KR20160004467A - Mobile advertising engine and 3d deformable object modeling method thereof - Google Patents

Abstract

According to an embodiment of the present invention, a real-time 3D deformable object modeling method of a mobile advertising engine is provided. The real-time 3D deformable object modeling method of the mobile advertising engine comprises the steps of: forming a first border surrounding the entire appearance of a deformable object; forming a plurality of second borders surrounding a portion of the deformable object; and conducting a first collision test on the second borders of the first border when a collision occurs on the first border, and performing a triangle node-based detailed collision test on the second borders when a collision occurs in the second borders. Therefore, a mobile 3D advertising engine system that may operate with a low performance processor and a real-time 3D deformable object modeling method of the mobile 3D advertising engine system may be provided.

Description

[0001] MOBILE ADVERTISING ENGINE AND 3D DEFORMABLE OBJECT MODELING METHOD THEREOF [0002]

A mobile advertisement engine and a method for modeling a real time three-dimensional deformed object of the mobile advertisement engine. More particularly, the present invention relates to a mobile advertisement engine that can be implemented as a low-performance processor, and a method for modeling a real-time three-dimensional deformed object of the mobile advertisement engine.

Recently, smart devices are rapidly evolving to shake the world, and the market size is also rapidly expanding. According to the Korea Institute of Information and Communications (KIIS) 's August 2013 statistics, the penetration rate of smart devices in Korea is more than 61% of Korea' s households and it is announced that the number of subscribers has already exceeded 35 million in July 2013. In addition, as of 1H03, the domestic smart contents market is estimated to have increased by about 22% from KRW 1.9 trillion in 2012 to KRW2.2 trillion by 2012, and it is expected to reach KRW 3.5 trillion by 2015.

In addition, smart device hardware technology is developing rapidly from Apple's 1-core CPU in 2009 to recently 4-core CPU from Samsung. The development of this hardware technology has been rapidly shifting to the smart device environment where the high specification contents based on the existing PC and console platform are relatively limited in terms of time and space in combination with the expansion of the mobile game and application market.

In response to such environmental changes, there is a growing interest in real-time representation of 3D deformed objects applicable to various applications. However, although the performance of smart devices has been greatly improved, it still has limited performance compared to general PC, so there is a limit to apply the latest PC based 3D deformation object simulation methods. Therefore, it is necessary to study the modeling and simulation method of 3D deformed object considering the performance of the latest smart device.

A mobile advertisement engine that can be implemented as a low-performance processor, and a real-time three-dimensional deformed object modeling method of the mobile advertisement engine.

An embodiment includes a first boundary forming unit forming a first boundary that surrounds an entire outer shape of a deformed object; A second boundary forming unit forming a plurality of second boundaries surrounding a part of the deformed object; And performing a first collision check on the second boundary of the specific first boundary when a collision occurs at a specific first boundary and if a collision occurs at a specific second boundary, And a collision checking performing unit for performing a triangle-node-based detailed collision check.

The collision checking unit may find collision-susceptible nodes by applying a spatial hashing method to the specific second boundary, and may perform the collision inspection based on the triangle-node only for the nodes.

The first and second boundaries may have a spherical shape.

Meanwhile, the embodiment includes a step of forming a first step boundary to cover the entire outer shape of the deformed object; Forming a plurality of two-step boundaries surrounding a portion of the deformed object; And performing a first collision check on the second boundary of the first boundary when a collision occurs at a particular first boundary, and if a collision occurs at a particular second boundary, And performing node-based detailed collision checking on the node-based detailed collision model.

The step of performing the collision checking includes: detecting nodes that are likely to collide by applying a spatial hashing method to a specific second boundary, and performing the collision check based on the triangle-node only for the nodes .

The first and second boundaries may have a spherical shape.

According to the mobile 3D advertisement system of the embodiment, a mobile 3D advertisement engine system capable of operating with a low-performance processor and a real-time three-dimensional deformed object modeling method of the mobile advertisement engine can be implemented.

1A to 1C are views illustrating a collision detection method in which an empty space is brought into close contact with a surface of a 3D deformed object by applying an AABB to process a quick collision,
2 is a block diagram of a processor for producing a 3D advertisement,
FIG. 3 is a flowchart showing a two-step collision detection method considering morphological characteristics of a deformed object,
FIG. 4 is a view showing an example of FIG. 2,
FIGS. 5A and 5B are views illustrating a positional change of a two-level boundary according to the movement of an object;
6 is a view simulating a free fall situation of various objects.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily carry out the present invention. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. In order to clearly illustrate the present invention, parts not related to the description are omitted, and similar parts are denoted by like reference characters throughout the specification.

