KR20000049906A - Making method of avatar to come under substantial user in cyber space and system for the same - Google Patents

Abstract

PURPOSE: A method and a system for creating avatar corresponding to real user in cyber space are provided to create a second model mostly adapted to an user for represent the user himself in internet shopping mall. CONSTITUTION: A method for creating avatar corresponding to real user in cyber space comprises: a first step human characters of various humane body profiles; a second step extracting only an image of user's face from an image file; a third step receiving the data of the user's body profile to access to a character of humane characters stored in the first step nearest to the data; a forth step overlapping the image extracted in the second step to the face of a humane character accessed in the third step; and a fifth step matching the image overlapped in the forth step to the face of the third step, thereby creating one avatar.

Description

Making method of avatar to come under substantial user in cyber space and system for the same}

The present invention relates to a virtual human (avatar) generation method that operates in a virtual space on the Internet, in particular, virtual humans representing a specific user, such as Internet shopping malls are created to be substantially the same as the user cyber to have a representative The present invention relates to a method and system for creating a virtual human being corresponding to a user who is detrimental in space.

In general, an application technique using a three-dimensional (3D) object model is applied in the field of video coding and computer graphics, and is currently on the Internet, which is of much interest. The application has a wide range of applications due to the infinite development potential using the constructed virtual space as an environment.

In addition, as e-commerce using the Internet, ie shopping malls, has become more active, it provides the convenience of allowing users to purchase various items in a virtual space built on the Internet instead of going directly to the mall. Expressing, packaging and displaying various products through computer graphics as described above is becoming intimate enough to not feel the novelty or rejection.

At this time, the way of expressing various products on the existing shopping mall is not the actual 3D image, but dozens of 2D images of the corresponding product, when the user clicks and moves the mouse, it rotates in the desired direction and the effect is as if it is a 3D image. It introduces a product by giving a, and it still contains technical limitations of computer graphics.

Therefore, a method of showing a specific model using a product using video image data is also used for a product that is determined to be unclear of the function or appearance of the product by the computer graphic technology.

However, this method requires a lot of time and labor by requiring 2D images and video of numerous products.

Moreover, with the proliferation of multimedia in recent years, methods such as caricatures have been used to express individuals in various programs as well as interactive chatting or e-mail. This phenomenon is due to the increase in the desire for users to express themselves in the virtual world due to the spread of the virtual world due to the Internet.

However, the current caricatures or virtual human beings are not able to express themselves properly, and there are still too many shortcomings to use for e-commerce. The reason for this is that when a user wants to purchase a specific product in an e-commerce using the Internet, the user sees the product through another specific model, so that the product is suitable for himself, personal taste and body such as clothing or shoes and accessories. For products that tend to vary in size or body type, it is impossible to determine whether they fit well with the user.

In general, in order to express a human, it is necessary to properly express the facial features and various physical characteristics of the human body, that is, the body sizes. In e-commerce, a user may directly touch or wear various objects by hand. Because you can not buy through, you do not feel the sense that you can feel in the real world.

In order to solve these problems, there is an Internet site operated by a company called "Land's End". As shown in FIG. 1, simple data from the user, ie, shoulder, waist, hip, chest, etc. The size of the input is divided into "Large", "Medium", "Small", etc., and the hairstyle, skin color and face model can be selected as desired by the user.

However, although the method of generating virtual humans according to the diversification of these options has advanced considerably compared to the conventional method of generating virtual humans, it is not only possible to accurately input data for each user having a unique appearance and body shape for every individual entity. Since it is a virtual face model instead of a user's own face model, there is a problem that an intuitive and realistic comparison cannot be made from the user.

In addition, it is all about applying a specific product to a specific model without considering the face and body characteristics of the user at all, which implies that the user has no way of knowing how the product fits or fits the user.

