WO2023158370A2 - 表情包生成方法及设备 - Google Patents

表情包生成方法及设备 Download PDF

Info

Publication number
WO2023158370A2
WO2023158370A2 PCT/SG2023/050062 SG2023050062W WO2023158370A2 WO 2023158370 A2 WO2023158370 A2 WO 2023158370A2 SG 2023050062 W SG2023050062 W SG 2023050062W WO 2023158370 A2 WO2023158370 A2 WO 2023158370A2
Authority
WO
WIPO (PCT)
Prior art keywords
expression
emoticon
component
target
package
Prior art date
Application number
PCT/SG2023/050062
Other languages
English (en)
French (fr)
Other versions
WO2023158370A3 (zh
Inventor
曾伟宏
王旭
刘晶
桑燊
刘海珊
Original Assignee
脸萌有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 脸萌有限公司 filed Critical 脸萌有限公司
Publication of WO2023158370A2 publication Critical patent/WO2023158370A2/zh
Publication of WO2023158370A3 publication Critical patent/WO2023158370A3/zh

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G06F16/5866—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using information manually generated, e.g. tags, keywords, comments, manually generated location and time information
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T11/00—2D [Two Dimensional] image generation
    • G06T11/60—Editing figures and text; Combining figures or text
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T13/00—Animation
    • G06T13/20—3D [Three Dimensional] animation
    • G06T13/40—3D [Three Dimensional] animation of characters, e.g. humans, animals or virtual beings
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Definitions