Throughout the specification, when a part is referred to as being "connected" to another part, it includes not only "directly connected" but also "electrically connected" with another part in between .

Throughout the specification, when an element is referred to as "comprising ", it means that it can include other elements as well, without excluding other elements unless specifically stated otherwise. Also, the terms " part, "" module," and " module ", etc. in the specification mean a unit for processing at least one function or operation and may be implemented by hardware or software or a combination of hardware and software have.

Hereinafter, preferred embodiments according to the present invention will be described in detail with reference to the accompanying drawings.

Existing physics-based dynamic simulations have resulted in many technological advances in terms of stability and speed due to the continuous development of computer simulation related technologies. Especially, the study on the deformation object modeling method for expressing the deformation accurately when the external / internal force is applied to the 3D deformation object has been steadily progressed. In this regard, the finite element method is relatively accurate compared to other methods of predicting deformed objects, but it takes a long time to calculate, so it is mainly used for designing buildings and predicting car crashes that require accurate prediction regardless of time Has come.

Compared to this, the mass-spring method, which is relatively fast but poorly accurate, is mainly applied to real-time applications. In addition, although the accuracy is slightly lower, the boundary element method and the finite volume method, which complement the merits and demerits of the two methods, have been applied for simulation modeling of 3D deformed objects and have a physical basis The Meshless deformation method, which is fast but fast, has also been widely used.

Numerical integration methods for simulating 3D deformed objects have been applied implicit integration instead of explicit comparisons, which are relatively fast computations, in order to stabilize the system in most cases.

Existing 3D deformed object simulations have been used to reproduce physical phenomena more realistically in computer games and animations, and have been used for simulation of clothes, 3D collision objects, which covers most of human bodies in advertisements and movies, Surgical simulations have been used in a variety of fields, such as practicing doctors practicing surgery as in the actual situation, training in airplane or automobile training, training in the treatment of claustrophobia or claustrophobia.

However, most of the existing researches are based on PC based dynamic simulation method. Smart devices, which have remarkably low computing power, mainly use very simple 2D based technologies as in the example of Angry Bird, which has been largely successful as a recent game application.

In general, smart devices have limited computing power and screen size, so they do not require the results of very sophisticated and accurate transformation and deformation of objects, which is the goal of PC-based 3D deformed object simulation methods.

Instead, Smart Devices is committed to providing real-time simulation speeds for practical application to a variety of games and applications. Therefore, it is very important to provide real-time 3D deformed object motion and deformation results through an optimized algorithm considering the performance of current smart devices.

The finite element method, the boundary element method, and the finite volume method, which require a relatively long computation time, are not suitable for predicting the motion and deformation of a 3D deformed object in real time in a smart device.

Therefore, a relatively fast mass-spring method is suitable, but since most 3D deformed objects are bulky, deformations are easily distorted when external / internal forces are applied.

To prevent this, there is a method of applying a volume holding method through a mathematical operation or a method of installing a spring inside a deformed body. However, it is not easy to design a spring suitable for a deformed object or maintain a volume, and further operation is required.

Accordingly, the present invention applies FFD-AABB (Free Form Deformation-Axis-Aligned Bounding Box) modeling method which can physically, quickly, and simply express the motion and the deformation as well as maintaining the volume of the object with only the surface information of the 3D deformed object Implements a 3D transformed object.

The general FFD method is a sophisticated physics-based simulation method that transforms the shape of an object by using the force acting on the control point of the control grid connected by a spring, but it has the advantage of calculating the object deformation quickly, This method is suitable for 3D deformed object simulation.

The FFD method calculates the changed position (s, t, u) of the deformed object node by using the following equation (1) and obtains the final changed position p 'of the deformed object node using the following equation (2). This is calculated by using the difference of the center position of the FFD Grid including the node at the previous node position p = (x, y, z) of the object, C0 = (x0, y0, z0).

(S, t, u) are substituted into the equation (2) to obtain the final position p 'of the deformed object node, and the values of (s, t, u) Is used in Equation (2).

FIG. 1 is a view illustrating a collision sensing method in which an empty space is closely contacted with a surface of a 3D deformed object by applying an AABB to process a rapid collision.

That is, FIG. 1A shows an object before deformation, FIG. 1B shows an FFD grid, and FIG. 1C shows an FFD-AABB grid.

In the FFD-AABB method, as shown in FIG. 1B and FIG. 1C, AABB is applied to process a fast collision in a conventional FFD method, and the Grid is closely attached to the surface of the 3D deformed object.

