coordinate system in the human body model of the current posture and the optimized discrete curve, each topological line of the clothing model of the current posture is optimized to retain the clothing details of the clothing model after operation;

[0170] Furthermore, according to the reference vertices of the clothing model of the current posture obtained by affine transformation of other vertices except the discrete curves in the third discrete curve set in the initial posture model, the triangular faces corresponding to the clothing model of the current posture are constructed based on the reference vertices; and the change information of each triangular face obtained based on the coordinate system change information, as well as the associated constraint conditions and gradient optimization between each triangular face, are optimized to make the triangular faces smooth;

[0171] Finally, all optimized triangles and optimized topological lines are applied to the mesh model of the clothing model in the initial posture, so as to obtain the clothing form of the current clothing model.

[0172] The three-dimensional clothing model binding device provided by the present invention is described below. The three-dimensional clothing model binding device described below and the three-dimensional clothing model binding method described above can be referred to each other.

[0173] Fig. 9 A schematic diagram of the structure of a three-dimensional clothing model binding device provided in an embodiment of the present application; Fig. 9 As shown, the device comprises:

[0174] An acquisition unit 910 is used to acquire a first discrete curve set, a second discrete curve set and a third discrete curve set; the first discrete curve set is extracted from a human body model in an initial posture, the second discrete curve set is extracted from a human body model in a current posture, and the third discrete curve set is extracted from a clothing model in the initial posture;

[0175] A transformation unit 920, configured to perform an affine transformation on the third discrete curve set according to the first discrete curve set and the second discrete curve set to obtain a fourth discrete curve set;

[0176] The optimization unit 930 is used to optimize each topological line of the clothing model in the current posture according to the coordinate system change information of each skeletal node coordinate system in the human body model in the current posture and the fourth discrete curve set;

[0177] The morphology generation unit 940 is used to update the mesh model of the clothing model in the initial posture according to each optimized topological line and the coordinate system change information, and generate a clothing morphology corresponding to the clothing model in the current posture.

[0178] The device provided in this embodiment performs affine transformation on the first discrete curve set extracted from the human body model in the initial posture, the second discrete curve set extracted from the human body model in the current posture, and the third discrete curve set extracted from the clothing model in the initial posture, so as to accurately simulate and obtain the fourth discrete curve set corresponding to the clothing model in the current posture, and optimizes and generates the clothing form corresponding to the clothing model in the current posture by combining the coordinate system change information of the coordinate system of each bone node in the human body model in the current posture and the fourth discrete curve set, so as to guide the clothing form change of the clothing model outside the human body model according to the posture change of the human body model, so that the clothing form corresponding to the clothing model in the current posture correctly keeps synchronous deformation with the human body model, thereby greatly reducing the repair time required by the binder, improving the binding efficiency of the three-dimensional clothing model, and improving the binding effect.

[0179] In some embodiments, the transform unit is further configured to:

[0180] According to each grid coordinate system in the human body model of the current posture, each topological line of the clothing model of the current posture is obtained by affine transformation.

[0181] In some embodiments, the optimization unit is specifically configured to:

[0182] Performing curve optimization on each discrete curve in the fourth discrete curve set, and optimizing each topological line according to each optimized discrete curve; and,

[0183] According to the associated constraints between the topological lines, the coordinate system change information of the coordinate system of each skeletal node is transmitted to each topological line through skinning information interpolation, and each topological line is optimized.

[0184] In some embodiments, the optimization unit is further configured to:

[0185] According to a three-dimensional clothing model binding method provided by the present invention, the curve optimization of each discrete curve in the fourth discrete curve set includes:

[0186] In the fourth discrete curve set, a first curve to be optimized associated with a first skeletal joint point, a second skeletal joint point, and a third skeletal joint point in the human body model of the current posture is obtained;

[0187] Optimizing the distance scaling coefficient corresponding to the first curve to be optimized according to the distance values ​​of each target point pair between the first connection curve between the first skeletal joint point and the second skeletal joint point and the second connection curve between the second skeletal joint point and the third skeletal joint point;

[0188] Performing curve optimization on the first curve to be optimized according to the optimized distance scaling coefficient;

[0189] Among them, the first skeletal joint point and the second skeletal joint point are two adjacent skeletal joint points, the second skeletal joint point and the third skeletal joint point are two adjacent skeletal joint points, and the second skeletal joint point is a skeletal joint point in an action state; the target point pair includes a first target point in the first connecting curve and a second target point in the second connecting curve, and the first target point and the second target point are on the same horizontal plane.