  • an embodiment of the present disclosure provides a method for generating an emoticon package, including: acquiring a material map of a plurality of components on an avatar; determining the global position of the component according to the material map; the following target pose; generating the emoticon package according to the material map, the global position and the target pose; wherein, in the emoticon package, the expression change of the avatar includes changing from the initial expression to the Describe the target expression.
  • an embodiment of the present disclosure provides an emoticon package generation device, including: an acquisition unit, configured to acquire a material map of a plurality of components on an avatar; a position determination unit, configured to determine the component according to the material map The global position of; A posture determination unit is used to determine the target posture of the component under the target expression; a generation unit is used to generate the expression package according to the material map, the global position and the target posture, wherein , in the expression package, the expression of the avatar changes from an initial expression to the target expression.
  • an embodiment of the present disclosure provides an electronic device, including: at least one processor and a memory; the memory stores computer-executable instructions; the at least one processor executes the computer-executable instructions stored in the memory, so that the At least one processor executes the emoticon package generation method described in the first aspect above.
  • an embodiment of the present disclosure provides a computer-readable storage medium, where computer-executable instructions are stored in the computer-readable storage medium, and when the processor executes the computer-executable instructions, the above-mentioned first aspect is implemented. Method for generating emoticons.
  • a computer program product in the fifth aspect, according to one or more embodiments of the present disclosure, includes computer-executable instructions, and when the processor executes the computer-executable instructions, the above-mentioned first aspect is realized.
  • a computer program is provided. When a processor executes the computer program, the method for generating emoticons as described in the first aspect above is implemented.
  • the emoticon package generation method and device obtained the material maps of multiple components on the avatar, determine the global position of the components according to the material maps of the components, determine the target posture of the components under the target expression, and determine the target posture of the components according to the material maps of the components. , the global position and the target pose to generate an emoticon package.
  • the expression change of the avatar in the expression package includes the change from the initial expression to the target expression. Therefore, the user only needs to prepare the material map of the parts on the avatar, and does not need to design the expression of the avatar, and does not need to care about how to combine multiple frames of images, which effectively improves the production efficiency of expression packs and reduces the difficulty of production.
  • Fig. 1 is an example diagram of an application scenario provided by an embodiment of the present disclosure
  • Fig. 2 is a schematic flow diagram of a method for generating emoticons provided by an embodiment of the present disclosure
  • Fig. 3a is an example diagram of a material map of multiple components
  • Fig. 1 is an example diagram of an application scenario provided by an embodiment of the present disclosure
  • Fig. 2 is a schematic flow diagram of a method for generating emoticons provided by an embodiment of the present disclosure
  • Fig. 3a is an example diagram of a material map of multiple components
  • FIG. 3b An example diagram of component classification and component naming
  • Figure 4 is a schematic flow diagram of the emoticon package generation method provided by the embodiment of the present disclosure
  • Figure 5 is an expression diagram of an animation character image provided by the embodiment of the disclosure
  • FIG. 7 is a schematic diagram of the hardware structure of the electronic device provided in the embodiment of the present disclosure.
  • Avatar a virtual character depicted by an image in a computing device, such as an anime character.
  • Parts of the virtual image components of the virtual image; for example, the eyes, nose, mouth, etc. of an animation character are all parts of the animation character.
  • Material map of a component the layer on which the component is drawn; wherein, different components can correspond to different material maps, that is, different components can correspond to different layers, so as to improve the flexibility of component combination.
  • the global position of the component the image position of the component in the emoticon in the emoticon package, where the emoticon includes multiple A virtual image obtained by combining parts.
  • Posture of parts In the emoticon package, the expression of the avatar will change, and this change can be subdivided into changes in the posture of the parts, for example, changes in the degree of inclination and bending of the parts, so the posture of the parts can include The degree of inclination, bending, stretching, etc.
  • the idea of the embodiment of the present disclosure is provided: In the process of making a dynamic emoticon package, the user usually needs to use a drawing tool to draw a multi-frame sketch, and then combine it into a dynamic emoticon package. This process takes a long time and has high technical threshold requirements.
  • embodiments of the present disclosure propose a method and device for generating emoticons.
  • the corresponding dynamic emoticons are generated based on the material maps, positions and postures of the multiple components.
  • the user only needs to prepare the material maps of multiple parts on the avatar, and does not need to consider the drawing of each frame of emoticons, which effectively reduces the difficulty of making emoticons and improves the production efficiency.
  • Figure 1 is An example diagram of an application scenario provided by an embodiment of the present disclosure. As shown in Figure 1, the application scenario is a dynamic emoticon package production scenario.
  • the user can prepare material maps of multiple parts on the avatar on the terminal 101, and the terminal 101 can create dynamic emoticons based on the material maps of the multiple parts.
  • the terminal 101 may send the material maps of the multiple parts to the server 102, and the server 102 may create a dynamic emoticon package based on the material maps of the multiple parts.
  • the user wants to make a unique and interesting dynamic emoticon package, he can click on the terminal to enter the emoticon pack creation page provided by the chat application, and enter some cartoon animals designed by himself on the emoticon pack creation page , anime characters and other avatars, or you can also input some material pictures of multiple parts of avatars such as cartoon animals and anime characters that are publicly authorized and available, and get the prepared emoticons through the emoticon package production program Bag.
  • the emoticon package generation method and device provided by the embodiments of the present disclosure will be described in conjunction with the application scenario shown in FIG. 1 . It should be noted that the above application scenarios are only shown to facilitate understanding of the spirit and principle of the present disclosure, and the implementation manners of the present disclosure are not limited in this regard.
  • the embodiments of the present disclosure can be applied to any applicable scene.
  • the electronic equipment may be a terminal or a server.
  • the terminal may be a personal digital assistant (personal digital assistant, PDA for short) device, a handheld device with a wireless communication function (such as a smart phone, a tablet computer), a computing device (such as a personal computer (personal computer, PC for short)), a vehicle-mounted devices, wearable devices (such as smart watches, smart bracelets), smart home devices (such as smart display devices), etc.
  • the server may be an integral server or a distributed server spanning multiple computers or computer data centers.
  • FIG. 2 is a first schematic flow diagram of a method for generating an emoticon package provided by an embodiment of the present disclosure.
  • the emoticon package generation method includes:
  • material maps of multiple components on the avatar may be acquired, and the multiple components belong to the same avatar.
  • the user can input material images of multiple parts through the input control on the emoticon pack creation page; for another example, the component material maps of multiple avatars can be displayed on the emoticon pack creation page, from which the user can select multiple parts of the same avatar material map.
  • the material maps corresponding to different components have the same size, and for each component, the position of the component in the material map is The global position of the component. Therefore, the global position of the component can be determined by determining the position of the component in the material map. Therefore, the accuracy of the global position is improved by making the size of the material map consistent and the position of the part in the material map determines the global position of the part in the expression map.
  • the global position of the component is randomly determined within the corresponding position range of the component, wherein different position ranges are set in advance for different components.
  • the target expression includes one or more expressions, such as happy, angry, sad and so on.
  • the target emoticon input by the user may be received, or the target emoticon selected by the user among one or more emoticons may be acquired, or the target emoticon may be determined as the default emoticon.
  • the action gestures corresponding to the multiple components under one or more expressions the action gestures of the multiple components under the target expression are determined.
  • the action gestures of the multiple components under the target expression are called the target gestures of the multiple components. For example, get the user's input text "Happy”, determine the target expression "Happy” according to the input text, and determine the target expression "Happy” under the head in the action posture of multiple parts under multiple expressions , face and other parts of the action posture.
  • an emoticon package according to the material map, the global position, and the target pose, and the expression change of the avatar in the emoticon package includes the change from the initial expression to the target expression.
  • the initial expression refers to the expression in the expression package at the initial moment, that is, the expression shown in the first frame image in the expression package or the image at time 0.
  • the target posture of the component is the posture of the component under the target expression, and the expression of the avatar in the expression package will change from the initial expression to the target expression, which means that the posture of each component in the expression package is also gradual. Therefore, after determining the global position and target pose of multiple components, for each component, the pose of the component at multiple moments can be determined based on the target pose of the component.
  • the material maps of the multiple components are combined to obtain the expression graph at that time.
  • emoticons at multiple moments are obtained.
  • an emoticon package is obtained.
  • the expression change of the avatar in the emoticon package also includes changing from the target expression to the initial expression, that is, the expression change of the avatar in the emoticon package is from the initial expression to the target expression, and then from the target expression to the initial expression. For example, the avatar went from not smiling to smiling, and from smiling to not smiling.
  • the global position of the component is determined, based on the global position of the component and the target pose of the component under the target expression, the expression package is obtained, and the expression map corresponding to the target expression can also be obtained .
  • the user only needs to input the material map of the component to obtain a high-quality emoticon package, which effectively improves the production efficiency of the emoticon package and the emoticon map, reduces the difficulty of making the emoticon pack and the emoticon map, and improves the user experience.
  • the virtual image includes an avatar, especially an anime character.
  • the production of emoticon packs of anime characters is more difficult, and it is usually necessary to draw 3D dynamic effects through 2D images.
  • the user can obtain the dynamic expression package of the animation character image by inputting the material pictures of multiple parts on the animation character image, and in the process of making the expression package, the position and posture of each component are taken into account, which not only improves It improves the production efficiency of dynamic emoticons for animation characters, reduces the difficulty of production, and ensures the quality of emoticons.
  • necessary and non-essential components are preset.
  • the necessary components are necessary components for making the emoticon package of the avatar
  • the non-essential components are optional components for making the emoticon package of the avatar.
  • S201 includes: Acquiring material maps of multiple necessary components on the avatar. Specifically, the user can be informed in advance of the necessary components for making emoticons, for example, display the name of the necessary components on the emoticon creation page, and for example, mark whether the component is a necessary component around the input control corresponding to the component; Material drawings of these necessary components must be entered.
  • the user may also input material maps of non-essential components in addition to inputting material maps of necessary components, so as to further improve and enrich the avatar.
  • the necessary parts may include eyebrow parts, upper eyelash parts, pupil parts, mouth parts and face parts. Among them, these components can accurately depict the appearance of the animation character, and can also vividly express various emotions, which is conducive to ensuring the integrity of the virtual image and improving the vividness of the expression of the virtual image.
  • non-essential components may include at least one of the following: foreground components, hair components, head decoration components, lower eyelash components, eye white components, nose components, ear components, body components, and background components.
  • the foreground component refers to the component located in front of the avatar according to the spatial relationship.
  • multiple component categories are preset. A plurality of component categories may be displayed prior to acquiring the footage of the plurality of components on the avatar. Therefore, it is convenient for the user to input the material map of the component according to the category of the component.
  • the component category can be divided into multiple levels of categories, and when the component category is divided into two levels, the component category can be divided into a parent category and a subcategory under the parent category.
  • the parent class includes at least one of the following: a foreground component, a hair component, a head component, a body component, and a background component.
  • Subclasses under hair parts include at least one of the following: head decoration parts, front hair parts, ear front hair parts, ear back hair parts, back hair parts; subclasses under head parts include head decoration parts, eyebrow parts , Eye Parts, Nose Parts, Mouth Parts, Face Parts, Ear Parts.
  • subcategories can be further divided into different categories.
  • the subcategories under eye parts may include at least one of the following: upper eyelash parts, lower eyelash parts, pupil parts, and eye white parts.
  • FIG. 3a is an example diagram of material diagrams of multiple components.
  • the material maps corresponding to the eyebrow parts, upper eyelash parts, pupil parts, mouth parts, face parts, and body parts of the anime character are given. It can be seen that these material maps are of the same size. The material images of these parts are combined and spliced to obtain the corresponding animation character image.
  • a component may correspond to one or more material graphs.
  • the avatar has multiple head decoration parts, so the head decoration parts can correspond to multiple material images.
  • the material graph corresponds to a unique image identifier, that is, different material graphs correspond to different image identifiers. Therefore, in the process of generating emoticon packs according to the material maps of components, the material maps and the components corresponding to the material maps can be distinguished through image identification.
  • the image identifier includes an image name.
  • the image names of the multiple material images corresponding to the foreground component are foreground one, foreground two, respectively; the hair decoration part
  • the image names of the multiple material images corresponding to the pieces are hair decoration part 1, hair decoration part 2, ..., and so on.
  • FIG. 3b is an example diagram of component classification and component naming, wherein the left area shows multiple components, and the right area shows the naming methods of material maps under multiple component types, "Layer” refers to the material image, and "png" is the image format of the material image.
  • a possible implementation of S202 includes: determining the circumscribing matrix of the component in the material map of the component; determining the global position of the component according to the circumscribing matrix of the component. Therefore, by solving the circumscribed matrix of the component, the accuracy of the global position of the component is improved.
  • the circumscribed matrix of the component can be identified in the material diagram of the component, and the position of the circumscribed matrix of the component in the material diagram can be obtained.
  • the position of the circumscribing matrix in the material graph includes the pixel coordinates of the four vertices of the circumscribing matrix in the material graph.
  • the image position of the component in the material map reflects the global position of the component, so the global position of the component can be determined to be the position of the circumscribed matrix of the component.
  • the image channel of the material image of the component includes a position channel.
  • the channel value of the pixel point in the position channel reflects whether the pixel point is located in the pattern area of the component. For example: If the channel value of the pixel point in the position channel is 1, it is determined that the pixel point is located in the pattern area; if the channel value of the pixel point in the position channel is 0, it is determined that the pixel point is not located in the pattern area.
  • the circumscribed matrix of the component in the material image can be determined through the values of the position channels of multiple pixel points in the material image, which improves the accuracy of the circumscribed matrix.
  • the material image of the part is an RGBA four-channel image, that is, the image channels of the material image of the part include R channel, G channel, B channel and A channel.
  • the R channel, the G channel, and the B channel are the red, green, and blue color channels of the image respectively
  • the A channel is the position channel of the image. Therefore, the channel value of each pixel in the A channel can be obtained in the material image of the component, and the circumscribed matrix of the component can be determined according to the channel value of the A channel of each pixel.
  • the bounding matrix of the component may also be a minimum bounding rectangle (MBR) of the component, so as to improve the accuracy of the global position of the component.
  • MLR minimum bounding rectangle
  • a possible implementation of S203 includes: determining the expression action corresponding to the target expression according to the preset correspondence between multiple expression types and expression actions, and the target expression corresponding The expression action includes the target pose of multiple parts under the target expression.
  • the preset correspondence between multiple expression types and emoticon actions professionals can pre-set in advance, so as to reduce the difficulty of making emoticon packages.
  • different types of expressions correspond to different facial expressions, and the facial expressions include gestures of multiple preset components. Under different expression types, the preset components may be the same or different.
  • the expression type of "happy” includes the action postures of the eyebrow parts, upper eyelash parts, pupil parts, mouth parts, and face parts.
  • the action postures of the eyebrow parts, upper eyelash parts and mouth parts can all be curved upwards
  • the "question” expression can also include emoji parts (such as "question mark"), and the action posture of the mouth part can be presented as straight or with the corners of the mouth down .
  • the target expression type to which the target expression belongs can be determined among multiple expression types.
  • the preset corresponding relationship between multiple expression types and facial expressions determines the facial expressions corresponding to the target facial expression type, that is, the facial expressions corresponding to the target facial expressions, and finds the expressions of multiple parts on the avatar from the facial expressions corresponding to the target facial expressions.
  • Action pose that is, to obtain the target pose of multiple components.
  • the action posture of the component includes the attitude angle of the component, and the attitude angle may include at least one of the following: pitch angle, yaw angle, and roll angle. Therefore, by combining the position of the component and the attitude angle of the component, accurate Express the expression of the avatar in the emoticon package.
  • FIG. 4 is a second schematic flow diagram of a method for generating an emoticon package provided by an embodiment of the present disclosure.
  • the emoticon package generation method includes:
  • the action posture of the component at multiple moments can refer to the foregoing embodiments, and details are not repeated here.
  • the expressions at these multiple moments include the initial expression and the target expression, and also include the differences between the initial expression and the target expression. Change expressions. Therefore, in order to reflect the expressions at the multiple moments, it is necessary to determine the action postures of the multiple components at the multiple moments.
  • the process of changing the expression of the avatar in the expression package from the initial expression to the target expression is a process in which the action posture of the component gradually increases to the target posture, and this process is slow and non-linear.
  • this embodiment uses a periodic function to determine the action posture of the component at multiple moments.
  • the target posture is processed through a periodic function to obtain the action posture of the component at multiple moments within the period of the periodic function.
  • the expression change process of the avatar in the emoticon package only includes the process of changing from the initial expression to the target expression, then the target posture of the component is the action posture of the component at the last moment; In the process of changing the expression to the target expression, and then changing from the target expression to the initial expression, the target posture of the component is the action posture of the component at the intermediate moment.
  • different components adopt the same periodic function, so that different components have the same posture change range at the same time, and the posture changes of different components in the emoticon package are more harmonious.
  • a possible implementation of S404 includes: according to a periodic function, determining that the component is Further, in the case where the action posture includes pitch angle (pitch), yaw angle (yaw) and roll angle (roll), V ⁇ itch .
  • Xixu aw, Bai roil are the pitch angle, yaw angle and roll angle of the action attitude of component a2 at time r.
  • the expression of the component at multiple moments can be determined through the periodic function according to the number of image frames of the expression package and the frame rate of the expression package Weights.
  • the expression weight corresponding to the components under each frame of the emoticon map in the emoticon pack can be more accurately determined, thereby improving the accuracy of the action posture of the components in each frame of the label map sex.
  • the input data can be determined according to the image frame number of the expression package and the frame rate of the expression package, and the input data can be input into the periodic function to obtain the expression weights of the components at multiple moments.