In order to successfully simulate 3D deformed objects in real time on a smart device, geometric modeling should be performed so that the deformed objects can be represented concretely and the shape of the object can be well expressed and deformed. To do this, we conducted experiments on the modeling limit (number of maximum nodes and edges of 3D object) and the number of objects that can be processed in real time on iPhone 5.

Numerical integration for predicting the next state of a 3D deformed object usually uses Implicit Integration to maintain a stable simulation system. However, in Smart devices, Explicit Integration (Implicit Integration) is faster than Implicit Integration It is appropriate to apply Euler Integration or Verlet Integration method and apply stable simulation time interval. Therefore, the present invention can perform fast and stable simulation in a smart device which has insufficient computation ability compared to a PC by applying the Euler Integration method.

On the other hand, if the resolution of collision detection and collision is not accurate in 3D deformed object simulation, the overlapping phenomenon of objects that do not occur in the real world occurs, so that the simulation results are not realistic at all. Therefore, in dynamic simulation involving complex deformed objects, it is necessary to improve the accuracy and speed of calculation in collision detection and resolution.

In general, for precise collision detection, the collision detection centered on the base unit (node-edge) is required, which is the most computationally expensive step in dynamic simulation. However, because smart devices have a small screen size and limited computing power, it is even more important to quickly find and resolve collisions at the right level of accuracy than correct collision detection and resolution.

Accordingly, as shown in FIG. 2, a processor for producing a 3D advertisement is operated through a modeler for a 3D object. 3D models and animations require 3D professional modeling programs, such as 3D Max, Maya, and Blender.

Such a processor may include a configuration as shown in FIG.

Referring to FIG. 2, the processor for producing a 3D advertisement includes a first boundary forming unit 100, a second boundary forming unit 200, and a collision checking performing unit 300.

The first boundary forming part 100 forms a first step boundary surrounding the whole outer shape of the deformed object and the second boundary forming part 200 forms a plurality of two step boundary that surrounds a part of the deformed object.

At this time, if a collision occurs at the specific one-step boundary, the collision inspection performing unit 300 performs a first collision inspection on the two-step boundary of the specific one-step boundary, and collision occurs at the specific two- , A triangle-node based collision check is performed on the specific two-step boundary.

FIGS. 3 and 4 are flowcharts and an example of a two-step collision detection method considering morphological characteristics of a deformed object.

As shown in FIGS. 2 to 4, the FFD-AABB method is applied to a two-step bounding sphere inspection method considering the morphological characteristics of deformed objects for fast collision inspection and processing.

First, when the inspection is started, a first-level boundary L1 is created from the vertex data of the object, and a second-level boundary L2 is formed (s10).

The first-level boundary L1 is represented by red in Fig. 3, and is formed so as to surround the entire outer shape of the deformed object.

The second-level boundary L2 is represented by blue in FIG. 3, and is constituted by a few numbers within the first-level boundary L1.

Next, when a collision is detected in a specific first-level boundary L1, a collision check is performed using a second-level boundary L2 of the first-level boundary L1 in step s30.

Next, when a collision is detected in the specific two-level boundary L2, s60 is searched for only a portion corresponding to the specific two-level boundary L2.

In this way, the time consumed in collision inspection is minimized through two-stage segmentation. Also, for the detected two - level boundary region, we apply fast spatial hashing method to find the nodes with possibility of collision and then perform triangle - node - based collision detection only for these nodes. The results show that this method shortens the collision inspection time.

Most deformed objects consist of concave and convex shapes with many complex bends and pine shapes. Concave shaped objects require relatively finer collision tests than other objects. Since concave shapes may collide with each other due to the force when colliding with another part of the object, the boundary should be designed considering the shape when modeling the object. Therefore, when creating the bounding sphere of the concave shaped deformed objects, it is necessary to model them so as to include the vertices of the object as much as possible without overlapping the boundary sphere as much as possible according to the shape of the object.

5 is a diagram illustrating a positional change of a two-level boundary according to the movement of an object.

FIG. 5A shows a boundary before the alphabet K collides with the floor, and FIG. 5B shows a boundary where the position changes after collision with the floor.

 5A and 5B, the position of the second-stage boundary is effectively changed when the position of the vertex is changed according to the shape change of the object due to collision of the deformed object of the alphabet K with the floor.

The implemented method reads the data file of the position information (node, edge) of the object and the data file of the position information of the boundary sphere and uses it for the simulation.

After the information of the deformed object is loaded first, the first step boundary is created by using the vertex data of the object, and then the second step boundary of the object is created by referring to the first step boundary pit data file.