[0190] In some embodiments, the optimization unit is further configured to:

[0191] In the fourth discrete curve set, a second curve to be optimized associated with a fourth skeletal joint point and a fifth skeletal joint point in the human body model of the current posture is obtained;

[0192] Optimizing the distance of each line segment in the second curve to be optimized according to the distance value between each third target point in the third connecting curve between the fourth skeletal joint point and the fifth skeletal joint point and the target vertex of the second curve to be optimized;

[0193] Among them, the target vertex is the vertex with the shortest mapping distance to the fifth skeletal joint point; the fourth skeletal joint point and the fifth skeletal joint point are two adjacent skeletal joint points, and there is at least one skeletal joint point in an action state.

[0194] In some embodiments, the optimization unit is further configured to:

[0195] According to the average coordinate information of adjacent vertices of the vertices of each discrete curve in the fourth discrete curve set, Laplace smoothing optimization is performed on each discrete curve in the fourth discrete curve set.

[0196] In some embodiments, the morphology generation unit is specifically configured to:

[0197] According to the coordinate system change information, obtaining change information of each grid corresponding to the clothing model of the current posture;

[0198] Optimizing each of the grids according to the change information of each of the grids, the associated constraints between each of the grids, and the gradient optimization to obtain each optimized grid;

[0199] The mesh model is updated according to each of the optimized meshes and each of the optimized topological lines to generate the clothing form.

[0200] In some embodiments, the acquisition unit is specifically configured to:

[0201] From the human body model in the initial posture, discrete curves perpendicular to the axial direction of the coordinate system of each bone node in the human body model in the initial posture, and / or discrete curves parallel to the axial direction of the coordinate system of each bone node in the human body model in the initial posture are extracted to construct the first discrete curve set.

[0202] In some embodiments, the acquisition unit is further configured to:

[0203] From the human body model in the initial posture, extract the discrete curve in the tangential direction of the axial direction of the target bone node coordinate system in the human body model in the initial posture, the discrete curve in the target topological line in the human body model in the initial posture, and at least one of the parallel discrete curves evenly distributed in other areas except the local area where the target bone node coordinate system is located, and construct the first discrete curve set.

[0204] In some embodiments, a deformation unit is further included, specifically configured to:

[0205] Determining at least one target deformation algorithm from a plurality of deformation algorithms according to a morphology optimization requirement corresponding to the clothing model of the current posture;

[0206] According to the at least one target deformation algorithm, the garment form is deformed and optimized;

[0207] Among them, the multiple deformation algorithms include multiple ones of deformation gradient algorithm, Laplace gradient, curvature gradient and conformal gradient.

[0208] In some embodiments, the transform unit is specifically configured to:

[0209] For each discrete curve in the third discrete curve set, the following steps are performed:

[0210] In the first discrete curve set, a first associated curve of the current discrete curve is obtained; the projection distance from the vertex of the current discrete curve to the first associated curve is the shortest;

[0211] acquiring, from the second discrete curve set, a second association curve associated with the first association curve;

[0212] According to a first direction vector of a tangent direction of a first associated point in the first associated curve, a second direction vector of a tangent direction of a second associated point in the second associated curve, an angle between the first direction vector and the second direction vector, a vertex in the current discrete curve, the first associated point, the second associated point, and a distance scaling coefficient corresponding to the current discrete curve, an affine transformation curve corresponding to the current discrete curve is obtained; the first associated point is a point that matches a vertex of the current discrete curve, and the second associated point is a point that matches the first associated point;

[0213] The fourth discrete curve set is constructed according to the affine transformation curve corresponding to each discrete curve in the third discrete curve set.

[0214] Fig.10 An example of a physical structure diagram of an electronic device is shown in FIG. Fig.10 As shown, the electronic device may include: a processor (processor) 1010, a communication interface (Communications Interface) 1020, a memory (memory) 1030 and a communication bus 1040, wherein the processor 1010, the communication interface 1020, and the memory 1030 communicate with each other through the communication bus 1040. The processor 1010 can call the logic instructions in the memory 1030 to execute the three-dimensional clothing model binding method, which includes: obtaining a first discrete curve set, a second discrete curve set and a third discrete curve set; the first discrete curve set is extracted from the human body model in an initial posture, the second discrete curve set is extracted from the human body model in a current posture, and the third discrete curve set is extracted from the clothing model in the initial posture; according to the first discrete curve set and the second discrete curve set, the third discrete curve set is affine transformed to obtain a fourth discrete curve set; according to the coordinate system change information of the coordinate system of each bone node in the human body model in the current posture and the fourth discrete curve set, each topological line of the clothing model in the current posture is optimized; according to each optimized topological line and the coordinate system change information, the mesh model of the clothing model in the initial posture is updated to generate a clothing form corresponding to the clothing model in the current posture.