  • the frame sequence of the emoticon corresponding to the moment in the emoticon packet can be determined, and the ratio of the frame sequence of the emoticon corresponding to the moment in the emoticon packet to the frame rate of the emoticon packet is determined as the moment
  • the periodic function is determined according to the duration of the emoticon pack, so as to improve the rationality and accuracy of the periodic function when used for emoticon pack generation.
  • the duration of the emoticon package can be determined as the period of the periodic function, or twice the duration of the emoticon package can be determined as the period of the periodic function.
  • the periodic function is a sine function.
  • the changing law of the function value of the sine function is similar to the changing law of the expression, Using the sine function to participate in determining the action posture of the component at multiple moments can improve the accuracy and fluency of the action posture of the component at multiple moments, thereby improving the accuracy and fluency of the virtual image expression in the emoticon package.
  • the maximum function value of the sine function corresponds to the target expression
  • the process of changing the function value from 0 to the maximum function value is equivalent to the process of changing the expression of the avatar from the initial expression to the target expression.
  • the process of changing from the maximum function value to 0 is equivalent to the process of changing the expression of the avatar from the target expression to the initial expression.
  • the weight of the expression of the component at multiple moments can be determined by combining the number of image frames of the expression package and the frame rate of the expression package.
  • the expression weight weight o of the component at time t can be obtained by the above formula.
  • the duration of the emoticon package is 1 second
  • the expression of the avatar in the emoticon package changes from the initial time to the target time, and From the target time to the initial time, the duration of the emoticon package is equivalent to half a cycle of the sine function, so the cycle of the sine function is 2 seconds.
  • the periodic function can be expressed as: weight
  • S405. Generate an emoticon package according to the material map, the global position, and the action postures of the components at multiple moments.
  • the material map of multiple components on the avatar is obtained, the global position of the component is determined according to the material map, the target posture of the component under the target expression is determined, and the component is determined at multiple times according to the target posture and the periodic function.
  • the action posture of the component according to the global position of the component and the action posture of the component at multiple moments, generates an emoticon package. Therefore, the feature that the change rule of the function value of the periodic function is close to the change rule of the dynamic expression of the expression is used to improve the accuracy and fluency of the action state of the component at multiple moments, thereby improving the quality of the produced expression package.
  • a possible implementation includes: According to The global position and the action posture of the components at multiple moments, through the driving algorithm, determine the position and shape of the material map on each frame image in the expression package, and obtain the expression package.
  • the driving algorithm is used to drive the material graph. Specifically, it is used to drive the material graph of the component to the corresponding position and corresponding shape according to the global position of the component and the action posture of the component, and then the driven The material images of constitute the emoticons in the emoticon package.
  • the component image can be obtained from the material map of the component, the component image is divided into a plurality of rectangular image areas, the vertices of each image area are obtained, and the depth value of each vertex is determined,
  • the depth values corresponding to different components can be preset, or the front and rear positional relationship of the material map can be determined based on the image identification (such as the image name) of the material map, and then the corresponding depth value can be determined.
  • the facial feature information can be determined according to the global positions of multiple components, and according to the multiple components Determine the rotation matrix of each material image for the action posture at multiple moments, and perform displacement transformation and rotation on the material image according to the facial feature information and the rotation matrix of the material image.
  • the facial feature information related to multiple key points can be determined based on the global positions of multiple key points (such as eyebrows, eyes, pupils, and mouth) on the material map of the component, so as to improve the accuracy of determining facial feature information. Stability, thereby improving the stability of expression.
  • facial feature information such as left/right eyebrow moving height, left/right eye opening height, mouth opening size, mouth width and so on.
  • the maximum deformation value of the multiple key points may be determined based on the fixed facial feature information.
  • the maximum deformation value of the key point of the face may include an upper limit value and a lower limit value of the key point movement.
  • the upper limit value of the eyes is the eigenvalue when the eyes are open
  • the lower limit value is the eigenvalue when the eyes are closed.
  • the corresponding feature value when the key point changes can be determined in the facial feature information of the key point, according to the corresponding feature value and key point when the key point changes
  • the maximum deformation value corresponding to the point determines the deformation value of the key point, that is, the displacement value of the key point, and drives the position change of the key point according to the displacement value of the key point and performs rendering to realize the deformation of the key point. And rotate the material image according to the rotation matrix of the material image. In this way, the driving of the material map of the component is completed, and the automatic generation of the emoticon package is realized.
  • the morphology can be used to fill in the image at this time, so as to improve the expression package generation effect. For example, using morphology to automatically generate images of the upper and lower eyelids, and images of the oral cavity.
  • the freeze-frame emoticon is the emoticon with the largest expression range of the avatar in the emoticon pack.
  • the production efficiency of dynamic emoticons and static freeze-frame emoticons is improved, the difficulty of production is reduced, and the experience of making emoticons for users is improved.
  • the target expression as "smile” as an example
  • the expression of the avatar in the emoticon package changes from not smiling to "smiling” and then from “smiling” to not smiling, and the "smiling" expression map of the avatar can be obtained from the emoticon graph.
  • FIG. 5 shows the facial expressions of the animation characters when they are expressionless (that is, the expression diagrams under the initial expression) and the animation characters in the "angry", “black line”, “smiling", “doubt”, “ Freeze emoticons of various target expressions such as “shy”, “surprised” and “wink”.
  • emoticons of anime characters under these target expressions can be generated, for example, emoticons ranging from expressionless to angry and then from angry to expressionless.
  • FIG. 6 is a structural block diagram of an emoticon pack generation device provided by an embodiment of the present disclosure. For ease of description, only parts related to the embodiments of the present disclosure are shown. Referring to FIG.
  • the emoticon package generation device includes: an acquisition unit 601, a position determination unit 602, a posture determination unit 603, and a generation unit 604.
  • the obtaining unit 601 is used to obtain the material map of multiple parts on the avatar;
  • the position determination unit 602 is used to determine the global position of the part according to the material map;
  • the pose determination unit 603 is used to determine the target pose of the part under the target expression ;
  • the generation unit 604 is used to generate an emoticon package according to the material map, the global position and the target posture, wherein, in the emoticon package, The expression of the avatar changes from the initial expression to the target expression.
  • the generation unit 604 in the process of generating emoticons according to the material map, the global position and the target pose, is specifically configured to: determine the action pose of the component at multiple moments according to the target pose and the periodic function; The material map, the global position, and the action postures of the parts at multiple moments generate an emoticon package, wherein the initial expression of the avatar in the emoticon package at the initial moment among the multiple moments is the initial expression.
  • the generating unit 604 in the process of determining the action posture of the component at multiple moments according to the target posture and the periodic function, is specifically configured to: determine the expression weight of the component at multiple moments according to the periodic function; According to the expression weights and target poses of the parts at multiple moments, the action poses of the parts at multiple moments are determined. In some embodiments, in the process of determining the expression weights of components at multiple moments according to the periodic function, the generation unit 604 is specifically configured to: according to the number of image frames of the emoticon package and the frame rate of the emoticon package, through the periodic function , to determine the expression weight of the part at multiple moments.
  • the emoticon pack generating device further includes: a function determining unit (not shown in the figure), configured to determine a periodic function according to the duration of the emoticon pack; wherein, the periodic function is a sine function.
  • the generation unit 604 in the process of generating emoticons according to the material map, the global position and the action posture of the component at multiple moments, the generation unit 604 is specifically configured to: According to the global position and the action posture of the component at multiple moments, Through the driving algorithm, the position and shape of the material map on each frame image in the expression package are determined, and the expression package is obtained.
  • the posture determination unit 603 is specifically configured to: determine the facial expression corresponding to the target facial expression according to the preset correspondence between multiple types of facial expressions and facial expressions, The expression action corresponding to the target expression includes the target gesture.
  • the position determining unit 602 is specifically configured to: determine the circumscribing matrix of the component in the material graph; determine the global position according to the circumscribing matrix.
  • FIG. 7 it shows a schematic structural diagram of an electronic device 700 suitable for implementing the embodiments of the present disclosure.
  • the electronic device 700 may be a terminal device or a server.
  • the terminal equipment may include but not limited to mobile phones, notebook computers, digital broadcast receivers, personal digital assistants (Personal Digital Assistant, PDA for short), tablet computers (Portable Android Device, PAD for short), portable multimedia players (Portable MediaPlayer, PMP for short), mobile terminals such as vehicle-mounted terminals (such as vehicle-mounted navigation terminals), and fixed terminals such as digital TVs and desktop computers.
  • PDA Personal Digital Assistant
  • PMP portable multimedia players
  • mobile terminals such as vehicle-mounted terminals (such as vehicle-mounted navigation terminals)
  • fixed terminals such as digital TVs and desktop computers.
  • the electronic device shown in FIG. 7 is only an example, and should not limit the functions and scope of use of the embodiments of the present disclosure. As shown in FIG.
  • an electronic device 700 may include a processing device (such as a central processing unit, a graphics processing unit, etc.) 701, which may be stored in a read-only memory (Read Only Memory, ROM for short) 702 or from a storage device 708 loads the programs in the random access memory (Random Access Memory, RAM for short) 703 to execute various appropriate actions and processes.
  • a processing device such as a central processing unit, a graphics processing unit, etc.
  • ROM Read Only Memory
  • RAM Random Access Memory
  • the input/output (Input/Output, I/O) interface 705 is also connected to the bus 704 o
  • the following devices can be connected to the I/O interface 705: including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer , an input device 706 such as a gyroscope; an output device 707 including a liquid crystal display (Liquid Crystal Display, LCD for short), a speaker, a vibrator, etc.; a storage device 708 including a magnetic tape, a hard disk, etc.; and a communication device 709 o communication device 709 may allow the electronic device 700 to communicate wirelessly or wiredly with other devices to exchange data. While FIG.
  • FIG. 7 shows electronic device 700 having various means, it should be understood that implementing or having all of the illustrated means is not a requirement. More or fewer means may alternatively be implemented or provided.
  • the processes described above with reference to the flowcharts can be implemented as computer software programs.
  • embodiments of the present disclosure include a computer program product, which includes a computer program carried on a computer-readable medium, where the computer program includes program codes for executing the methods shown in the flowcharts.
  • the computer program may be downloaded and installed from a network via communication means 709, or from storage means 708, or MROM 702.
  • the method in the embodiment of the present disclosure is limited to Mass transmission, including but not limited to: electric wire, optical cable, RF (Radio Frequency, Radio Frequency), etc., or any suitable combination of the above.
  • the above-mentioned computer-readable medium may be included in the above-mentioned electronic device, or may exist independently without being incorporated into the electronic device.
  • the above-mentioned computer-readable medium carries one or more programs, and when the above-mentioned one or more programs are executed by the electronic device, the electronic device is made to execute the methods shown in the above-mentioned embodiments.
  • Computer program code for carrying out the operations of the present disclosure may be written in one or more programming languages, or combinations thereof, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional Procedural programming language-such as "C" or similar programming language.
  • the program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server.
  • the remote computer may be connected to the user computer through any kind of network including a local area network (LAN) or a wide area network (WideAreaNetwork, WAN), or may be connected to an external computer (for example, using Internet service provider to connect via the Internet).
  • LAN local area network
  • WideAreaNetwork WideAreaNetwork
  • each block in a flowchart or block diagram may represent a module, program segment, or portion of code that contains one or more logical functions for implementing specified executable instructions.
  • the functions marked in the boxes may also Occurs in a different order than noted in the drawings. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or they may sometimes be executed in the reverse order, depending upon the functionality involved.
  • each block in the block diagrams and/or flowcharts, and combinations of blocks in the block diagrams and/or flowcharts can be implemented by a dedicated hardware-based system that performs specified functions or operations. , or may be implemented by a combination of special purpose hardware and computer instructions.
  • the units involved in the embodiments described in the present disclosure may be implemented by means of software or by means of hardware. Wherein, the name of the unit does not constitute a limitation on the unit itself under certain circumstances, for example, the first obtaining unit may also be described as "a unit that obtains at least two Internet Protocol addresses".
  • the functions described herein above may be performed at least in part by one or more hardware logic components.
  • exemplary types of hardware logic components include: Field Programmable Gate Arrays (Field Programmable Gate Arrays, FPGAs), Application Specific Integrated Circuits (ASICs), Application Specific Standard Products (Application Specific Standard Product, ASSP), System on Chip (System on Chip, SOC), Complex Programmable Logic Device (Complex Programmable Logic Device, CPLD) and so on.
  • a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device.
  • a machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium.
  • a machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any suitable combination of the foregoing. More specific examples of machine-readable storage media would include one or more wire-based electrical connections, portable computer disks, hard disks, Random Access Memory (RAM), Read Only Memory (ROM), Erasable Programmable Read Only Memory (EPROM or flash memory), optical fiber, compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.
  • RAM Random Access Memory
  • ROM Read Only Memory
  • EPROM Erasable Programmable Read Only Memory
  • CD-ROM compact disk read-only memory
  • a method for generating emoticons including: acquiring material maps of multiple components on the avatar; determining the global positions of the components according to the material maps ; determine the target posture of the component under the target expression; generate the expression package according to the material map, the global position and the target posture; wherein, in the expression package, the expression of the avatar Changing includes changing from an initial expression to said target expression.
  • generating the emoticon package according to the material map, the global position and the target pose includes: determining the The action posture of the component at multiple moments; the emoticon package is generated according to the material map, the global position, and the action posture of the component at multiple moments, wherein the avatar in the emoticon package is at the The expression at the initial moment among the multiple moments is the initial expression.
  • the determining the action posture of the component at multiple moments according to the target posture and the periodic function includes: determining the component at multiple times according to the periodic function the expression weight at multiple moments; and determine the action posture of the component at multiple moments according to the expression weights of the component at multiple moments and the target posture.
  • the determining the expression weight of the component at multiple moments according to the periodic function includes: according to the number of image frames of the expression package and the frames of the expression package The expression weight of the component at multiple moments is determined through the periodic function. According to one or more embodiments of the present disclosure, before determining the action posture of the component at multiple moments according to the target posture and the periodic function, it also includes: according to the duration of the emoticon package, determining the A periodic function; wherein, the periodic function is a sine function.
  • generating the emoticon package according to the material map, the global position, and the action gestures of the components at multiple moments includes: according to the global position and the The action poses of the above parts at multiple moments state, through a driving algorithm, determine the position and shape of the material map on each frame image in the emoticon package, and obtain the emoticon package.
  • the determining the target posture of the component under the target expression includes: determining the expression action corresponding to the target expression according to the preset correspondence between a plurality of expression types and expression actions , the expression action corresponding to the target expression includes the target gesture.
  • the determining the global position of the component according to the material graph includes: determining a circumscribing matrix of the component in the material graph; according to the circumscribing matrix, The global location is determined.
  • an emoticon package generation device including: an acquisition unit, configured to acquire material images of multiple components on the avatar; a position determination unit, configured to the material map, to determine the global position of the component; a pose determining unit, to determine the target pose of the component under the target expression; a generating unit, to determine the target pose according to the material map, the global position and the target pose , generating the expression package, wherein, in the expression package, the expression of the avatar changes from an initial expression to the target expression.
  • the generating unit in the process of generating the emoticon package according to the material map, the global position and the target pose, is specifically configured to: according to the The target posture and the periodic function determine the action posture of the component at multiple moments; generate the emoticon package according to the material map, the global position and the action posture of the component at multiple moments, wherein the The expression of the avatar in the expression package at the initial moment of the plurality of moments is the initial expression.
  • the generating unit in the process of determining the action posture of the component at multiple moments according to the target posture and the periodic function, is specifically configured to: according to the The periodic function determines the expression weight of the component at multiple moments; determines the action posture of the component at multiple moments according to the expression weight of the component at multiple moments and the target posture.
  • the generating unit in the process of determining the expression weights of the component at multiple moments according to the periodic function, is specifically configured to: according to the expression package The number of image frames and the frame rate of the expression package determine the weight of the expression of the component at multiple moments through the periodic function.
  • the emoticon package generation device further includes: a function determination unit, configured to determine the periodic function according to the duration of the emoticon package; wherein, the periodic function is a sine function.
  • the generating unit specifically It is configured to: determine the position and shape of the material map on each frame image in the emoticon package through a driving algorithm according to the global position and the action posture of the component at multiple moments, and obtain the emoticon package.
  • the posture determination unit in the process of determining the target posture of the component under the target expression, is specifically configured to: correspond to the preset correspondence between the multiple expression types and the expression actions Relationship, determine the expression action corresponding to the target expression, the expression action corresponding to the target expression includes the target gesture.
  • the position determining unit in the process of determining the global position of the component according to the material map, is specifically configured to: in the material map, determine the a circumscribing matrix of the component; and determining the global position according to the circumscribing matrix.
  • an electronic device including: at least one processor and a memory; the memory stores computer-executable instructions; the at least one processor executes the memory-stored The computer executes instructions, so that the at least one processor executes the above first aspect or various possible designs of the first aspect.
  • the emoticon package generation method or, execute the model training method described in the second aspect or various possible designs of the second aspect.
  • a computer-readable storage medium stores computer-executable instructions, and when a processor executes the computer-executable instructions, Realize the emoticon package generation method described in the above first aspect and various possible designs of the first aspect.
  • a computer program product is provided, the computer program product includes computer-executable instructions, and when a processor executes the computer-executable instructions, the first aspect and In the first aspect, various possible design methods for generating emoticons are described.
  • a computer program is provided.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Library & Information Science (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Processing Or Creating Images (AREA)
  • Image Generation (AREA)