Since the boundary of the second stage is generated based on the vertex close to the center of the boundary of the first stage when the first boundary is created, if the position of the vertex of the object changes due to external collision or force, the center position of the collision sphere also changes.

Equation (3) shows a method of obtaining the center position c of the first-level boundary.

 The center of gravity of the first stage boundary is calculated by dividing the sum of all vertices by the number of nodes.

6 is a view simulating a free fall situation of various objects.

In order to analyze the limitations of 3D deformable object simulation that can be implemented in smart devices, we analyzed the number of objects that can be processed in real time on various objects and the complexity of deformed objects based on the latest smart device, the iPhone 5.

The iPhone 5 is based on the Apple A6 system chipset and consists of a two-core CPU with 1,300Mhz operating speed and an SGX543MP3 chipset for graphics processing. The GPU consists of three cores with 325Mhz speed have.

Therefore, the graphical effect can be expressed with relatively high performance as compared with other smart devices.

In addition, OpenGL | ES 1.X, version 2.0 and 3D graphics libraries are available for 2D and 3D rendering.

Experiments were carried out in such a situation that objects fall freely from a certain height to the floor with respect to various objects such as alphabet, note, cone, and tire shape as shown in FIG. 6. In the simulation, the collision problem between deformed objects, The calculation time was measured including the process of deforming when it collided.

As shown in Table 1, the average fps (frames per second) and the lowest fps were measured for 700 frames using 1 to 5 deformed objects.

In order to detect and resolve the collision of the 3D deformed object, additional computation is required. Therefore, at the moment of collision of the deformed objects, the performance is relatively low compared to the average fps.

That is, the minimum fps required for successful real-time 3D deformed object simulation is measured and presented.

transform
Number of objects Average number of vertices Average
Number of triangles Average
Number of FFD cells Average
fps lowest
fps One 2,809 5,636 100 17.46 17.46 2 1,016 2,030 97 17.59 15.13 3 509 1,015 97 17.91 16.03 4 215 426 89 17.04 15.17 5 214 426 95 14.41 12.39

The average number of vertices, the number of triangles, and the number of FFD cells in Table 1 mean the average number of vertices, triangles, and FFD cells used to represent objects in each simulation experiment.

Therefore, the experiments using three deformed objects were modeled using 509 vertices, 1,015 triangles, and 97 FFD cells per object, and the average was 17.91 fps during the simulation, and the lowest complexity of 16.03 fps Respectively.

In general, about 17 fps is required to represent the smooth motion of 3D deformed objects applied to the implementation of animations and games in smart devices. Table 1 shows how to model objects in real-time 3D deformed object modeling. And to provide appropriate guidelines on whether to do so.

The physically-based FFD-AABB algorithm implemented through the present invention detects and resolves the collision of deformed objects through a two-step boundary sphere inspection method based on the morphological analysis of the 3D deformed object, thereby realizing the realistic and immersive It is possible to provide a high application development environment.

In addition, it can provide guidelines for 3D deformed object modeling that can achieve execution speed to express smooth deformed object motion in smart device.

Therefore, if the deformed objects are modeled according to the present invention, it is possible to express the motion of a natural deformed object in real time in the game and the 3D object simulation application developed in the mobile device.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit or scope of the invention as defined in the following claims. There will be.

Claims (6)

A first boundary forming part forming a first boundary to surround the entire outer shape of the deformed object;
A second boundary forming unit forming a plurality of second boundaries surrounding a part of the deformed object; And
A first collision check is performed on the second boundary of the specific first boundary when a collision occurs at a specific first boundary and when a collision occurs at a specific second boundary, And a collision check performing unit for performing a collision check of the node based detailed collision. The method according to claim 1,
Wherein the collision checking unit includes:
A mobile hashes the node with a potential collision by applying a spatial hashing method to the specific second boundary, and performs the collision check based on the triangle-node only for the nodes. 3. The method of claim 2,
Wherein the first boundary and the second boundary have a sphere shape. Forming a first stage boundary surrounding the entire outer shape of the deformed object;
Forming a plurality of two-step boundaries surrounding a portion of the deformed object; And
A first collision check is performed on the second boundary of the first boundary when a collision occurs at a specific first boundary, and when a collision occurs at a specific second boundary, And performing node-based detailed collision checking on the node-based detailed collision model. 5. The method of claim 4,
The step of performing the collision checking includes:
Applying a spatial hashing method to the specific second boundary to find nodes with possible collision, and
Performing the triangle-based collision check on the nodes only
Dimensional deformed object model of a mobile advertisement engine. 5. The method of claim 4,
Wherein the first boundary and the second boundary have a sphere shape.

Images