Method for generating pixel-level three-dimensional character action material

Abstract

The invention provides a pixel-level three-dimensional character action material generation method. The generation method comprises the following steps: constructing a target person model and a contour stroke thereof by utilizing a pre-constructed person model component library and related materials in a Unreal engine; and endowing the target person model with a normal mapping material. And based on a virtual camera, recording image, action and normal information of the target person model in real time when the action blueprint is executed into two NewCanvasRenderTarget 2D dynamic pictures. And carrying out pixelated processing on the dynamic pictures, carrying out cooperative playing of the two dynamic pictures by combining parallel light in a second environment so as to generate real-time illumination interaction, independently exporting a target person model in each frame in real time, and storing the target person model as a corresponding pixel-level three-dimensional person action material. Through the technical scheme provided by the invention, the pixel character action model with stereoscopic impression can be manufactured at low cost, and the image, action and light and shadow interaction effect of the target character model can be conveniently adjusted in the manufacturing process and subsequent players.

Description

Technical Field

[0001] This application relates to the field of production of pixel virtual character materials in pixel games, and specifically relates to a method for generating pixel-level three-dimensional character action materials. Background Art

[0002] With the improvement and development of hardware conditions, games are now increasingly developing towards the 3A level. Today, although pixel games are simple, creating game products with simple pixel graphics and conveying the creativity of the author is a unique artistic enjoyment, and relevant game works still have a considerable number of users / enthusiasts.

[0003] For pixel-level games, the creation of character action materials is one of the main tasks, which largely determines the quality of the later game works and the user experience. Currently, pixel-level character action materials are mainly constructed in the following way: Set the camera and lighting in software such as Maya or Max, then give the character model a predetermined animation, and render a sequence of frame images of the complete set of actions, turning off the anti-aliasing option during rendering. Then use software engines such as Maya, Max, Unreal, or Unity to perform pixelization post-processing on the rendered entire screen or picture.

[0004] In the above method, the modification environment of the character shape, character actions, etc. is all in Maya or Max. Since the sequence of frames is rendered in Maya or Max software by presetting the camera perspective and lighting, the pre-designed character clothing styles, facial fatness, hair styles, and action performances are rendered into pictures. Therefore, after the rendered screen or picture is imported into the Unreal or Unity engine, players cannot change the clothing combination, hair style, and action performance at any time. Moreover, traditional 2D Spine animations cannot show the details of the 3D actions of the character model. For example, during the turning process of the character model, the relative angle between the body and the camera is changing, and the corresponding three-dimensional information cannot be shown in the 2D Spine animation.

[0005] In addition, in this method, when performing pixelization processing on the rendered screen using the Unreal or Unity engine, all elements in the entire screen (including characters, environment, special effects, sky, etc.) need to be processed simultaneously. However, in the actual production process, often only the character needs to be pixelized (the environment or special effects need other processing performances). Obviously, performing pixelization post-processing on all elements of the entire screen consumes a very large amount of GPU and is not very efficient.

[0006] It is worth mentioning that since each frame needs to be rendered during the entire process of producing sequence frame materials in Maya or Max, once there are any modifications, it is necessary to return to the software to remake them and then re-import them into the Unreal or Unity engine. The production cycle of making sequence frame materials of human character action models in Maya or Max and then importing them into the rendering engine is also much longer than directly capturing 3D human character models in the engine, and the real-time performance is much worse.

[0007] Term Explanation:

[0008] Unreal Engine: Unreal is the abbreviation of UNREAL ENGINE. Developed by Epic, it is one of the world's most well-known and widely licensed game engines.

[0009] Unity Engine: Unity is a real-time 3D interactive content creation and operation platform. All creators, including those in game development, art, architecture, automotive design, and film and television, can turn their creativity into reality with the help of Unity. The Unity platform provides a complete set of software solutions for creating, operating, and monetizing any real-time interactive 2D and 3D content, and supports platforms including mobile phones, tablets, PCs, game consoles, augmented reality, and virtual reality devices.