Abstract

摘要本公开实施例提供一种表情包生成方法及设备，该方法包括：获取虚拟形象上多个部件的素材图；根据素材图，确定部件的全局位置；确定部件在目标表情下的目标姿态；根据素材图、全局位置和目标姿态，生成表情包；其中，在表情包中，虚拟形象的表情变化包括从初始表情变化至目标表情。从而，用户仅需要输入虚拟形象上多个部件的素材图，即可生成虚拟形象的动态表情包，提高了表情包制作效率，降低了表情包制作难度。

Description

表 情 包生 成方 法及 设 备 相关申请的交叉引用 本申请要求于 2022年 02月 16日提交中国专利局、 申请号为 202210141281.7、 申请名称 为 “表情包生成方法及设备”的中国专利申请的优先权，其全部内容通过引用结合在本文中。 技术领域 本公开实施例涉及计算机技术领域， 尤其涉及一种表情包生成方法及设备。 背景技术 以静态图像、动态图像等方式呈现的表情包形象生动、趣味性强， 深受用户喜爱， 除了在 聊天中使用表情包， 制作表情包也成为部分用户的喜好。 目前，表情包的制作需由专业画手通过绘画工具进行绘制。尤其针对动态的表情包，设计 者需要设计虚拟形象， 设计虚拟形象的运动、渐变、运动对齐等， 并逐帧绘制， 最终播放形成 动态的表情包。 整个制作过程耗费较多的时间和精力， 而且对绘画技术的要求较高。 因此， 如何降低表情包制作难度是目前亟需解决的问题。 发明内容 本公开实施例提供一种表情包生成方法及设备， 以克服表情包制作难度较高的问题。 第一方面， 本公开实施例提供一种表情包生成方法， 包括： 获取虚拟形象上多个部件的素材图； 根据所述素材图， 确定所述部件的全局位置； 确定所述部件在目标表情下的目标姿态； 根据所述素材图、 所述全局位置和所述目标姿态， 生成所述表情包； 其中，在所述表情包中，所述虚拟形象的表情变化包括从初始表情变化至所述目标表情。 第二方面， 本公开实施例提供一种表情包生成设备， 包括： 获取单元， 用于获取虚拟形象上多个部件的素材图； 位置确定单元， 用于根据所述素材图， 确定所述部件的全局位置； 姿态确定单元， 用于确定所述部件在目标表情下的目标姿态； 生成单元,用于根据所述素材图、所述全局位置和所述目标姿态，生成所述表情包，其中， 在所述表情包中， 所述虚拟形象的表情从初始表情变化至所述目标表情。 第三方面， 本公开实施例提供一种电子设备， 包括： 至少一个处理器和存储器； 所述存储器存储计算机执行指令； 所述至少一个处理器执行所述存储器存储的计算机执行指令， 使得所述至少一个处理器 执行如上第一方面所述的表情包生成方法。 第四方面，本公开实施例提供一种计算机可读存储介质，所述计算机可读存储介质中存储 有计算机执行指令，当处理器执行所述计算机执行指令时，实现如上第一方面所述的表情包生 成方法。 第五方面，根据本公开的一个或多个实施例，提供了一种计算机程序产品，所述计算机程 序产品包含计算机执行指令，当处理器执行所述计算机执行指令时，实现如上第一方面所述的 表情包生成方法。 第六方面，根据本公开的一个或多个实施例，提供了一种计算机程序， 当处理器执行所述 计算机程序时， 实现如上第一方面所述的表情包生成方法。 本公开实施例提供的表情包生成方法及设备，获取虚拟形象上多个部件的素材图，根据部 件的素材图确定部件的全局位置，确定部件在目标表情下的目标姿态，根据部件的素材图、全 局位置和目标姿态，生成表情包。其中，表情包中虚拟形象的表情变化包括从初始表情变化至 目标表情。 因此，用户仅需准备虚拟形象上部件的素材图， 无需进行虚拟形象的表情设计， 无 需关心多帧图像如何组合， 有效地提高了表情包制作效率并降低了制作难度。 附图说明 此处的附图被并入说明书中并构成本说明书的一部分， 示出了符合本申请的实施例， 并 与说明书一起用于解释本申请的原理。 图 1为本公开实施例提供的一种应用场景的示例图； 图 2为本公开实施例提供的表情包生成方法流程示意图一； 图 3a为多个部件的素材图的不例图； 图 3b为部件分类和部件命名的示例图； 图 4为本公开实施例提供的表情包生成方法的流程示意图二； 图 5为本公开实施例提供的动漫人物形象的表情图； 图 6为本公开实施例提供的模型确定设备的结构框图； 图 7为本公开实施例提供的电子设备的硬件结构示意图。 具体实施方式 为使本公开实施例的目的、 技术方案和优点更加清楚， 下面将结合本公开实施例中的附 图， 对本公开实施例中的技术方案进行清楚、完整地描述， 显然， 所描述的实施例是本公开一 部分实施例，而不是全部的实施例。基于本公开中的实施例，本领域普通技术人员在没有作出 创造性劳动前提下所获得的所有其他实施例， 都属于本公开保护的范围。 首先， 对公开实施例涉及的词语进行解释：
( 1 ) 虚拟形象： 在计算设备中通过图像描绘的虚拟角色， 比如动漫角色。
(2)虚拟形象的部件： 虚拟形象的组成部分； 比如， 动漫角色的眼睛、 鼻子、 嘴巴等均 属于动漫角色的部件。
(3 ) 部件的素材图： 绘制有该部件的图层； 其中， 不同部件可对应不同素材图， 即不同 部件可对应不同图层， 以提高部件组合的灵活性。
(4) 部件的全局位置： 部件在表情包中的表情图内的图像位置， 其中， 表情图上包括多 个部件组合得到的虚拟形象。
（5） 部件的姿态： 在表情包中， 虚拟形象的表情会发生变化， 该变化可以细化为部件的 姿态变化， 比如， 部件的倾斜程度、弯曲程度的变化， 所以部件的姿态可以包括部件的倾斜程 度、 弯曲程度、 拉伸程度等。 其次， 提供本公开实施例的构思： 在制作动态表情包的过程中，用户通常需要利用绘图工具绘制多帧画稿，再组合成动态表 情包。 这个过程耗费时间长， 且有着较高的技术门槛要求。 为解决上述问题，本公开的实施例提出了一种表情包生成方法及设备。在本公开的实施例 中，通过获取虚拟形象上多个部件的素材图，再确定多个部件的位置和姿态，基于多个部件的 素材图、位置和姿态来生成相应的动态表情包。在整个过程中，用户仅需准备虚拟形象上多个 部件的素材图，无需考虑每帧表情图的绘制，有效地降低了表情包的制作难度并提高了制作效 率 O 参考图 1 , 图 1为本公开实施例提供的一种应用场景的示例图。 如图 1所示，应用场景为动态表情包制作场景，在该场景下，用户可在终端 101上准备虚 拟形象上多个部件的素材图，终端 101基于该多个部件的素材图，制作动态的表情包，或者， 终端 101可将该多个部件的素材图发送至服务器 102,由服务器 102基于多个部件的素材图制 作动态的表情包。 示例性的，在聊天场景中，用户想要制作独特有趣的动态表情包，可在终端上点击进入聊 天应用程序提供的表情包制作页面， 在该表情包制作页面可以输入一些自己设计的卡通动物、 动漫人物等虚拟形象的多个部件的素材图，或者，也可以输入一些公开授权可用的卡通动物、 动漫人物等虚拟形象的多个部件的素材图， 通过表情包制作程序得到制作好的表情包。 下面，结合图 1所示的应用场景，描述本公开的实施例提供的表情包生成方法及设备。需 要注意的是，上述应用场景仅是为了便于理解本公开的精神和原理而示出，本公开的实施方式 在此方面不受任何限制。 相反， 本公开的实施方式可以应用于适用的任何场景。 需要说明的是，本公开的实施例可应用于电子设备， 电子设备可以为终端或者服务器。其 中， 终端可以是个人数字助理（personal digital assistant, 简称 PDA）设备、 具有无线通信功能 的手持设备（例如智能手机、 平板电脑）、 计算设备（例如个人电脑（personal computer, 简称 PC））、 车载设备、 可穿戴设备（例如智能手表、 智能手环）、 智能家居设备（例如智能显示设 备）等。其中，服务器可以是整体式服务器或是跨多计算机或计算机数据中心的分散式服务器。 服务器还可以是各种类别的， 例如但不限于， 网络服务器， 应用服务器， 或数据库服务器， 或 代理服务器。 参考图 2, 图 2为本公开实施例提供的表情包生成方法流程示意图一。如图 2所示， 该表 情包生成方法包括：
5201、 获取虚拟形象上多个部件的素材图。 本实施例中，可获取用户输入的多个部件的素材图，该多个部件属于同一虚拟形象。例如, 用户可通过表情包制作页面的输入控件输入多个部件的素材图；又如，可在表情包制作页面显 示多个虚拟形象的部件素材图， 用户可从中选取同一虚拟形象的多个部件的素材图。
5202、 根据素材图， 确定部件的全局位置。 本实施例中，不同部件对应的素材图的大小一致，针对各部件，部件在素材图中的位置即 部件的全局位置，因此，可通过确定部件在素材图中的位置，得到确定部件的全局位置。从而， 通过素材图大小一致、部件在素材图中的位置决定部件在表情图中的全局位置的方式，提高全 局位置的准确性。 除上述方式外， 可选的， 在部件对应的位置范围内随机确定部件的全局位置， 其中， 预先 为不同的部件设置不同的位置范围。
5203、 确定部件在目标表情下的目标姿态。 其中， 目标表情包括一种或多种表情， 例如为开心、 生气、 悲伤等。 本实施例中，可接收用户输入的目标表情，或者，可获取用户在一个或多个表情中选中的 目标表情，或者，确定目标表情为默认表情。在一个或多个表情下多个部件对应的动作姿态中， 确定目标表情下多个部件的动作姿态，为便于区分，将目标表情下多个部件的动作姿态称为多 个部件的目标姿态。 例如，获取用户的输入文本 “开心”，根据该输入文本确定目标表情为 “开心 ”这一表情， 在多个表情下多个部件的动作姿态中， 确定 “开心 ”这一目标表情下头部、脸部等部件的动作 姿态。
5204、根据素材图、全局位置和目标姿态， 生成表情包，表情包中虚拟形象的表情变化包 括从初始表情变化至目标表情。 其中， 初始表情是指表情包中在初始时刻下的表情， 即表情包中的第一帧图像或者 0时 刻下的图像所表现的表情。 本实施例中，部件的目标姿态是部件在目标表情下的姿态，而在表情包中虚拟形象的表情 会从初始表情变化至目标表情， 意味着表情包中各部件的姿态也是渐变的。所以，在确定多个 部件的全局位置和目标姿态之后，针对各部件，可基于部件的目标姿态，确定部件在多个时刻 下的姿态。接着，针对各个时刻， 基于多个部件的全局位置和多个部件在该时刻下的姿态， 对 多个部件的素材图进行组合， 得到该时刻下的表情图。如此， 得到多个时刻下的表情图。通过 组合多个时刻下的表情图， 得到表情包。 可选的，表情包中虚拟形象的表情变化还包括从目标表情变化至初始表情，即表情包中虚 拟形象的表情变化为从初始表情变化至目标表情，再从目标表情变化至初始表情。例如，虚拟 形象从不笑到微笑， 再从微笑到不笑。 本公开实施例中，基于虚拟形象上多个部件的素材图，确定部件的全局位置，基于部件的 全局位置和目标表情下部件的目标姿态，得到表情包，还可以得到目标表情对应的表情图。可 见，用户仅需输入部件的素材图， 即可得到质量较高的表情包，有效地提高了表情包和表情图 的制作效率， 降低了表情包和表情图的制作难度， 提高了用户体验。 下面， 在图 2提供的实施例的基础上， 提供多个可行的扩展实施例。