An object of the present invention for solving the above problems is to receive a 2D face image of the user directly from the user in order to have a virtual person representing a specific user, such as an Internet shopping mall is created substantially the same as the corresponding user The virtual image corresponding to the detrimental user in the cyber space by using the user's face image as it is and inputting the user's body size to generate a body model suitable for the input size to generate a second user's own model most suitable for the user. The present invention provides a method and system for producing human.

1 is an exemplary view for explaining a conventional virtual human formation method,

2 is a flow chart of a virtual human creation process according to the present invention;

3 is an exemplary flowchart of a face extraction process according to the present invention;

4 is an exemplary view of image data processing for explaining a face extraction method according to the present invention;

5 is an exemplary configuration module of a website server for managing cyber space used as a virtual human creation and activity stage according to the present invention;

6 is an exemplary diagram of a process of inputting body size data of a virtual human according to the present invention;

7 is an exemplary shopping example using virtual humans according to the present invention.

Features of the present invention for achieving the above object, a first process for storing a human character of various body types, a second process for extracting only the image of the user's face from the image file for any user input, and the user A third step of accessing the character closest among the human characters stored in the first step by receiving data on the X-shape of the first shape, and the second part of the face of the human character accessed in the third step; The fourth process of overlapping the extracted image of the user's face, and the image of the user's face overlapped in the fourth process is matched to the face part of the human character accessed in the third process in an optimal state to create a virtual human And a fifth process of forming.

The above object and various advantages of the present invention will become more apparent from the preferred embodiments of the present invention described below with reference to the accompanying drawings by those skilled in the art.

First, a brief description of the technical concept to be applied in the present invention, the face of the virtual man, the biggest disadvantage of the conventional virtual man is that the face of the mannequin as a basic model. Accordingly, the present invention is focused on the fact that the virtual human may represent the user more if the virtual mannequin face is used as the actual face of the user.

Therefore, after extracting only the face part from the ID photos that are commonly possessed by the user, the virtual face of the user best suited to the input body size of the user is upgraded to create the virtual face of the user. It is an object of the present invention.

In addition, the present invention proposes an internet shopping mall in a virtual space in which a virtual human of the above-described concept can act.

Hereinafter, exemplary embodiments of the present invention will be described with reference to the accompanying drawings.

Referring to FIG. 2 attached to the process to which the present invention is applied on a certain shopping mall as in the above-described technical concept, in step S100, the user creates his photo as an image file (using a digital camera or a scanner). Thereafter, after accessing the Internet website provided by the present invention through the step S200, the user uploads the photo file generated in the step S100 and inputs the user's own body size. At this time, his or her body type and its features may also be input.

As described above, the website according to the present invention receiving the user's own photo and body size through step S200 extracts only the image of the face portion from the user's photo file uploaded through step S300, and simultaneously through step S400. The model most similar to the input body size of the user is generated.

Thereafter, in step S500, a virtual human body is formed after overlapping the face image of the user extracted in step S300 on the face portion of the model generated through the process of step S400.

The virtual human representing the user formed through the process of step S500 is active in the virtual space (for example, a shopping mall) on the website provided by the present invention through the process of step S600.

Looking at the flow of the technical idea according to the present invention in detail, the extraction process of the face image is shown by way of example in the accompanying FIG.

As shown in FIG. 3, a description of a series of processes for extracting only an image of a face after generating an image file through a scanner or the like, which is commonly held in general, such as an ID photo, has been performed at home and abroad. Proposed papers among them are 『Sun Hwang Kyu, Jae Ho Lee, Hyung Joon Kim, Hoe Yul Kim] Real-Time Face Tracking Using Color Information”, Journal of Information Processing Society, Vol. 26, No. 2, pp. 368-370, 1999. "

In the example of the ID picture in Figure 3 is generally because the social life of one or two pictures of the user is commonly owned by the user, mainly used for the type of ID picture 3 x 4 cm ² , 4 x 5 cm ² This is because the sizes of 5 x 6 cm ² and the like are almost standardized.