[0010] Spine Animation: It is a 2D bone animation editing tool developed for priority. Only one picture is needed. In the Spine software, by binding the picture to the bones and then controlling the bones to achieve animation.

[0011] RenderTarget: A RenderTarget is a texture that can be read and written at runtime. It allows us to store information in the RGBA channels of a TextureObject at runtime and read and use it when needed. Summary of the Invention

[0012] In view of the above-mentioned deficiencies in the current production of pixel-level human action models, the present invention provides a method for generating pixel-level three-dimensional human action materials. Through the said generation method, the art R & D personnel of the game can adjust the pixel human image and actions in real time during the production process of pixel-level human action materials, and can conveniently produce three-dimensional human action materials.

[0013] The technical solution provided by the present invention is implemented as a method for generating pixel-level three-dimensional human action materials. The said generation method includes the following steps:

[0014] S1. Produce a library of human model components for users to combine and import them into the Unreal engine, and build a target human model based on the said library of human model components;

[0015] S2. Use the model vertex normal expansion and double-sided material to create the outer contour of the target character model;

[0016] S3. Create a new camera for capturing the target character model's movements (RenderCapture) and place the target character model in the camera's field of view.

[0017] S4. Create a normal map material for baking the vertex normal material of the target character model and assign it to the target character model so that the target character model generates corresponding dynamic light and shadow changes when the angle of the first ambient light changes;

[0018] S5 uses the animation blueprint to drive the target character model to make the corresponding action, and the camera captures the action image in real time, as well as the normal information of each vertex of the target character model and stores it in two new NewCanvasRenderTarget2D dynamic images;

[0019] S6. Based on the texture UV coordinate values of the captured action image, some pixel values are removed to create a self-luminous material, and assigned to each frame of the action image in the dynamic picture;

[0020] S7. Utilize the RenderTarget component of the Unreal engine to obtain in real time the real-time normal map of the target character model in each frame of the action image during the collaborative playback of the two dynamic images, and perform DOTProduct processing with the parallel light in the second environment to generate real-time lighting interaction with the environment, and independently display the target character model in each frame from the surface and save it as the corresponding pixel-level three-dimensional character action material.

[0021] Furthermore, in step S3, the target character model is placed in the shooting field of view of the camera, which is implemented as follows: setting the camera to an orthogonal camera, or setting the focal length of the camera to 45-50, and ensuring that the target character model is located in a safe area in the camera field of view when executing the animation blueprint; rotating the vertices of the target character model to a coordinate system facing the camera capture direction through the matrix calculation in the material.

[0022] Furthermore, in step S5, the camera captures the action image in real time, and the normal information of each vertex on the target character model is stored in two newly created NewCanvasRenderTarget2D dynamic pictures, which is implemented as follows: giving the target character model an animation cartoon plane material to record the image and action of the target character model in real time and save it in the first newly created NewCanvasRenderTarget2D dynamic picture; giving the target character model a corresponding normal material to capture the normal information map of the character's action in different time periods in real time and save it in the second newly created NewCanvasRenderTarget2D dynamic picture.

[0023] Furthermore, in step S6, some pixel values are eliminated based on the texture UV coordinate values of the captured action image, which is achieved by amplifying the texture UV coordinate values of the action image by a first multiple, converting them into integer data, and then reducing them by the first multiple to perform pixel elimination.

[0024] Furthermore, the production method further includes: setting the self-luminous material in step S6 to a size obtained by dividing the material by the number of pixel grids, which is consistent with the real-time RenderTarget size of the camera.

[0025] Furthermore, in step S7, the target character model in each frame is independently displayed from the surface and saved, which is achieved by: during the collaborative playback of the two dynamic pictures, the contour channel information is connected to the opacitymask channel, so that the target character model is independently displayed from the surface and saved.