( 1 ) 关于虚拟形象 在一些实施例中，虚拟形象包括虚拟人物形象，尤其是动漫人物形象。相较于其他类型的 表情包， 动漫人物形象的表情包的制作难度较高， 通常需要通过 2D图像绘制出 3D的动态效 果。本实施例中，用户通过输入动漫人物形象上多个部件的素材图， 即可得到动漫人物形象的 动态表情包，而且在表情包的制作过程中，考虑到了各个部件的位置以及姿态， 既提高了动漫 人物形象的动态表情包的制作效率， 降低了制作难度， 还确保了表情包的质量。
(2) 关于部件 在一些实施例中， 预先设置有必要部件和非必要部件。 其中，必要部件是制作虚拟形象的表情包所必需的部件，非必要部件是制作虚拟形象的表 情包的可选部件。用户在输入多个部件的素材图时必须输入所有必要部件的素材图，以确保表 情包中虚拟形象的完整性。 此时， S201 的一种可能的实现方式包括： 获取虚拟形象上多个必要部件的素材图。 具体 的，可预先告知用户制作表情包的必要部件， 比如，在表情制作页面上显示必要部件的名称， 又如，在部件对应的输入控件周围标记该部件是否为必要部件；用户在制作表情包时必须输入 这些必要部件的素材图。 从而，通过将部件区分为必要部件和非必要部件，提高表情包制作的成功率和表情包的制 作效果。当然， 用户也可以在输入必要部件的素材图之余， 输入非必要部件的素材图， 以进一 步完善和丰富虚拟形象。 可选的，在虚拟形象为动漫人物形象的情况下，必要部件可包括眉毛部件、上睫毛部件、 瞳孔部件、嘴巴部件和脸部部件。其中，通过这些部件可以准确描绘出动漫人物形象的外貌， 还可以生动地表达出多种情绪， 有利于确保虚拟形象的完整性并提高虚拟形象的表情生动性。 可选的， 非必要部件可包括如下至少一种： 前景部件、头发部件、头部装饰部件、 下睫毛 部件、 眼白部件、 鼻子部件、 耳朵部件、 身体部件、 背景部件。 从而， 通过这些非必要部件使 得虚拟形象更具细节。 其中， 前景部件是指根据空间关系位于虚拟形象前面的部件。 在一些实施例中，预先设置多个部件类别。在获取虚拟形象上多个部件的素材图之前，可 显示多个部件类别。从而， 便于用户按照部件类别来输入部件的素材图。其中， 部件类别可以 分为多个层次的类别， 部件类别分为两个层次时， 部件类别可分为父类和父类下的子类。 可选的，父类包括如下至少一种：前景部件、头发部件、头部部件、身体部件、背景部件。 头发部件下的子类包括如下至少一种： 头部装饰部件、前发部件、耳前发部件、耳后发部件、 后发部件；头部部件下的子类包括头部装饰部件、眉毛部件、眼睛部件、鼻子部件、嘴巴部件、 脸部部件、 耳朵部件。 进一步的，子类还可以进一步划分为不同类别。具体的， 眼睛部件下的子类可包括如下至 少一种： 上睫毛部件、 下睫毛部件、 瞳孔部件、 眼白部件。 作为示例的， 如图 3a所示， 图 3a为多个部件的素材图的示例图。 在图 3a中， 给出了动 漫人物形象的眉毛部件、上睫毛部件、 瞳孔部件、 嘴巴部件、脸部部件、 身体部件这些部件分 别对应的素材图， 可以看出， 这些素材图大小一致， 通过这些部件的素材图组合拼接， 可以得 到对应的动漫人物形象。
(3 ) 关于素材图 在一些实施例中，一个部件可对应一个或多个素材图。比如，虚拟形象有多个头部装饰部 件， 所以头部装饰部件可对应多个素材图。 在一些实施例中，素材图对应唯一的图像标识， 即不同的素材图对应不同的图像标识。从 而，在根据部件的素材图生成表情包的过程中，可通过图像标识来区分素材图以及区分素材图 对应的部件。 可选的， 图像标识包括图像名称。 比如， 前景部件对应的多个素材图的图像名称分别为前景一、前景二、 ; 头发装饰部 件对应的多个素材图的图像名称分别为头发装饰部件一、 头发装饰部件二、 ……， 等等。 作为示例的， 如图 3b所示， 图 3b为部件分类和部件命名的示例图， 其中， 左侧区域显示 了多个部件，右侧区域显示了多个部件类型下的素材图的命名方式，“图层 ”是指素材图，“png” 为素材图的图像格式。 从图 3b可以看出： 1 ) “前景” 可以对应多个图层， 图像命名可以为前景 _1、 前景 _2等; 2) “头发装饰”可以对应多个图层， 图像命名可以为头发装饰 _1、 头发装饰 _2等； 3 ) “前发” 可以对应多个图层， 图像命名可以为前发 _1、 前发 _2等； 4) “耳前发” 可以对应多个图层， 图像命名可以为耳前发 _1、 耳前发 _2等； 5) “后发” 可以对应多个图层， 图像命名可以为后 发 _1、 后发 _2等； 6) “头部装饰” 可以对应多个图层， 图像命名可以为头部装饰 _1、 头部装 饰 _2等； 7) “眉毛”可以对应多个图层， 多个图层可以合并为一个 png, 即多个素材图可以合 并为一个素材图， 图像命名可以为眉毛 _1 ; ……等等。 如此， 可为不同类别的部件下的素材图 提供不同的命名， 为同一类别的部件下的不同素材图提供不同的命名。 在此不一一描述。
(4) 关于全局位置的确定 在一些实施例中， S202 的一种可能的实现方式包括： 在部件的素材图中， 确定部件的外 接矩阵；根据部件的外接矩阵，确定部件的全局位置。从而，通过求解部件的外接矩阵的方式， 提高部件的全局位置的准确性。 本实现方式中，可在部件的素材图中识别出部件的外接矩阵，得到部件的外接矩阵在素材 图中的位置。其中，外接矩阵在素材图中的位置包括外接矩阵的四个顶点在素材图中的像素点 坐标。接着， 由于所有部件的素材图大小一致， 部件在素材图中的图像位置反映部件的全局位 置， 所以， 可确定部件的全局位置为部件的外接矩阵的位置。 可选的， 部件的素材图的图像通道包括位置通道。 其中， 在素材图中， 像素点在位置通道的通道值反映出该像素点是否位于部件的图案区 域。 比如： 如果像素点在位置通道的通道值为 1 , 则确定该像素点位于图案区域； 如果像素点 在位置通道的通道值为 0, 则确定该像素点不位于图案区域。 所以， 可通过素材图中多个像素 点的位置通道的取值， 确定出部件在素材图中的外接矩阵， 提高了外接矩阵的准确性。 进一步的，部件的素材图为 RGBA四通道图像，即部件的素材图的图像通道包括 R通道、 G 通道、 B通道和 A通道。 其中， R通道、 G通道、 B通道分别为图像的红、 绿、 蓝三种颜色 通道， A通道为图像的位置通道。 因此， 可在部件的素材图中获取各像素点在 A通道的通道值， 根据各像素点的 A通道的 通道值， 确定部件的外接矩阵。 例如， 确定 A通道的通道值为 1 的所有像素点， 确定包含这 些像素点的外接矩阵为部件的外接矩阵。 进一步 的， 部件的外接矩阵还可为部件的最小外接矩形 (minimum bounding rectangle, MBR), 以提高部件全局位置的准确性。
(5) 关于部件姿态的确定 在一些实施例中， S203 的一种可能的实现方式包括： 根据多个表情类型与表情动作的预 设对应关系，确定目标表情对应的表情动作， 目标表情对应的表情动作包括多个部件在目标表 情下的目标姿态。 其中，根据多个表情类型与表情动作的预设对应关系可以由专业人员预先设置， 以降低表 情包制作难度。 在该预设对应关系中，不同的表情类型对应不同的表情动作，表情动作中包括多个预设部 件的动作姿态。 在不同的表情类型下， 预设部件可以相同也可以不同。 例如， “开心”这一表 情类型下包括眉毛部件、上睫毛部件、 瞳孔部件、嘴巴部件、脸部部件这些部件的动作姿态， 眉毛部件、 上睫毛部件和嘴巴部件的动作姿态均可为弯曲上扬的状态； “疑问”这一表情下除 了上述 “开心 ”表情类型下的那些部件之外， 还可以包括表情符号部件（比如 “问号”）， 嘴巴 部件的动作姿态可呈现为笔直或者嘴角向下。 本实现方式中， 可在多个表情类型中， 确定目标表情所属的目标表情类型。接着， 多个表 情类型与表情动作的预设对应关系，确定目标表情类型对应的表情动作，即目标表情对应的表 情动作，从目标表情对应的表情动作中，查找到虚拟形象上多个部件的动作姿态， 即得到多个 部件的目标姿态。 在一些实施例中，部件的动作姿态包括部件的姿态角度，姿态角度可包括如下至少一种： 俯仰角、偏航角、翻滚角， 从而， 可通过结合部件的位置和部件的姿态角度， 准确地表现出表 情包中虚拟形象的表情。 在部件的动作姿态包括部件的姿态角度的情况下，可选的，初始表情为虚拟形象上的部件 的姿态角度为 0时的表情。 参考图 4, 图 4为本公开实施例提供的表情包生成方法的流程示意图二。如图 4所示， 表 情包生成方法包括：
5401、 获取虚拟形象上多个部件的素材图。
5402、 根据素材图， 确定部件的全局位置。
5403、 确定部件在目标表情下的目标姿态。 其中， S401-S403的实现原理和技术效果可参照前述实施例， 不再赘述。
5404、 根据目标姿态和周期性函数， 确定部件在多个时刻的动作姿态。 其中， 部件的动作姿态可以参考前述实施例， 不再赘述。 本实施例中，由于表情包中的多帧表情图呈现了虚拟形象在多个时刻下的表情，这多个时 刻下的表情包括初始表情和目标表情，还包括初始表情与目标表情之间的变化表情。所以，为 体现该多个时刻下的表情， 需要确定多个部件分别在多个时刻下的动作姿态。 表情包中虚拟形象的表情从初始表情变化至目标表情的过程是部件的动作姿态逐渐增大 至目标姿态的过程，该过程是缓慢且非线性的。在虚拟形象的表情变化过程还包括从目标表情 变化至初始表情的过程时，该变化过程是周期性的。因此，为更准确地贴合表情动态变化功率, 本实施例采用周期性函数来确定部件在多个时刻的动作姿态。 本实施例中，针对各部件，通过周期性函数对目标姿态进行处理，得到部件在周期性函数 的周期内的多个时刻下的动作姿态。 其中，若表情包中虚拟形象的表情变化过程仅包括从初始表情变化至目标表情的过程，则 部件的目标姿态为部件在最后时刻的动作姿态； 若表情包中虚拟形象的变化过程包括从初始 表情变化至目标表情、再从目标表情变化至初始表情的过程，则部件的目标姿态为部件在中间 时刻的动作姿态。 在一些实施例中，不同部件采用同一周期性函数，使得不同部件在同一时刻下姿态的变化 幅度一致， 表情包中不同部件的姿态变化更为和谐。 在一些实施例中， S404的一种可能的实现方式包括： 根据周期性函数， 确定部件在多个
Figure imgf000010_0001
进一步的， 在动作姿态包括俯仰角 （pitch）、 偏航角 （yaw） 和翻滚角 （roll） 的情况下,
Figure imgf000010_0002