[0215] In addition, the logic instructions in the above-mentioned memory 1030 can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. According to such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0216] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the three-dimensional clothing model binding method provided by the above methods, and the method includes: obtaining a first discrete curve set, a second discrete curve set and a third discrete curve set; the first discrete curve set is extracted from a human body model in an initial posture, the second discrete curve set is extracted from a human body model in a current posture, and the third discrete curve set is extracted from a clothing model in the initial posture; according to the first discrete curve set and the second discrete curve set, the third discrete curve set is affine transformed to obtain a fourth discrete curve set; according to the coordinate system change information of each bone node coordinate system in the human body model in the current posture and the fourth discrete curve set, each topological line of the clothing model in the current posture is optimized; according to each optimized topological line and the coordinate system change information, the mesh model of the clothing model in the initial posture is updated to generate a clothing form corresponding to the clothing model in the current posture.

[0217] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the three-dimensional clothing model binding method provided by the above-mentioned methods, the method comprising: obtaining a first discrete curve set, a second discrete curve set and a third discrete curve set; the first discrete curve set is extracted from a human body model in an initial posture, the second discrete curve set is extracted from a human body model in a current posture, and the third discrete curve set is extracted from a clothing model in the initial posture; according to the first discrete curve set and the second discrete curve set, the third discrete curve set is affine transformed to obtain a fourth discrete curve set; according to the coordinate system change information of the coordinate system of each bone node in the human body model in the current posture and the fourth discrete curve set, each topological line of the clothing model in the current posture is optimized; according to each optimized topological line and the coordinate system change information, the mesh model of the clothing model in the initial posture is updated to generate a clothing form corresponding to the clothing model in the current posture.

[0218] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0219] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0220] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A three-dimensional clothing model binding method, characterized in that: include: Acquire a first discrete curve set, a second discrete curve set and a third discrete curve set; the first discrete curve set is extracted from a human body model in an initial posture, the second discrete curve set is extracted from a human body model in a current posture, and the third discrete curve set is extracted from a clothing model in the initial posture; According to the first discrete curve set and the second discrete curve set, performing affine transformation on the third discrete curve set to obtain a fourth discrete curve set; Optimizing each topological line of the clothing model in the current posture according to the coordinate system change information of each skeletal node coordinate system in the human body model in the current posture and the fourth discrete curve set; According to each optimized topological line and the coordinate system change information, the mesh model of the clothing model in the initial posture is updated to generate a clothing form corresponding to the clothing model in the current posture; The optimizing of each topological line of the clothing model of the current posture according to the coordinate system change information of each skeletal node coordinate system in the human body model of the current posture and the fourth discrete curve set comprises: Performing curve optimization on each discrete curve in the fourth discrete curve set, and optimizing each topological line according to each optimized discrete curve; and, According to the associated constraints between the topological lines, the coordinate system change information of the coordinate system of each skeletal node is transmitted to each topological line through skinning information interpolation, and each topological line is optimized.

2. The three-dimensional clothing model binding method according to claim 1, characterized in that: The step of obtaining each topological line of the clothing model of the current posture includes: According to each grid coordinate system in the human body model of the current posture, each topological line of the clothing model of the current posture is obtained by affine transformation.

3. The three-dimensional clothing model binding method according to claim 1, characterized in that: The performing curve optimization on each discrete curve in the fourth discrete curve set includes: In the fourth discrete curve set, a first curve to be optimized associated with a first skeletal joint point, a second skeletal joint point, and a third skeletal joint point in the human body model of the current posture is obtained; Optimizing the distance scaling coefficient corresponding to the first curve to be optimized according to the distance values ​​of each target point pair between the first connection curve between the first skeletal joint point and the second skeletal joint point and the second connection curve between the second skeletal joint point and the third skeletal joint point; Performing curve optimization on the first curve to be optimized according to the optimized distance scaling coefficient; Among them, the first skeletal joint point and the second skeletal joint point are two adjacent skeletal joint points, the second skeletal joint point and the third skeletal joint point are two adjacent skeletal joint points, and the second skeletal joint point is a skeletal joint point in an action state; the target point pair includes a first target point in the first connecting curve and a second target point in the second connecting curve, and the first target point and the second target point are on the same horizontal plane.

4. The three-dimensional clothing model binding method according to claim 1, characterized in that: The performing curve optimization on each discrete curve in the fourth discrete curve set includes: In the fourth discrete curve set, a second curve to be optimized associated with a fourth skeletal joint point and a fifth skeletal joint point in the human body model of the current posture is obtained; Optimizing the distance of each line segment in the second curve to be optimized according to the distance value between each third target point in the third connecting curve between the fourth skeletal joint point and the fifth skeletal joint point and the target vertex of the second curve to be optimized; Among them, the target vertex is the vertex with the shortest mapping distance to the fifth skeletal joint point; the fourth skeletal joint point and the fifth skeletal joint point are two adjacent skeletal joint points, and there is at least one skeletal joint point in an action state.

5. The three-dimensional clothing model binding method according to claim 1, characterized in that: The performing curve optimization on each discrete curve in the fourth discrete curve set includes: According to the average coordinate information of adjacent vertices of the vertices of each discrete curve in the fourth discrete curve set, Laplace smoothing optimization is performed on each discrete curve in the fourth discrete curve set.