[0026] The method for generating pixel-level three-dimensional character action materials provided by the present invention performs all processing, except for the character model components, in the UNreal engine. In UNreal, combined with specific processing steps (and corresponding data structures / components), not only can production personnel easily and in real time modify or adjust pixel-level character action materials when producing them or when applying them to games, but also, compared to pixel-level character action models produced by existing methods, the pixel-level character action models produced by the generation method provided by the present invention have a three-dimensional sense and good light and shadow interaction with the ambient light system. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without creative work.

[0028] Figure 1 In one embodiment, the present invention provides a flowchart of a method for generating a pixel-level three-dimensional human action model. Specific embodiments

[0029] The following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Usually, the components of the embodiments of the present application described and shown in the accompanying drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application claimed, but only represents the selected embodiments of the present application. All other embodiments obtained by those skilled in the art based on the embodiments of the present application without creative efforts belong to the scope of protection of the present application.

[0030] It should be noted that similar reference numerals and letters indicate similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. At the same time, in the description of the present application, terms such as "first" and "second" are only used for distinguishing descriptions and cannot be understood as indicating or implying relative importance.

[0031] As Figure 1 shown, in one embodiment, the method for generating pixel-level three-dimensional human action materials provided by the present invention includes the following steps:

[0032] S1. Produce a library of human model components for users to combine and import them into the UNreal engine, and construct a target human model based on the library of human model components. In order to make the generated human characters more diverse, in this step, existing drawing software can be used to produce human model components such as hair, face, coat, pants, shoes, etc. that allow players or human action model producers to freely combine; then these human model components are imported into the Unreal engine for subsequent use.

[0033] S2. Use model vertex normal expansion and double-sided materials to create an outline stroke for the target human model. The reason for this step is mainly that later it is necessary to separately extract the target human model from the playing dynamic picture based on the outline of the target human model.

[0034] S3. Create a new camera for capturing the actions of the target human model (RenderCapture), and place the target human model within the shooting field of view of this camera.

[0035] Furthermore, the target character model is placed in the shooting field of view of the camera, which is achieved by: setting the camera to an orthogonal camera, or setting the focal length of the camera to 45-50, and ensuring that the target character model is always located in the safe area of the camera field of view when executing the animation blueprint; rotating the vertices of the target character model to the coordinate system facing the camera capture direction through the matrix calculation in the material (the default coordinate direction of the object in the Unreal engine is the Z axis facing upward, and general pixel games use a horizontal movement perspective).

[0036] S4. Create a normal map material for baking the target character model's vertex normals and assign it to the target character model. This allows the target character model to dynamically change lighting and shadows when the first ambient lighting angle changes. This step primarily captures the normal map of the target character model as it performs an action, ensuring a three-dimensional effect when subsequently creating the target character's action model.

[0037] S5. Use the animation blueprint to drive the target character model to perform corresponding actions. The camera captures the action images in real time, as well as the normal information of each vertex on the target character model, and stores them in two newly created NewCanvasRenderTarget2D dynamic images. The dynamic images can store information such as the target character's image (color, outline, depth, etc.), actions, and normal map captured by the camera in real time, and provide players with the ability to independently modify the target character model when the game engine renders the relevant game.

[0038] Preferably, the camera captures the action image in real time, and the normal information of each vertex on the target character model is stored in two newly created NewCanvasRenderTarget2D dynamic pictures, which is implemented as follows: giving the target character model an animation cartoon plane material to record the image and action of the target character model in real time and save it in the first newly created NewCanvasRenderTarget2D dynamic picture; giving the target character model a corresponding normal material to capture the normal information map of the character's action in different time periods in real time and save it in the second newly created NewCanvasRenderTarget2D dynamic picture.

[0039] S6. Based on the texture UV coordinate values of the captured motion image, some pixel values are culled to create a self-luminous material, which is then applied to each frame of the motion image in the dynamic image. Pixel culling primarily removes some gradient pixel values, which is a common pixel processing technique and will not be further described here.