V^itch.席顼 aw、 Bai roil分别为部件 a2在时刻 r时刻的动作姿态中的俯仰角、 偏航角和翻滚 角。 可选的，在根据周期性函数确定部件在多个时刻的表情权重的过程中，可根据表情包的图 像帧数和表情包的帧率， 通过周期性函数， 确定部件在多个时刻的表情权重。从而， 通过在周 期性函数中结合表情包的图像帧数和帧率， 更准确地确定表情包中每帧表情图下部件对应的 表情权重， 进而提高每帧标签图中部件的动作姿态的准确性。 本可选方式中，可根据表情包的图像帧数和表情包的帧率，确定输入数据，将输入数据输 入至周期性函数中， 得到部件在多个时刻的表情权重。 进一步的，针对各个时刻，可确定该时刻对应的表情图在表情包中的帧序，将该时刻对应 的表情图在表情包中的帧序与表情包的帧率的比值，确定为该时刻对应的输入数据，将该时刻 对应的输入数据输入至周期性函数中， 得到部件在该时刻的表情权重。 可选的，根据表情包的时长确定周期性函数，以提高周期性函数用于表情包生成时的合理 性和准确性。 本可选方式中，可将表情包的时长确定为周期性函数的周期，也可以将表情包的时长的两 倍确定为周期性函数的周期， 具体要根据周期性函数的函数值变化范围进行确定。 可选的，周期性函数为正弦函数。由于正弦函数的函数值变化规律与表情变化规律相似， 采用正弦函数来参与确定部件在多个时刻下的动作姿态， 可以提高部件在多个时刻下的动作 姿态的准确性和流畅性， 进而提高表情包中虚拟形象表情的准确性和流畅性。 其中，周期性函数为正弦函数时，正弦函数的最大函数值与目标表情对应， 函数值从 0变 化至最大函数值的过程相当于虚拟形象的表情从初始表情变化至目标表情的过程， 函数值从 最大函数值变化至 0的过程相当于虚拟形象的表情从目标表情变化至初始表情的过程。 进一步的， 在周期性函数为正弦函数的情况下， 周期性函数可表示为： f(x) = sin(wx) 其中， T = 2ir/|w| , T为周期， x为正弦函数的输入， w为参数。 在上述公式的基础上，可结合表情包的图像帧数和表情包的帧率，确定部件在多个时刻的 表情权重， 此时， 周期性函数可表示为： weight = sin (w — ) 其中， fps为表情包的帧率， 1表示第 1帧图像。 假设第 1帧图像对应 t时刻， 则可以通过 上述公式求得部件在 t时刻的表情权重 weight o 假设表情包的时长为 1 秒， 表情包中虚拟形象的表情从初始时刻变化至目标时刻， 又从 目标时刻变化至初始时刻，则表情包的时长相当于正弦函数的半个周期，所以正弦函数的周期 为 2秒。 此时， 周期性函数可表示为: weight
Figure imgf000011_0001
S405、 根据素材图、 全局位置和部件在多个时刻的动作姿态， 生成表情包。 其中， S405的实现原理和技术效果可参照前述实施例， 不再赘述。 本公开实施例中，获取虚拟形象上多个部件的素材图，根据素材图确定部件的全局位置， 确定部件在目标表情下的目标姿态，根据目标姿态和周期性函数，确定部件在多个时刻的动作 姿态， 根据部件的全局位置和部件在多个时刻的动作姿态， 生成表情包。从而， 利用周期性函 数的函数值变化规律与表情动态变化规律接近的特点， 提高部件在多个时刻的动作状态的准 确性和流畅性， 进而提高制作的表情包质量。
(6) 关于表情包生成 在一些实施例中，在根据部件的素材图、部件的全局位置和部件在多个时刻的动作姿态， 生成表情包的过程中， 一种可能的实现方式包括： 根据全局位置和部件在多个时刻的动作姿 态， 通过驱动算法， 确定素材图在表情包中各帧图像上的位置和形状， 得到表情包。 本实施例中，驱动算法用于对素材图进行驱动，具体的，用于根据部件的全局位置和部件 的动作姿态，将部件的素材图驱动至相应的位置和相应的形状，进而由驱动后的素材图构成表 情包中的表情图。 可选的， 在驱动算法中， 针对各部件， 可从部件的素材图中获得部件图像， 将部件图像划 分为多个矩形的图像区域，得到各个图像区域的顶点，确定各个顶点的深度值，使得部件图像 在视觉上呈现类似 3维的效果， 使得表情包中虚拟形象更为立体， 提高表情包生成效果。 其中，可以预设不同部件对应的深度值，也可以基于素材图的图像标识 (比如图像名称)， 确定素材图的前后位置关系， 进而确定相应深度值。 可选的， 在驱动算法中， 可根据多个部件的全局位置， 确定脸部特征信息， 根据多个部件 在多个时刻的动作姿态， 确定各个素材图的旋转矩阵， 根据脸部特征信息和素材图的旋转矩 阵， 对素材图进行位移变换和旋转。 其中，可基于部件的素材图上的多个关键点（比如眉毛、眼睛、瞳孔、嘴巴）的全局位置， 来确定与多个关键点相关的脸部特征信息，以提高定脸部特征信息的稳定性，进而提高表情的 稳定性。 其中， 脸部特征信息例如左 /右眉毛移动高度、 左 /右眼张开高度、 嘴巴张开大小、 嘴 巴宽度等。 本可选方式中，在得到与多个关键点相关的定脸部特征信息之后，可基于定脸部特征信息 确定多个关键点的最大形变值。其中，脸部关键点的最大形变值可包括关键点运动的上极限值 和下极限值。比如，眼睛的上极限值为眼睛睁开时的特征值，下极限值为眼睛闭上时的特征值。 本可选方式中， 针对各关键点， 可以在关键点的脸部特征信息中， 确定关键点变化时（比 如眼睛上下眨动） 对应的特征值， 根据关键点变化时对应的特征值和关键点对应的最大形变 值，确定关键点的形变值，也即关键点的位移值，根据关键点的位移值来驱动关键点的位置变 化并进行绘制渲染，实现对关键点的形变。并根据素材图的旋转矩阵对素材图进行旋转。如此, 完成对部件的素材图的驱动， 实现表情包的自动生成。 可选的，在驱动过程中，考虑到部件在形变时会产生空白或者间隙，此时可以利用形态学 进行图像填补， 以提高表情包生成效果。 比如， 利用形态学自动生成上下眼皮的图像、 口腔的 图像。 通过上述各实施例， 既可以得到表情包，也可以得到表情包中虚拟形象的各帧表情图，尤 其是可以得到虚拟形象的定格表情图，即虚拟形象的表情为目标表情的表情图。由于表情包中 虚拟形象的表情是从初始表情变化至目标表情再从目标表情变化至初始表情，所以，定格表情 图是表情包中虚拟形象表情幅度最大的表情图。从而，提高了动态的表情包和静态的定格表情 图的制作效率， 降低了制作难度， 提高了用户的表情包制作体验。 以目标表情为 “微笑 ”为例， 表情包中虚拟形象的表情从不笑到 “微笑 ”再从 “微笑 ”到 不笑， 可以从表情图中获取到虚拟形象的 “微笑” 的表情图。 作为示例的， 图 5给出了动漫人物形象无表情时的表情图（即初始表情下的表情图）以及 动漫人物形象在 “生气”、 “黑线”、 “微笑”、 “疑问”、 “害羞”、 “惊讶”、 “眨眼”多种目标表情 下的定格表情图。其中，基于本公开任一实施例提供的表情包生成方法，可生成动漫人物形象 在这些目标表情下的表情包， 例如， 从无表情到生气再从生气到无表情的表情包。 在图 5中， 除了动漫形象的五官之外，动漫人物形象的眼镜、动漫人物形象上的表情符号 （如 “生气 ”对应的定格表情图中的生气符号， 又如 “疑问 ”对应的定格表情图中的问号， 又 如 “眨眼 ”对应的定格表情图中的星星）也属于动漫人物形象上的部件， 同样可以通过上述实 施例来确定位置和姿态。 对应于上文实施例的表情包生成方法， 图 6为本公开实施例提供的表情包生成设备的结 构框图。为了便于说明， 仅示出了与本公开实施例相关的部分。参照图 6, 表情包生成设备包 括： 获取单元 601、 位置确定单元 602、 姿态确定单元 603和生成单元 604。 获取单元 601 , 用于获取虚拟形象上多个部件的素材图； 位置确定单元 602, 用于根据素材图， 确定部件的全局位置； 姿态确定单元 603 , 用于确定部件在目标表情下的目标姿态； 生成单元 604,用于根据素材图、全局位置和目标姿态，生成表情包，其中，在表情包中， 虚拟形象的表情从初始表情变化至目标表情。 在一些实施例中， 在根据素材图、全局位置和目标姿态， 生成表情包的过程中， 生成单元 604具体用于： 根据目标姿态和周期性函数， 确定部件在多个时刻的动作姿态； 根据素材图、 全局位置和部件在多个时刻的动作姿态，生成表情包，其中，表情包中虚拟形象在多个时刻中 的初始时刻下的表情为初始表情。 在一些实施例中，在根据目标姿态和周期性函数，确定部件在多个时刻的动作姿态的过程 中， 生成单元 604具体用于： 根据周期性函数， 确定部件在多个时刻的表情权重； 根据部件在 多个时刻的表情权重和目标姿态， 确定部件在多个时刻的动作姿态。 在一些实施例中，在根据周期性函数，确定部件在多个时刻的表情权重的过程中，生成单 元 604具体用于：根据表情包的图像帧数和表情包的帧率，通过周期性函数，确定部件在多个 时刻的表情权重。 在一些实施例中， 表情包生成设备还包括： 函数确定单元（图中未示出）， 用于根据表情 包的时长， 确定周期性函数； 其中， 周期性函数为正弦函数。 在一些实施例中，在根据素材图、全局位置和部件在多个时刻的动作姿态，生成表情包的 过程中， 生成单元 604具体用于： 根据全局位置和部件在多个时刻的动作姿态， 通过驱动算 法， 确定素材图在表情包中各帧图像上的位置和形状， 得到表情包。 在一些实施例中， 在确定部件在目标表情下的目标姿态的过程中， 姿态确定单元 603 具 体用于：根据多个表情类型与表情动作的预设对应关系，确定目标表情对应的表情动作， 目标 表情对应的表情动作包括目标姿态。 在一些实施例中，在根据素材图，确定部件的全局位置的过程中，位置确定单元 602具体 用于： 在素材图中， 确定部件的外接矩阵； 根据外接矩阵， 确定全局位置。 本实施例提供的表情包生成设备， 可用于执行上述表情包生成方法的实施例的技术方案, 其实现原理和技术效果类似， 本实施例此处不再赘述。 参考图 7, 其示出了适于用来实现本公开实施例的电子设备 700的结构示意图，该电子设 备 700可以为终端设备或服务器。其中，终端设备可以包括但不限于诸如移动电话、笔记本电 脑、数字广播接收器、个人数字助理（Personal Digital Assistant,简称 PDA）、平板电脑（Portable Android Device, 简称 PAD）、 便携式多媒体播放器（Portable MediaPlayer, 简称 PMP）、 车载 终端 （例如车载导航终端） 等等的移动终端以及诸如数字 TV、 台式计算机等等的固定终端。 图 7示出的电子设备仅仅是一个示例,不应对本公开实施例的功能和使用范围带来任何限制。 如图 7所示， 电子设备 700可以包括处理装置 （例如中央处理器、 图形处理器等） 701 , 其可以根据存储在只读存储器 （Read Only Memory， 简称 ROM） 702中的程序或者从存储装 置 708加载到随机访问存储器 （Random Access Memory, 简称 RAM） 703中的程序而执行各 种适当的动作和处理。 在 RAM 703中， 还存储有电子设备 700操作所需的各种程序和数据。 处理装置 701、 ROM 702以及 RAM 703通过总线 704彼此相连。 输入 /输出 （Input/Output, I/O） 接口 705也连接至总线 704 o 通常， 以下装置可以连接至 I/O接口 705： 包括例如触摸屏、 触摸板、 键盘、 鼠标、 摄像 头、麦克风、加速度计、陀螺仪等的输入装置 706；包括例如液晶显示器（Liquid Crystal Display , 简称 LCD）、 扬声器、 振动器等的输出装置 707； 包括例如磁带、 硬盘等的存储装置 708； 以 及通信装置 709 o通信装置 709可以允许电子设备 700与其他设备进行无线或有线通信以交换 数据。 虽然图 7示出了具有各种装置的电子设备 700, 但是应理解的是， 并不要求实施或具备 所有示出的装置。 可以替代地实施或具备更多或更少的装置。 特别地， 根据本公开的实施例， 上文参考流程图描述的过程可以被实现为计算机软件程 序。例如，本公开的实施例包括一种计算机程序产品，其包括承载在计算机可读介质上的计算 机程序，该计算机程序包含用于执行流程图所示的方法的程序代码。在这样的实施例中，该计 算机程序可以通过通信装置 709从网络上被下载和安装， 或者从存储装置 708被安装， 或者 M ROM 702被安装。 在该计算机程序被处理装置 701执行时， 执行本公开实施例的方法中限