6. The three-dimensional clothing model binding method according to any one of claims 1 to 5, characterized in that: The updating of the mesh model of the clothing model in the initial posture according to each optimized topological line and the coordinate system change information to generate a clothing form corresponding to the clothing model in the current posture includes: According to the coordinate system change information, obtaining change information of each grid corresponding to the clothing model of the current posture; Optimizing each of the grids according to the change information of each of the grids, the associated constraints between each of the grids, and the gradient optimization to obtain each optimized grid; The mesh model is updated according to each of the optimized meshes and each of the optimized topological lines to generate the clothing form.

7. The three-dimensional clothing model binding method according to any one of claims 1 to 5, characterized in that: The first discrete curve set is obtained according to the following steps: From the human body model in the initial posture, discrete curves perpendicular to the axial direction of the coordinate system of each bone node in the human body model in the initial posture, and / or discrete curves parallel to the axial direction of the coordinate system of each bone node in the human body model in the initial posture are extracted to construct the first discrete curve set.

8. The three-dimensional clothing model binding method according to any one of claims 1 to 5, characterized in that: The first discrete curve set is obtained according to the following steps: From the human body model in the initial posture, extract the discrete curve in the tangential direction of the axial direction of the target bone node coordinate system in the human body model in the initial posture, the discrete curve in the target topological line in the human body model in the initial posture, and at least one of the parallel discrete curves evenly distributed in other areas except the local area where the target bone node coordinate system is located, and construct the first discrete curve set.

9. The three-dimensional clothing model binding method according to any one of claims 1 to 5, characterized in that: After updating the mesh model of the clothing model in the initial posture according to each optimized topological line and the coordinate system change information to generate the clothing form corresponding to the clothing model in the current posture, the method further includes: Determining at least one target deformation algorithm from a plurality of deformation algorithms according to a morphology optimization requirement corresponding to the clothing model of the current posture; According to the at least one target deformation algorithm, the garment form is deformed and optimized; Among them, the multiple deformation algorithms include multiple ones of deformation gradient algorithm, Laplace gradient, curvature gradient and conformal gradient.

10. The three-dimensional clothing model binding method according to any one of claims 1 to 5, characterized in that: The step of performing affine transformation on the third discrete curve set according to the first discrete curve set and the second discrete curve set to obtain a fourth discrete curve set includes: For each discrete curve in the third discrete curve set, the following steps are performed: In the first discrete curve set, a first associated curve of the current discrete curve is obtained; the projection distance from the vertex of the current discrete curve to the first associated curve is the shortest; acquiring, from the second discrete curve set, a second association curve associated with the first association curve; According to a first direction vector of a tangent direction of a first associated point in the first associated curve, a second direction vector of a tangent direction of a second associated point in the second associated curve, an angle between the first direction vector and the second direction vector, a vertex in the current discrete curve, the first associated point, the second associated point, and a distance scaling coefficient corresponding to the current discrete curve, an affine transformation curve corresponding to the current discrete curve is obtained; the first associated point is a point that matches a vertex of the current discrete curve, and the second associated point is a point that matches the first associated point; The fourth discrete curve set is constructed according to the affine transformation curve corresponding to each discrete curve in the third discrete curve set.

11. A three-dimensional clothing model binding device, characterized in that: include: An acquisition unit, configured to acquire a first discrete curve set, a second discrete curve set, and a third discrete curve set; the first discrete curve set is extracted from a human body model in an initial posture, the second discrete curve set is extracted from a human body model in a current posture, and the third discrete curve set is extracted from a clothing model in the initial posture; a transformation unit, configured to perform an affine transformation on the third discrete curve set according to the first discrete curve set and the second discrete curve set to obtain a fourth discrete curve set; an optimization unit, configured to optimize each topological line of the clothing model in the current posture according to the coordinate system change information of each skeletal node coordinate system in the human body model in the current posture and the fourth discrete curve set; A morphology generation unit, configured to update the mesh model of the clothing model in the initial posture according to each optimized topological line and the coordinate system change information, and generate a clothing morphology corresponding to the clothing model in the current posture; The optimization unit is specifically used for: Performing curve optimization on each discrete curve in the fourth discrete curve set, and optimizing each topological line according to each optimized discrete curve; and, According to the associated constraints between the topological lines, the coordinate system change information of the coordinate system of each skeletal node is transmitted to each topological line through skinning information interpolation, and each topological line is optimized.

12. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the three-dimensional clothing model binding method according to any one of claims 1 to 10 is implemented.

13. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the three-dimensional clothing model binding method according to any one of claims 1 to 10 is implemented.

14. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the three-dimensional clothing model binding method according to any one of claims 1 to 10 is implemented.

Images