[0040] S7. Use the RenderTarget component of the Unreal engine to obtain the real-time normal map of the target character model in each frame of the action image during the collaborative playback of the two dynamic images in real time, and perform DOTProduct processing with the parallel light in the second environment to generate real-time lighting interaction with the environment. Display the target character model in each frame independently from the patch and save it as corresponding pixel-level three-dimensional character action materials. Preferably, the above real-time lighting interaction is generated based on the lambert lighting model.

[0041] Further, the realization of displaying and saving the target character model in each frame independently from the patch is as follows: during the collaborative playback of the two dynamic images, by connecting the contour channel information to the opacity mask channel, the target character model therein is displayed and saved independently from the patch.

[0042] As described above, the above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for generating pixel-level stereoscopic human action materials, characterized in that, The generation method comprises the following steps: S1. Create a character model component library for user combinations and import it into the UNreal engine. Based on the character model component library, build a target character model; S2. Use the model vertex normal expansion and double-sided material to create the outer contour of the target character model; S3. Create a new camera for capturing the target character model's movements (RenderCapture) and place the target character model in the camera's field of view. S4. Create a normal map material for baking the vertex normal material of the target character model and assign it to the target character model so that the target character model generates corresponding dynamic light and shadow changes when the angle of the first ambient light changes; S5 uses the animation blueprint to drive the target character model to make the corresponding action, and the camera captures the action image in real time, as well as the normal information of each vertex of the target character model and stores it in two new NewCanvasRenderTarget2D dynamic images; S6 removes some pixel values based on the texture UV coordinate values of the captured motion image, creates a self-luminous material, and assigns it to each frame of the motion image in the dynamic picture; S7. Utilize the RenderTarget component of the Unreal engine to obtain in real time the real-time normal map of the target character model in each frame of the action image during the collaborative playback of the two dynamic images, and perform DOTProduct processing with the parallel light in the second environment to generate real-time lighting interaction with the environment, and independently display the target character model in each frame from the surface and save it as the corresponding pixel-level three-dimensional character action material.

2. The generation method according to claim 1, characterized in that, In step S3, the target character model is placed in the shooting field of view of the camera, which is achieved by setting the camera to an orthographic camera or setting the focal length of the camera to 45-50, and ensuring that the target character model is located in a safe area in the field of view of the camera when executing the animation blueprint; The vertices of the target character model are rotated to the coordinate system facing the direction of the camera capture through the matrix calculation in the material.

3. The generation method according to claim 1, wherein In step S5, the camera captures the action image in real time, and the normal information of each vertex on the target character model is stored in two newly created NewCanvasRenderTarget2D dynamic pictures, which is implemented as follows: giving the target character model an animation cartoon plane material to record the image and action of the target character model in real time and save it in the first newly created NewCanvasRenderTarget2D dynamic picture; giving the target character model a corresponding normal material to capture the normal information map of the character's action in different time periods in real time and save it in the second newly created NewCanvasRenderTarget2D dynamic picture.

4. The generation method according to claim 3, wherein In step S6, some pixel values are eliminated based on the texture UV coordinate values of the captured motion image, which is achieved by magnifying the texture UV coordinate values of the motion image by a first multiple, converting them into integer data, and then reducing them by the first multiple to perform pixel elimination.

5. The generation method according to claim 4, wherein The generation method further includes: setting the self-luminous material in step S6 to a size obtained by dividing the material by the number of pixel grids, which is consistent with the real-time RenderTarget size of the camera.

6. The generation method according to claim 5, wherein The realization of independently displaying the target character model in each frame from the patch and saving it as corresponding pixel-level three-dimensional character action materials is as follows: during the collaborative playback of the two dynamic images, by connecting the contour channel information to the opacity mask channel, the target character model is independently displayed from the patch and saved as corresponding pixel-level three-dimensional character action materials.

Images