VIRTUAL FACE GENERATING METHOD

Technical Field The present invention relates to a face generating method, and more particularly, to a virtual face generating method .

Background Art Generally, examples of generating a virtual face from a facial image obtained by a camera include a method of forming a facial image using an image mosaic technique or an image warping technique that depends on simple computer graphic technology. The image mosaic technique is a technique of combining an entire image with two or more images to obtain a transformed image. The image warping technique is a technique of applying a mathematical equation to a distortion filter to distort an image, thereby transforming the image such that the pixels of the image have new position values.

For example, Korean Patent Publication No .2003-0064711

(published on August 2, 2003) discloses a face formation method using face recognition. In the method, when a user accesses a face formation server through a wired/wireless communication terminal to transmit a real facial picture,

the face of the user is authenticated through a face recognition device embedded in the wired/wireless communication terminal, and temporarily stored in a user face formation database (DB) . When a facial portion of the facial image to be transformed is selected using a keyboard or a mouse button, each facial portion is transformed using an image warping technique and stored, outputted, or transmitted.

Disclosure

Technical Problem

However, since the transformed facial portion of a virtual formation face obtained through an image mosaic technique or an image warping technique does not harmonize with the size and contrast of an entire facial image, a separate manual job is required and the formed facial image is difficult to check in real-time.

An object of the present invention is to provide a virtual face generating method of extracting facial biometric feature information for each portion from a facial image of a user obtained by an image obtaining device such as a camera and a scanner, allowing a transformed facial portion to harmonize with the size and contrast of an entire facial image using the extracted facial biometric feature information for each portion and to be easily checked in

real-time, and easily generating a virtual formation face, a facial montage, or a virtual baby face.

Technical Solution According to a first embodiment of the present invention, there is provided a virtual face generating method including: obtaining a facial image and normalizing positions of facial portions of the obtained facial image for each facial portion to perform a pre-process; extracting facial biometric feature information for each portion of the facial image in which the positions of the facial portions have been normalized for each facial portion; and after the extracting of the facial biometric feature information for the each portion, when a specific facial portion to be transformed is selected from the facial image, converting facial biometric feature information of the selected facial portion into facial biometric feature information of a corresponding facial portion stored in a feature information DB for each facial portion provided in advance to generate a virtual formation face.

According to a second embodiment of the present invention, there is provided a virtual face generating method including: obtaining a facial image and normalizing positions of facial portions of the obtained facial image for each facial portion to perform a pre-process; extracting facial biometric feature information for each portion of the

facial image in which the positions of the facial portions have been normalized for each facial portion; and after the extracting of the facial biometric feature information for the each portion, converting facial biometric feature information for each portion of the facial image into facial biometric feature information of a corresponding facial portion stored in a feature information DB for each montage portion provided in advance to generate a virtual montage face . According to a third embodiment of the present invention, there is provided a virtual face generating method including: obtaining a facial image and normalizing positions of facial portions of the obtained facial image for each facial portion to perform a pre-process; extracting facial biometric feature information for each portion of the facial image in which the positions of the facial portions have been normalized for each facial portion; after the extracting of the facial biometric feature information for the each portion of the facial image, when a face synthesis ratio for each portion for a virtual baby face generation is selected with respect to facial biometric feature information for each portion of the facial image and facial biometric feature information of a corresponding portion stored in a feature information DB for each portion of a lover's face provided in advance, synthesizing/converting facial biometric feature information for each portion of the

facial image so that it is suitable for the selected face synthesis ratio to generate a virtual baby face.

Advantageous Effects As described above, according to the present invention, an image mosaic technique or an image warping technique requiring a separate manual job, which has been used in a related art virtual face forming method, is not used for a facial portion to be transformed. According to the present invention, a facial portion transformed harmonizes with the size and contrast of an entire facial image by using face recognition technology of extracting facial biometric feature information for each facial portion. Also, a virtual formation face or a montage face, or a virtual baby face for a specific facial portion can be easily generated. In the case where the present invention is applied to 2- dimensional (D) and 3D virtual face formation simulation, a montage producing program, a face synthesis simulation for each portion, a virtual baby simulation, and a wired/wireless face formation program, an interest of a user can be increased.

Description of Drawings

FIG. 1 is a flowchart illustrating a virtual face generating method according to the present invention;

FIG. 2 is a view illustrating an embodiment of a process for generating a virtual formation face according to a virtual face generating method of the present invention;

FIG. 3 is a view illustrating an embodiment of a process for generating a virtual montage face according to a virtual face generating method of the present invention; and FIG. 4 is a view illustrating an embodiment of a process for generating a virtual baby face according to a virtual face generating method of the present invention.