Figure imgf000014_0001
质传输， 包括但不限于： 电线、 光缆、 RF（射频， Radio Frequency）等等， 或者上述的任意合 适的组合。 上述计算机可读介质可以是上述电子设备中所包含的；也可以是单独存在，而未装配入该 电子设备中。 上述计算机可读介质承载有一个或者多个程序， 当上述一个或者多个程序被该电子设备 执行时， 使得该电子设备执行上述实施例所示的方法。 可以以一种或多种程序设计语言或其组合来编写用于执行本公开的操作的计算机程序代 码, 上述程序设计语言包括面向对象的程序设计语言一诸如 Java、 Smalltalk, C++, 还包括常 规的过程式程序设计语言一诸如 “C”语言或类似的程序设计语言。程序代码可以完全地在用 户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算 机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的 情形中，远程计算机可以通过任意种类的网络 包括局域网（ Local Area Network ,简称 LAN） 或广域网 （WideAreaNetwork, 简称 WAN） 一连接到用户计算机， 或者， 可以连接到外部计 算机 （例如利用因特网服务提供商来通过因特网连接）。 附图中的流程图和框图，图示了按照本公开各种实施例的系统、方法和计算机程序产品的 可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、 程序段、或代码的一部分， 该模块、程序段、或代码的一部分包含一个或多个用于实现规定的 逻辑功能的可执行指令。也应当注意，在有些作为替换的实现中，方框中所标注的功能也可以 以不同于附图中所标注的顺序发生。 例如， 两个接连地表示的方框实际上可以基本并行地执 行， 它们有时也可以按相反的顺序执行， 这依所涉及的功能而定。 也要注意的是， 框图和 /或 流程图中的每个方框、 以及框图和 /或流程图中的方框的组合， 可以用执行规定的功能或操作 的专用的基于硬件的系统来实现， 或者可以用专用硬件与计算机指令的组合来实现。 描述于本公开实施例中所涉及到的单元可以通过软件的方式实现， 也可以通过硬件的方 式来实现。其中， 单元的名称在某种情况下并不构成对该单元本身的限定， 例如， 第一获取单 元还可以被描述为 “获取至少两个网际协议地址的单元”。 本文中以上描述的功能可以至少部分地由一个或多个硬件逻辑部件来执行。例如，非限制 性地，可以使用的示范类型的硬件逻辑部件包括：现场可编程门阵列 (Field Programmable Gate Array, FPGA)、 专用集成电路 (Application Specific Integrated Circuit, ASIC)、 专用标准产品 (Application Specific Standard Product, ASSP)、 片上系统 (System on Chip, SOC)、 复杂可编 程逻辑设备 (Complex Programmable Logic Device, CPLD) 等等。 在本公开的上下文中，机器可读介质可以是有形的介质，其可以包含或存储以供指令执行 系统、装置或设备使用或与指令执行系统、装置或设备结合地使用的程序。机器可读介质可以 是机器可读信号介质或机器可读储存介质。机器可读介质可以包括但不限于电子的、磁性的、 光学的、 电磁的、 红外的、 或半导体系统、 装置或设备， 或者上述内容的任何合适组合。机器 可读存储介质的更具体示例会包括基于一个或多个线的电气连接、便携式计算机盘、硬盘、随 机存取存储器 (RAM)、 只读存储器 (ROM)、 可擦除可编程只读存储器 (EPROM或快闪存 储器)、光纤、便捷式紧凑盘只读存储器 (CD-ROM)、光学储存设备、 磁储存设备、 或上述内 容的任何合适组合。 第一方面， 根据本公开的一个或多个实施例， 提供了一种表情包生成方法， 包括： 获取虚 拟形象上多个部件的素材图；根据所述素材图，确定所述部件的全局位置；确定所述部件在目 标表情下的目标姿态； 根据所述素材图、所述全局位置和所述目标姿态， 生成所述表情包； 其 中， 在所述表情包中， 所述虚拟形象的表情变化包括从初始表情变化至所述目标表情。 根据本公开的一个或多个实施例，所述根据所述素材图、所述全局位置和所述目标姿态， 生成所述表情包，包括：根据所述目标姿态和周期性函数，确定所述部件在多个时刻的动作姿 态； 根据所述素材图、所述全局位置和所述部件在多个时刻的动作姿态， 生成所述表情包， 其 中， 所述表情包中所述虚拟形象在所述多个时刻中的初始时刻下的表情为所述初始表情。 根据本公开的一个或多个实施例，所述根据所述目标姿态和周期性函数，确定所述部件在 多个时刻的动作姿态， 包括： 根据所述周期性函数， 确定所述部件在多个时刻的表情权重； 根 据所述部件在多个时刻的表情权重和所述目标姿态， 确定所述部件在多个时刻的动作姿态。 根据本公开的一个或多个实施例，所述根据所述周期性函数，确定所述部件在多个时刻的 表情权重， 包括： 根据所述表情包的图像帧数和所述表情包的帧率， 通过所述周期性函数， 确 定所述部件在多个时刻的表情权重。 根据本公开的一个或多个实施例，所述根据所述目标姿态和周期性函数，确定所述部件在 多个时刻的动作姿态之前， 还包括： 根据所述表情包的时长， 确定所述周期性函数； 其中， 所 述周期性函数为正弦函数。 根据本公开的一个或多个实施例，所述根据所述素材图、所述全局位置和所述部件在多个 时刻的动作姿态，生成所述表情包，包括：根据所述全局位置和所述部件在多个时刻的动作姿 态，通过驱动算法，确定所述素材图在所述表情包中各帧图像上的位置和形状，得到所述表情 包。 根据本公开的一个或多个实施例，所述确定所述部件在目标表情下的目标姿态，包括：根 据多个表情类型与表情动作的预设对应关系，确定所述目标表情对应的表情动作，所述目标表 情对应的表情动作包括所述目标姿态。 根据本公开的一个或多个实施例，所述根据所述素材图，确定所述部件的全局位置，包括: 在所述素材图中， 确定所述部件的外接矩阵； 根据所述外接矩阵， 确定所述全局位置。 第二方面， 根据本公开的一个或多个实施例， 提供了一种表情包生成设备， 包括： 获取单 元， 用于获取虚拟形象上多个部件的素材图； 位置确定单元， 用于根据所述素材图， 确定所述 部件的全局位置； 姿态确定单元， 用于确定所述部件在目标表情下的目标姿态； 生成单元， 用 于根据所述素材图、所述全局位置和所述目标姿态，生成所述表情包，其中，在所述表情包中， 所述虚拟形象的表情从初始表情变化至所述目标表情。 根据本公开的一个或多个实施例， 在所述根据所述素材图、 所述全局位置和所述目标姿 态， 生成所述表情包的过程中， 所述生成单元具体用于： 根据所述目标姿态和周期性函数， 确 定所述部件在多个时刻的动作姿态；根据所述素材图、所述全局位置和所述部件在多个时刻的 动作姿态，生成所述表情包，其中，所述表情包中所述虚拟形象在所述多个时刻中的初始时刻 下的表情为所述初始表情。 根据本公开的一个或多个实施例，在所述根据所述目标姿态和周期性函数，确定所述部件 在多个时刻的动作姿态的过程中，所述生成单元具体用于：根据所述周期性函数，确定所述部 件在多个时刻的表情权重；根据所述部件在多个时刻的表情权重和所述目标姿态，确定所述部 件在多个时刻的动作姿态。 根据本公开的一个或多个实施例，在所述根据所述周期性函数，确定所述部件在多个时刻 的表情权重的过程中，所述生成单元具体用于：根据所述表情包的图像帧数和所述表情包的帧 率， 通过所述周期性函数， 确定所述部件在多个时刻的表情权重。 根据本公开的一个或多个实施例，所述表情包生成设备还包括： 函数确定单元，用于根据 所述表情包的时长， 确定所述周期性函数； 其中， 所述周期性函数为正弦函数。 根据本公开的一个或多个实施例，在所述根据所述素材图、所述全局位置和所述部件在多 个时刻的动作姿态，生成所述表情包的过程中，所述生成单元具体用于：根据所述全局位置和 所述部件在多个时刻的动作姿态，通过驱动算法，确定所述素材图在所述表情包中各帧图像上 的位置和形状， 得到所述表情包。 根据本公开的一个或多个实施例， 在所述确定所述部件在目标表情下的目标姿态的过程 中，所述姿态确定单元具体用于：根据多个表情类型与表情动作的预设对应关系，确定所述目 标表情对应的表情动作， 所述目标表情对应的表情动作包括所述目标姿态。 根据本公开的一个或多个实施例，在所述根据所述素材图，确定所述部件的全局位置的过 程中， 所述位置确定单元具体用于： 在所述素材图中， 确定所述部件的外接矩阵； 根据所述外 接矩阵， 确定所述全局位置。 第三方面， 根据本公开的一个或多个实施例， 提供了一种电子设备， 包括： 至少一个处理 器和存储器；所述存储器存储计算机执行指令；所述至少一个处理器执行所述存储器存储的计 算机执行指令， 使得所述至少一个处理器执行如上第一方面或第一方面各种可能的设计所述 的表情包生成方法，或者，执行如第二方面或第二方面各种可能的设计所述的模型训练方法。 第四方面，根据本公开的一个或多个实施例，提供了一种计算机可读存储介质，所述计算 机可读存储介质中存储有计算机执行指令，当处理器执行所述计算机执行指令时，实现如上第 一方面以及第一方面各种可能的设计所述的表情包生成方法。 第五方面，根据本公开的一个或多个实施例，提供了一种计算机程序产品，所述计算机程 序产品包含计算机执行指令，当处理器执行所述计算机执行指令时，实现如第一方面以及第一 方面各种可能的设计所述的表情包生成方法。 第六方面，根据本公开的一个或多个实施例，提供了一种计算机程序， 当处理器执行所述 计算机程序时， 实现如第一方面以及第一方面各种可能的设计所述的表情包生成方法。 以上描述仅为本公开的较佳实施例以及对所运用技术原理的说明。 本领域技术人员应当 理解，本公开中所涉及的公开范围，并不限于上述技术特征的特定组合而成的技术方案， 同时 也应涵盖在不脱离上述公开构思的情况下， 由上述技术特征或其等同特征进行任意组合而形 成的其它技术方案。例如上述特征与本公开中公开的（但不限于）具有类似功能的技术特征进 行互相替换而形成的技术方案。 此外，虽然采用特定次序描绘了各操作，但是这不应当理解为要求这些操作以所示出的特 定次序或以顺序次序执行来执行。在一定环境下， 多任务和并行处理可能是有利的。同样地， 虽然在上面论述中包含了若干具体实现细节， 但是这些不应当被解释为对本公开的范围的限 制。在单独的实施例的上下文中描述的某些特征还可以组合地实现在单个实施例中。相反地， 在单个实施例的上下文中描述的各种特征也可以单独地或以任何合适的子组合的方式实现在 多个实施例中。 尽管巳经采用特定于结构特征和 /或方法逻辑动作的语言描述了本主题， 但是应当理解所 附权利要求书中所限定的主题未必局限于上面描述的特定特征或动作。相反，上面所描述的特 定特征和动作仅仅是实现权利要求书的示例形式。