Therefore, in the present invention, it is recommended to use an image scanned by a 3 x 4 cm ² ID photo that is used most by the general public to create a user's face, and preferably improve the quality of the face image when scanning the ID photo. In order to unify and unify, set the resolution to 200 to induce a scan.

With the transmitted user's face image and body size, the web server generates the model that is most similar to the user in real time. By showing the created user model on the web, users can enjoy the real-time charm and the pleasure of seeing their own model.

In addition, the posture of the model desired by the user may not always stand in an upright position but may be a side profile or a slightly oblique posture. Therefore, it also adds the ability to rotate the face by a certain angle to adapt the face according to the model's posture.

First, the image processing technique applied in the face image extraction process as described above will be briefly described. In the conventional image processing technique, the color image is gray scaled without using the color of the original image when the image is processed. It uses a lot of RAW files converted to Scale. On the other hand, general graphics file formats (BMP, GIF, JPG, TIF, etc.) have header information, and the program needs to analyze the header information, and the program length increases and complexity increases.

On the other hand, since the RAW file format does not have header information, there is no need to analyze headers for image processing, and since image data is directly processed, no additional code is required. The disadvantage of RAW files is that developers need to know the size of the video. The following figure compares RAW files and BMP files. RAW files contain only human face data, and BMP files contain header information that stores the size, color, and resolution of a human face.

Therefore, in the present invention, an edge is detected by a user directly manipulating a human face on a web while using a color image without applying a conventional image processing technique that converts and uses a color image to a RAW file. Here's an intuitive way to do it.

In general, how edges are detected in color images depends on how the developer defines them. That is, one of them is to define an edge as a discontinuity of image brightness.

In this case, edge detection is performed on the brightness channel of the color image in the HSI (Hue, Saturation, Intensity) space. Another definition of edges is that edges exist when there is a discontinuity in brightness in the red, green, and blue channels. Thus, detection of an edge can be defined as in Equation 1 below.

That is, in order to detect an edge in a color image, if the conventional method is to convert the image to a gray scale image using an RGB value and then compare the image with a brightness value, in the present invention, RGB is expressed as in Equation 2 below. The edge is detected as it is.

C (x, y) = [R _t (x, y)-R _s (x, y)] / R _t (x, y)> T _r &

[G _t (x, y)-G _s (x, y)] / G _t (x, y)> T _g &

[B _t (x, y)-B _s (x, y)] / B _t (x, y)> T _b

Here, R _t (x, y) means the red value of the pixel to be compared, and R _s (x, y) means the red value of the original pixel. And T _t means the tolerance of the pixels to be compared.

Hereinafter, a configuration of a web server provided by the present invention will be described with reference to FIG. 5.

The web server shown in FIG. 5 is a connection configuration of software modules, and since it can be changed into a wide variety of structures, it does not focus on the connection configuration and therefore does not give a reference number, and focuses on the functional characteristics of each module. Let's take a look at this.

First, look at the function of the part called "Apache Web Server". It performs the homepage functions of websites such as bulletin boards, guest books, logins, etc., manages the entire state, and sets the initial state of the screen. do. Also, it is the only gate of each linked component, which is the only part that can access the D / B (accessible at LOGIN with personal ID).

In addition, the data stored in the D / B stores the user's face image extracted from the character (Character), the costume, and the ID photo, the body size (chest, waist, hip, height input by the user) ) Data is delivered to the Codination (CODI) Engine. At this time, only a user who has registered as a user can read a character similar to his or her body by adjusting a body size, and a user who does not read a character set as a default.

In the above, we looked at the function of the part called "Apache Web Server". The function of the "M / W model and clothes D / B" stores the male / female model and the clothes D / B. A function of loading a model and clothes into the CODI engine by (enter body size).

The function of the "CODI Engine" portion of the user's body size (chest, waist, hips, height) can actually wear the user's face and clothes on the character corresponding to the size, the "Men / This is the part that synthesizes the characters imported from the model and costume D / B "and the processed faces from the Image Processing Server. That is, it has a function of matching the optimal state to the character's body by moving the face image processed in the image processing server.

In addition, if the skin color of the character loaded from "M / fe model and costume D / B" and the face image processed by the image processing server are not the same, the skin color matching operation keeps the skin color similar. Create a scroll bar to allow the user to choose a color each time the user moves the scroll bar, but it can be divided into whites, blacks, yellows, etc. so that the user can select a color such as the skin color of the face image processed by the image processing server. It is.

The organic combination of the functions described above forms the most suitable virtual human with the user. As shown in FIG. 7, the user is most suitable for himself by applying various products including clothes to his own model. You can quickly and quickly select the right products in real time.

In this case, FIG. 6 shows a process in which the user inputs his / her body size. The face image simply shows a state in which the extracted face of the user overlaps the face portion of the mannequin in the extracted state.

It's as if you're wearing a variety of clothing at a department store or clothing store and looking at the mirror to see if it fits you well.

While the invention has been shown and described in connection with specific embodiments thereof, it is well known in the art that various modifications and changes can be made without departing from the spirit and scope of the invention as indicated by the claims. Anyone who owns it can easily find out.

As described above, according to the present invention, if a virtual human generation method corresponding to a real user and a system according to the present invention are provided, the 2D face image extracted from only the head and face except the background from the user's ID photo is extracted as it is. It is the most necessary in an internet shopping mall that specializes in clothing, as it generates avatars that are most suitable for the user's body size and the like. In addition, as a linked product, the user can wear various costumes suitable for his or her body to express an image of the effect of taking a family picture with an individual as well as a lover or using a plurality of avatars. You can also make. In addition, various posters can be featured as main characters, and creative posters can be produced. And it can also be applied to chat, which is very popular among young teenagers.

Claims (5)

A first process of storing human characters of various body types; Extracting only an image of a corresponding face of a user from an image file of an input user; A third step of receiving data on the user's body type and accessing a character closest to the human character stored in the first step; A fourth step of overlapping the image of the user's face extracted through the second step on the face part of the human character accessed in the third step; And And a fifth process of matching an image of the face of the user overlapped in the fourth process to an optimal state with the face part of the human character accessed in the third process to form a virtual human. The virtual human generation method corresponding to the actual user in the. The method of claim 1, In the second process, in order to extract only the image of the user's face from the image file for the user, edge is detected by applying RGB as it is, C (x, y) = [R _t (x, y)-R _s (x, y)] / R _t (x, y)> T _r & [G _t (x, y)-G _s (x, y)] / G _t (x, y)> T _g & [B _t (x, y)-B _s (x, y)] / B _t (x, y)> T _b However, R _t (x, y) means the red value of the pixel to be compared, and R _s (x, y) means the red value of the original pixel. And T _t is the tolerance of the pixels to be compared. When there is a discontinuity of brightness in the red, green, and blue channels, the edge is present in the virtual space creation method corresponding to the actual user in the cyber space. The method of claim 1, And receiving data of the user's image file and body shape through the second process and the third process to be input in an online state through the Internet. When an image file of an arbitrary user inputted through an internet connection and data of a corresponding body type are inputted, only an image of the user's face is extracted from the image file, and the inputted body type is selected from among various pre-stored human characters. A cyber person using virtual humans formed by overlapping the image of the user's face extracted from the face part of the character in the optimal state and forming virtual humans in close proximity to the user as a subject for commerce activities in the virtual space of the Internet e-commerce shopping mall. A virtual human creation method corresponding to a real user in space. A web server that performs a homepage function of a website, manages an entire stage, sets an initial state of a screen, and accesses a D / B as a unique gate of modules which are linked components; A database storing data on male and female models and clothes, and outputting the corresponding models and clothes by inputting a user's body size; The user's face and clothes can be put on the character corresponding to the body size (chest, waist, hip, height) input by the user, and the characters loaded from the database and the face processed by the image processing server are synthesized. Gasing human generation system corresponding to the actual user in the cyber space, characterized in that it comprises a matching engine.