Best Mode

Hereinafter, first, second, and third embodiments of the present invention are described in more detail with reference to the accompanying drawings . FIG. 1 is a flowchart illustrating a virtual face generating method according to the present invention, and illustrates all of the first, second, and third embodiments of the present invention. Processes SlO, S20, and S30 are common processes performed in the three embodiments. On the first place, the facial image of a user is obtained by an image obtaining device such as a camera and a scanner (SlO) , the positions of facial portions such as eyes, a nose, a mouse, an eyebrow, and a jaw of the obtained facial image are normalized for each facial portion and a pre-process is performed (S20) , and facial biometric feature information of each portion of a normalized facial image in which the

positions of the facial portions have been normalized for each facial portion (S30) .

At this point, facial biometric feature information for each facial portion is extracted using one of local facial biometric feature extracting methods such as local feature analysis (LFA) , local non-negative matrix factorization

(LNMF) , and independent component analysis (ICA) . The local facial biometric feature extracting method can obtain facial biometric feature information for each portion of a facial image using feature vectors as base vectors. The feature vectors have a characteristic that the local region of a face such as eyes, a nose, a mouse, an eyebrow, and a jaw is emphasized.

The positions of facial portions of a facial image obtained by an image obtaining device such as a camera and a scanner are normalized for each facial portion, and facial biometric feature information for each portion of the normalized facial image is extracted (S10-S30) . Hereinafter, a process for performing virtual face generating methods according to the first, second, and third embodiments of the present invention will be described below in detail.

FIG. 2 is a view illustrating a first embodiment of a process for generating a virtual formation face according to a virtual face generating method of the present invention, and illustrates a process for generating a virtual formation face using facial biometric feature information for each

portion extracted from a facial image of a user obtained by an image obtaining device such as a camera and a scanner.

In the first embodiment of the present invention illustrated in FIG. 2, facial biometric feature information for each portion of the normalized facial image is extracted (S10-S30) . When a specific facial portion to be transformed is selected from the facial image (S40) , facial biometric feature information of the selected facial portion is converted into facial biometric feature information of a corresponding portion stored in a feature information DB (a DB table-type DB illustrated in FIG. 2) for each facial portion provided in advance (S46) to generate a virtual formation face (S44), and then separately stored if necessary (S46) , or the stored virtual formation face can be output to an external device such as a printer.

FIG. 3 is a view illustrating a second embodiment of a process for generating a virtual montage face according to a virtual face generating method of the present invention, and illustrates a process for generating a virtual montage face using facial biometric feature information for each portion extracted from the facial image of a user obtained by an image obtaining device such as a camera and a scanner.

In the second embodiment of the present invention illustrated in FIG. 3, facial biometric feature information for each portion of the normalized facial image is extracted (S10-S30) . Facial biometric feature information of the each

facial portion of the facial image is converted into facial biometric feature information of a corresponding portion stored in a feature information DB (a DB table-type DB illustrated in FIG. 3) for each portion of a montage provided in advance by a montage design expert to generate a virtual montage face (S50 and S52), and then the generated virtual montage face can be separately stored if necessary

(S54) , or the stored virtual montage face can be output to an external device such as a printer. At this point, the virtual montage face is generated through conversion on the basis of similarity information

(for example, Euclidian distance) between conversion coefficients of the facial biometric feature information of each portion of the facial image and feature information for each montage portion provided in advance.

FIG. 4 is a view illustrating a third embodiment of a process for generating a virtual baby face according to a virtual face generating method of the present invention, and illustrates a process for generating a virtual montage face using facial biometric feature information for each portion extracted from a facial image of a user obtained by an image obtaining device such as a camera and a scanner.

In the third embodiment of the present invention illustrated in FIG. 4, facial biometric feature information for each portion of the normalized facial image is extracted

(S10-S30) . When a face synthesis ratio for each portion for

virtual baby face generation is selected with respect to facial biometric feature information for each portion of the facial image and facial biometric feature information of a corresponding portion stored in a feature information DB for each portion of a lover's face provided in advance (S60) , facial biometric feature information for each portion of the facial image is synthesized/converted suitably for the selected face synthesis ratio to generate a virtual baby face (S62 and S64), and then the generated virtual baby face is separately stored if necessary (S66) , or the stored virtual baby face can be output to an external device such as a printer.

Although the preferred embodiments of the present invention have been disclosed for illustrative purpose, those skilled in the art will appreciate that various modifications, additions and substitutions can be made without departing from the scope and spirit of the invention as defined in the accompanying claims.