Claims

权 利 要 求 书
1. •种表情包生成方法， 包括： 获取虚拟形象上多个部件的素材图； 根据所述素材图， 确定所述部件的全局位置； 确定所述部件在目标表情下的目标姿态； 根据所述素材图、 所述全局位置和所述目标姿态， 生成所述表情包； 其中，在所述表情包中，所述虚拟形象的表情变化包括从初始表情变化至所述日标表情。
2. 根据权利要求 1所述的表情包生成方法， 其中所述根据所述素材图、 所述全局位置利 所述目标姿态， 牛 .成所述表情包， 包括： 根据所述日标姿态和周期性函数， 确定所述部件在多个时刻的动作姿态； 根据所述素材图、所述全局位置和所述部件在多个时刻的动作姿态， 生成所述表情包， 其 中， 所述表情包中所述虚拟形象花所述多个时刻中的初始时刻卜的表情为所述初始表情。
3. 根据权利要求 2所述的表情包生成方法， 其中所述根据所述目标姿态和周期性函数， 确定所述部件在多个时刻的动作姿态， 包括： 根据所述周期性函数， 确定所述部件在多个时刻的表情权重； 根据所述部件在多个时刻的表情权重和所述目标姿态， 确定所述部件在多个时刻的动作 姿态。
4. 根据权利要求 3所述的表情包生成方法， 其中所述根据所述周期性函数， 确定所述部 件在多个时刻的表情权重， 包括： 根据所述表情包的图像帧数和所述表情包的帧率，通过所述周期性函数，确定所述部件在 多个时刻的表情权更。
5. 根据权利要求 2至 4任一项所述的表情包生成方法， 其中所述根据所述目标姿态和周 期性函数， 确定所述部件在多个时刻的动作姿态之前， 还包括： 根据所述表情包的时长， 确定所述周期性函数； 其中， 所述周期性函数为正弦函数。
6. 根据权利要求 2至 5任一项所述的表情包生成方法， 其中所述根据所述素材图、 所述 全局位置和所述部件在多个时刻的动作姿态， 生成所述表情包， 包括： 根据所述全局位置和所述部件在多个时刻的动作姿态，通过驱动算法，确定所述素材图在 所述表情包中各帧图像上的位置和形状， 得到所述表情包。
7. 根据权利要求 1至 6任一项所述的表情包生成方法， 其中所述确定所述部件在目标表 情下的目标姿态， 包括： 根据多个表情类型与表情动作的预设对应关系，确定所述目标表情对应的表情动作，所述 目标表情对应的表情动作包括所述目标姿态。
8. 根据权利要求 1至 7任一项所述的表情包生成方法， 其中所述根据所述素材图， 确定 所述部件的全局位置， 包括： 在所述素材图中， 确定所述部件的外接矩阵； 根据所述外接矩阵， 确定所述全局位置。
9. —种表情包生成设备， 包括： 获取单元， ffl于获取虚拟形象上多个部件的素材图；
16
SUBSTITUTE SHEET RULE 26 位置确定单元， 用于根据所述素材图， 确定所述部件的全局位置； 姿态确定单元， 用于确定所述部件在目标表情下的目标姿态； 生成单元,用于根据所述素材图、所述全局位置和所述目标姿态，生成所述表情包，其中， 在所述表情包中， 所述虚拟形象的表情从初始表情变化至所述目标表情。
10. 一种电子设备， 包括： 至少一个处理器和存储器； 所述存储器存储计算机执行指令； 所述至少一个处理器执行所述存储器存储的计算机执行指令， 使得所述至少一个处理器 执行如权利要求 1至 8任一项所述的表情包生成方法。
11. 一种计算机可读存储介质， 所述计算机可读存储介质中存储有计算机执行指令， 当处 理器执行所述计算机执行指令时， 实现如权利要求 1至 8任一项所述的表情包生成方法。
12. 一种计算机程序产品， 所述计算机程序产品包含计算机执行指令， 当处理器执行所述 计算机执行指令时， 实现如权利要求 1至 8任一项所述的表情包生成方法。
13. 一种计算机程序， 当处理器执行所述计算机程序时， 实现如权利要求 1至 8任一项所 述的表情包生成方法。
17
SUBSTITUTE SHEET RULE 26
PCT/SG2023/050062 2022-02-16 2023-02-06 表情包生成方法及设备 WO2023158370A2 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210141281.7A CN116645449A (zh) 2022-02-16 2022-02-16 表情包生成方法及设备
CN202210141281.7 2022-02-16

Publications (2)

Publication Number Publication Date
WO2023158370A2 true WO2023158370A2 (zh) 2023-08-24
WO2023158370A3 WO2023158370A3 (zh) 2023-11-09

Family

ID=87579176

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/SG2023/050062 WO2023158370A2 (zh) 2022-02-16 2023-02-06 表情包生成方法及设备

Country Status (2)

Country Link
CN (1) CN116645449A (zh)
WO (1) WO2023158370A2 (zh)

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2001216525A (ja) * 2000-02-04 2001-08-10 Sharp Corp 画像処理装置
JP2002157605A (ja) * 2000-11-21 2002-05-31 Sharp Corp 画像処理装置、画像処理方法、及び画像処理用プログラムを記録した記録媒体
CN1256702C (zh) * 2003-12-31 2006-05-17 马堃 数字图像合成的方法
JP2007286669A (ja) * 2006-04-12 2007-11-01 Sony Corp 画像処理装置および方法、並びにプログラム
CN102270352B (zh) * 2010-06-02 2016-12-07 腾讯科技（深圳）有限公司 动画播放的方法和装置
WO2017029488A2 (en) * 2015-08-14 2017-02-23 Metail Limited Methods of generating personalized 3d head models or 3d body models

Also Published As

Publication number Publication date
CN116645449A (zh) 2023-08-25
WO2023158370A3 (zh) 2023-11-09

Similar Documents

Publication Publication Date Title
US11961189B2 (en) Providing 3D data for messages in a messaging system
US11776233B2 (en) Beautification techniques for 3D data in a messaging system
US11790621B2 (en) Procedurally generating augmented reality content generators
KR102624635B1 (ko) 메시징 시스템에서의 3d 데이터 생성
US11410401B2 (en) Beautification techniques for 3D data in a messaging system
EP3992919B1 (en) Three-dimensional facial model generation method and apparatus, device, and medium
US11825065B2 (en) Effects for 3D data in a messaging system
US20240095981A1 (en) Video generation method and apparatus, device and readable storage medium
US20230360184A1 (en) Image processing method and apparatus, and electronic device and computer-readable storage medium
KR20210113948A (ko) 가상 아바타 생성 방법 및 장치
CN110148191A (zh) 视频虚拟表情生成方法、装置及计算机可读存储介质
CN114445562A (zh) 三维重建方法及装置、电子设备和存储介质
WO2022147451A1 (en) Media content items with haptic feedback augmentations
WO2023092950A1 (zh) 虚拟场景的素材处理方法及装置、电子设备、存储介质和计算机程序产品
JP7467780B2 (ja) 画像処理方法、装置、デバイス及び媒体
WO2024051540A1 (zh) 特效处理方法、装置、电子设备及存储介质
WO2023158370A2 (zh) 表情包生成方法及设备
WO2021155666A1 (zh) 用于生成图像的方法和装置
WO2023158375A2 (zh) 表情包生成方法及设备
CN115714888B (zh) 视频生成方法、装置、设备与计算机可读存储介质
CN118864677A (zh) 角色动画生成方法、装置及电子设备

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE