KR20040009460A - System and method for constructing three dimensional montaged geometric face - Google Patents

Abstract

PURPOSE: A system and a method for obtaining a three-dimensional face montage are provided to produce a montage more similar to an actual face. CONSTITUTION: A system for obtaining a three-dimensional face montage includes a face shape storage unit(7-16), a selective reading unit, and a producing unit. The face shape storage unit stores a plurality of three-dimensional partial face shapes. The selective reading unit selectively reads a plurality of three-dimensional partial face shapes corresponding to a plurality of face characteristic information items from the face shape storage unit. The producing unit combines the selectively read plurality of three-dimensional facial face shapes to produce a three-dimensional face montage. The system further includes an index storage unit(6) that stores index information corresponding to the face characteristic information.

Description

System and method for constructing three dimensional montaged geometric face}

The present invention relates to a three-dimensional face montage acquisition system and method for producing a three-dimensional face montage by combining a three-dimensional partial face shape, and more particularly, a three-dimensional part corresponding to a plurality of face feature information input from the outside, respectively. The present invention relates to a three-dimensional face montage acquisition system and method for selectively reading a face shape from a database and combining the selected three-dimensional partial face shapes to produce a three-dimensional face montage.

For many years, montage production has been performed by combining several images or photographs to create a single screen or photograph for the purpose of reproducing a human face. The most common way of producing a montage is to draw a face shape that reflects the facial features by a person. Another method is to produce a montage image by fitting pieces of photographs of a part of a photograph containing the facial features. .

However, this montage production method has a limit made on a two-dimensional plane.

Accordingly, an object of the present invention is to overcome the above limitations, and selectively reads a plurality of three-dimensional partial face shapes corresponding to a plurality of face feature information input from the outside from a database and selects the plurality of selected three-dimensional images. It is to provide a three-dimensional face montage acquisition system and method to produce a three-dimensional face montage by combining the partial face shape, to reproduce the montage closer to the actual face.

In addition, the present invention provides a three-dimensional face montage by combining a plurality of three-dimensional partial face shapes corresponding to a plurality of face feature information using the three-dimensional partial face shape index information stored in a database, thereby making the three-dimensional face more quickly. Another object is to provide a three-dimensional face montage acquisition system and method for producing a montage.

In addition, the present invention, by storing the three-dimensional partial face shape in the database as a projection 3D image in the form of range data, a three-dimensional face montage that allows more three-dimensional partial face shape to be stored while taking up less capacity It is another object to provide an acquisition system and method.

In addition, the present invention, when inputting the correction request data for the produced three-dimensional face montage, edit the produced three-dimensional face montage in response to the input correction request data, but rotates the three-dimensional partial face shape to the minimum unit It is another object to provide a 3D face montage acquisition system and method that reproduces a montage closer to a real face by performing editing by copying, reducing, enlarging, moving the position, and the like. .

In addition, the present invention, by supporting the interpolation function to smoothly change the discontinuity line of the edge of any three-dimensional partial face shape of the produced three-dimensional face montage to a continuous line and to fill the occlusion region, Another object is to provide a three-dimensional face montage acquisition system and method for reproducing a montage closer to an actual face.

The present invention is characterized in that the three-dimensional partial face shape is a shape in which each part of the face including the forehead, the eye, the eyebrow, the nose, the middle ear, the mouth, the chin, and the ear is independently separated, thereby realizing more than the actual face. Another object is to provide a three-dimensional face montage acquisition system and method for reproducing a close montage.

In addition, the present invention is characterized in that the three-dimensional partial face shape is a shape for each part of the face divided horizontally to include the head and forehead, eyes and eyebrows, nose, mouth, and chin, respectively, Another object of the present invention is to provide a 3D face montage acquisition system and method for reproducing a closer montage and producing a 3D face montage more quickly.

1 is a configuration diagram of a first embodiment to which the present invention is applied,

2 is a configuration diagram of a second embodiment to which the present invention is applied;

3 is a flowchart of a first embodiment to which the present invention is applied;

4A and 4B are three-dimensional partial face shapes (full face shape and face part shape) according to the first embodiment,

Figure 4c is a standard comparison pattern of the three-dimensional partial face shape (face portion shape) of Figure 4b,

FIG. 5 is a detailed flowchart of S306 in FIG. 3;

FIG. 6 illustrates a reference standard comparison pattern generated from the three-dimensional partial face shape (whole face shape) of FIG. 4A.

Figure 7a illustrates a three-dimensional partial face shape (face portion) corresponding to the 'mouth',

FIG. 7B is edited by replacing the three-dimensional partial face shape (face partial shape) of FIG. 7A with the three-dimensional partial face shape (full face shape) of FIG. 4A.

FIG. 8 is a view of converting FIG. 7B into 3D polygon data.

9 is a three-dimensional partial face shape (face portion shape) according to the second embodiment,

10A and 10B show a three-dimensional partial face shape (face portion shape) cut out according to the second embodiment,

11 is a flowchart of a second embodiment to which the present invention is applied.

<Explanation of symbols for the main parts of the drawings>

1,101: client terminal 25,125: server computer

2,102: Display 3,20,103,120: I / O

5,105: Projection 3D Image Search Engine

6: index DB 7,107: whole face DB

8: forehead DB 9: first eye and eyebrow DB

10: No. 1 DB 11: Inseam DB

12: first mouth DB 13: first jaw DB

14: Ear DB 15: Head DB

16: sideburns DB 17,117: projection 3D image editing unit

18,118: converting section 19,119: three-dimensional polygon data memory section

21,121: 3D polygon data texture generator

22,122: 3D polygon data editing unit

23,123: inverse transform unit 24,124: projection 3D image memory unit

126: Head and Forehead DB 127: Second Eye and Eyebrow DB

128: second nose DB 129: second mouth DB

130: second jaw DB

The present invention for achieving the above object, the face shape storage means for storing a plurality of three-dimensional partial face shape is divided; Selective reading means for selectively reading out a plurality of three-dimensional partial face shapes corresponding to a plurality of face feature information from said face shape storing means; And manufacturing means for producing a three-dimensional face montage by combining the plurality of selected three-dimensional partial face shapes.

The present invention may further include index storage means for storing index information corresponding to the facial feature information, wherein the selective reading means reads index information corresponding to a plurality of facial feature information from the index storage means, and A three-dimensional face montage acquisition system characterized in that the three-dimensional partial face shape corresponding to the read index information is selectively read out from the face shape storage means.

In addition, the present invention, wherein the selected three-dimensional partial face shape is composed of the entire face shape and the face portion shape, the index information is the positional information of the standard comparison pattern in the face portion shape, the production means is A detection unit which detects the entire face shape among the three-dimensional partial face shapes read out selectively; A generation unit generating a reference standard comparison pattern corresponding to the positional information of the standard comparison pattern in the detected whole face shape; And a combination unit for matching the generated reference standard comparison pattern with the standard comparison pattern and combining the selected three-dimensional partial face shapes to solve the above-mentioned problems. .

In addition, the present invention is a three-dimensional face montage acquisition system, characterized in that the three-dimensional partial face shape is in the form of range data, solves the above problems.

In addition, the present invention, while editing the produced three-dimensional face montage in response to the inputted correction request data when inputting the correction request data for the three-dimensional face montage, rotates the three-dimensional partial face shape in a minimum unit, The first editing means is performed by performing at least one editing operation of copying, reducing, enlarging, and moving the position, or by performing an editing operation of replacing an arbitrary three-dimensional partial face shape among three-dimensional face montages with another three-dimensional partial face shape. A three-dimensional face montage acquisition system, further comprising the above-mentioned problem.

In addition, the present invention, when inputting the correction request data for the produced three-dimensional face montage edits the produced three-dimensional face montage in response to the input correction request data, any three of the produced three-dimensional face montage A three-dimensional face montage acquisition system, further comprising: second editing means for smoothing the discontinuity line of the edge of the dimensional partial face shape into a continuous line and filling and editing the occlusion region. Solve one task.

In the present invention, the three-dimensional partial face shape is a three-dimensional face montage, characterized in that each part of the face including the forehead, eyes and eyebrows, nose, human, mouth, chin, ears are separately divided As an acquisition system, the above-mentioned subject is solved.

In addition, the present invention, the three-dimensional selected partial face shape is composed of the entire face shape and the face portion shape, the manufacturing means, the selected read face portion shape to fit the left and right edge of the whole face shape Cutting part to cut off; And a combination unit for combining the cut-out face parts in a specified order, which solves the above-mentioned problems.

In addition, the three-dimensional face of the present invention is a three-dimensional face, characterized in that the shape for each part of the face divided horizontally so that the head and forehead, eyes and eyebrows, nose, mouth, and chin are included, respectively. As a montage acquisition system, the above-mentioned problem is solved.

In addition, the present invention for achieving the above object, the selective reading step of selectively reading out a plurality of three-dimensional partial face shape corresponding to each of the plurality of face feature information from the face shape database; And a manufacturing step of manufacturing a three-dimensional face montage by combining the plurality of selected three-dimensional partial face shapes.

Also, in the selective reading step, the index information corresponding to the plurality of facial feature information is read from an index database, and the three-dimensional partial face shape corresponding to the read index information is selectively read out from the face shape database. As a three-dimensional face montage acquisition method characterized in that, to solve the above problems.

In addition, the present invention, the selected three-dimensional partial face shape is composed of the entire face shape and the face portion shape, the index information is the position information of the standard comparison pattern in the face portion shape, the production step, the A detection step of detecting the entire face shape among the three-dimensional partial face shapes read out selectively; Generating a reference standard comparison pattern corresponding to the positional information of the standard comparison pattern in the detected whole face shape; And a combination step of combining the selected read-out three-dimensional partial face shape by matching the generated reference standard comparison pattern with the standard comparison pattern. Solve.

The present invention also provides a three-dimensional face montage obtaining method, wherein the three-dimensional partial face shape is in the form of range data.

In addition, the present invention, when inputting the correction request data for the produced three-dimensional face montage, performing the editing operation such as rotating, copying, reducing, enlarging, moving the position of the three-dimensional partial face shape in a minimum unit or three-dimensional A method for acquiring a three-dimensional face montage further comprising the step of performing an editing operation of replacing an arbitrary three-dimensional partial face shape among the facial montages with another three-dimensional partial face shape, thereby solving the above-mentioned problems. do.

In addition, the present invention, when inputting the correction request data for the produced three-dimensional face montage, smoothly changing the discontinuous line of the edge of any three-dimensional partial face shape of the produced three-dimensional face montage to a continuous line (smoothing) A method for acquiring a three-dimensional face montage, further comprising a second editing step of filling and editing an occlusion region.

In the present invention, the three-dimensional partial face shape is a three-dimensional face montage, characterized in that each part of the face including the forehead, eyes and eyebrows, nose, human, mouth, chin, ears are separately divided As an acquisition method, the above-mentioned problem is solved.

In addition, the present invention, the selected three-dimensional partial face shape is composed of the entire face shape and the face portion shape, the manufacturing step, the selected read face portion shape to fit the left and right edge of the whole face shape Cutting step of cutting; And a combination step of combining the cut-out face portions in a specified order, which solves the above-mentioned problems.

In addition, the three-dimensional face of the present invention is a three-dimensional face, characterized in that the shape for each part of the face divided horizontally so that the head and forehead, eyes and eyebrows, nose, mouth, and chin are included, respectively. As a montage obtaining method, the above-mentioned problem is solved.

In the embodiments to which the present invention is applied, the three-dimensional face montage obtaining system is a server computer, and transmits a plurality of facial feature information to the server computer, and is produced by the server computer. It is assumed that the subject selectively outputting the modification request data for the client is a client terminal connected to the server computer.

Therefore, a plurality of facial feature information is provided by the user using a character or image input device provided or connected to the client terminal (features small eyes, nose is dull, face elongated, forehead convex, mouth Big, chin sticking out).

1 is a configuration diagram of an embodiment to which the present invention is applied and receives a plurality of facial feature information through an input device such as a keyboard and a mouse, and requests for correction of a three-dimensional face montage by a user's manipulation. A client terminal 1 for outputting data; And storing a three-dimensional partial face shape in a range data form, and selectively reading out a three-dimensional partial face shape corresponding to a plurality of face feature information input from the client terminal 1, and selecting and selecting the three-dimensional partial face shape. And a server computer 25 that combines partial face shapes to produce a three-dimensional face montage.

The client terminal (1) includes a display unit (2) for receiving a three-dimensional face montage produced (or modified) by the server computer (25) from the server computer (25); And an I / O unit 3 for transmitting and receiving data to and from the server computer 25.

The server computer 25 includes: an I / O unit 20 for receiving a plurality of facial feature information and correction request data from the client terminal 1 and transmitting a produced (or edited) three-dimensional face montage to the client terminal; An index DB 6 storing index information corresponding to a plurality of facial feature information received from the client terminal 1; The three-dimensional partial face shape corresponding to the index information is stored as a projection 3D image in the form of range data, and the three-dimensional partial face shape includes the entire face, forehead, eyes and eyebrows, nose, middle weight, mouth, chin, and ear. Each of the DBs 7 to 16, characterized in that each part of the face, and the head and the whiskers are independently separated shapes; The respective index information corresponding to the received plurality of facial feature information is read from the index DB 6, and a three-dimensional partial face shape corresponding to the read index information is selected from the entire face DB 7 to sideburn DB 16. A projection 3D image search engine 5 to read out; A reference standard comparison pattern is generated from the entire face shape among the three-dimensional partial face shapes input from the projection 3D image retrieval engine 5, and the standard comparison pattern of the face partial shape among the three-dimensional partial face shapes and the generated reference standard. The three-dimensional partial face shape (the whole face shape and the face part shape) are combined by matching the comparison pattern, and the combined three-dimensional face montage is three-dimensional partial face shape in response to the correction request data input from the client terminal 1. Edit by rotating, copying, reducing, enlarging, moving the position, etc. in a minimum unit, or changing any 3D partial face shape among pre-made 3D face montages input from the projected 3D image memory unit 24. A projection 3D image editing unit 17 for substituting a three-dimensional partial face shape to edit a previously produced three-dimensional face montage; A conversion unit 18 for converting the edited three-dimensional face montage into three-dimensional polygon data; A three-dimensional polygon data memory unit (19) for temporarily storing the converted three-dimensional polygon data; A three-dimensional polygon data texture generator 21 for adding color information to the temporarily stored three-dimensional polygon data; 3D polygon data for editing the 3D polygon data in response to the correction request data received upon receiving the correction request data requesting to edit the 3D polygon data of the produced 3D face montage through the I / O unit 20 Editing section 22; An inverse transform unit 23 for inverting a three-dimensional face montage in the form of three-dimensional polygon data temporarily stored in the memory unit 19 into a three-dimensional face montage in the form of range data; And a projection 3D image memory unit 24 for temporarily storing the inverted three-dimensional face montage.

Among the plurality of DBs 7 to 16, there are forehead DBs 8 to ear DBs 14 as databases related to the biometric information of a human face part, and index DBs 6 are associated with these DBs 8 to 14. Index information stored in) includes the positional information of the standard comparison pattern of a rectangular shape having a certain size in the three-dimensional partial face shape stored in the DB (8 ~ 14) and described in detail as follows.

First, the standard comparison pattern of the forehead is designated at a position corresponding to the left / right center of the forehead, and in addition to the positional information of the standard comparison pattern, the curved features such as a round forehead and a flat forehead from curved surface information in the standard comparison pattern are also specified. Used as index information. The standard comparison pattern of the eye is specified at the center of the eye by extracting the top / bottom / left / right edges of the eye, and the geometric features that classify the size of the left / right width of the eye and the area of the eye are also used as index information. do. The standard comparison pattern of the eyebrows is specified above the center of the eye, the standard comparison pattern of the nose is assigned to the apex of the nose, and geometric features that classify the distance between the left and right edges of the nose are also index information. Are utilized. The standard comparison pattern of the mouth is specified at the left and right centers of the centerline where the upper and lower lips meet each other, and the distance between the left and right end points of the mouth and the distance between the upper lip chevron and upper and lower lip lines Geometric features representing thickness are also used as index information. In the standard comparison pattern of the jaw, an arbitrary position is specified at the left and right centers of the jaw, the ear is positioned below the auricle, and a geometric feature representing the upper and lower distances of the ear is also used as index information.

In addition, among the plurality of DBs 8 to 16, there are head DB 15 and whisker DB 16 which correspond to the shape of a human face and are not biometric information, and are related to these DBs 15 and 16. The index information stored in the index DB 6 corresponds to the head shape of the sports type, the farmer type, the long hair type, and the sideburns correspond to the sideburns of various shapes.

When the facial feature information corresponding to the non-stored index information is input from the client terminal 1 and the index 3D image search engine 5 cannot find the corresponding index information, the face part corresponding to the projected 3D image search engine 5 is randomly extracted. The star projection 3D image is selected and read from the entire face DB 7 to the sideburn DB 16, and index information corresponding to the selected projection 3D image is read out.

Hereinafter, the operation of the first embodiment to which the present invention configured as described above is applied will be described in detail in parallel with FIG. 3 which is a flowchart of the first embodiment to which the present invention is applied.

First, when the client terminal 1 transmits a plurality of facial feature information necessary for producing a three-dimensional face montage (S301) to the server computer 25, the transmitted plurality of facial feature information is transmitted to the I / O unit (1) of the server computer 25 ( 20), the received plurality of facial feature information is input to the projection 3D image search engine 5 in the server computer 25. Subsequently, the projection 3D image search engine 5 reads each index information corresponding to the plurality of received facial feature information from the index DB 6 (S302) and forms a three-dimensional partial face shape corresponding to the read index information. Is read out from the entire face DB 7 to the sideburn DB 16 (S303).

Further, even when index information corresponding to the plurality of received facial feature information is not found in the step S302, the projection 3D image search engine 5 displays an arbitrary three-dimensional partial face shape in the whole face DB 7 to sideburn DB 16. ) Randomly selected reading (S304).

Then, the projection 3D image retrieval engine 5 projects 3D partial face shapes (full face shape and face part shape) and index information corresponding to three-dimensional partial face shape (face part shape). The data is sent to the image editing unit 17 (S305).

The transmitted three-dimensional partial face shapes are shown in FIGS. 4A (whole face shape) and 4B (face part shape). Fig. 4a shows the whole face in the three-dimensional partial face shape, and Fig. 4b shows the face shape in the three-dimensional partial face shape (ehead-31, left / right eyebrows-32a / 32b, left / right eyes-33a /). 33b, nose-34, in-35, mouth-36, left / right ears-37a / 37b, jaw-38) and FIG. 4C shows the standard comparison pattern of FIG. 4B (brow-41, left / right eyebrows-42a / 42b). , Left / right eyes-43a / 43b, nose-44, in-45, mouth-46, left / right ears-47a / 47b, chin-48).

The projection 3D image editing unit 17 first generates a reference standard comparison pattern for each part from the entire face shape among the transmitted three-dimensional partial face shapes, and then transmits the reference standard comparison pattern to the position information of the generated reference standard comparison pattern. A three-dimensional face montage is produced by combining the three-dimensional partial face shapes by matching the index information (positional information of the standard comparison pattern) of the three-dimensional partial face shape (face partial shape).

The operation of the projection 3D image editing unit 17 for producing the three-dimensional face montage mentioned in S306 will be described in detail in parallel with FIG. 5 as follows.

The projection 3D image editing unit 17 detects the entire face shape (see FIG. 4A) among the three-dimensional partial face shapes transmitted from the projection 3D image search engine 5 (S501), and each face part shape from the detected whole face shape. A reference standard comparison pattern corresponding thereto is generated (S502).

That is, the projection 3D image editing unit 17 finds the point where the pixel value representing the height of the range data is the maximum in the z-axis direction from the detected whole face shape, and obtains an average value when the number of points is several, and returns it to the position of the peak of the nose. It recognizes (S502-1).

Subsequently, as the pixel value is lowered from the apex position of the nose up to the y-axis, the glacial position is recognized at the point where the pixel value is lowered again, and the pixel value corresponding to the pupil (eye) in the left / right direction from the glacial position is rapidly lowered. Recognizes the area as the position of the eye, recognizes the position where the pixel value is the maximum above a certain distance from the position of the eyebrow as the position of the eyebrow, and the center of the forehead of the area corresponding to the average value of the xy axis among the areas above the position of the left / right eyebrows (S502-2).

Next, there is a point where the pixel value decreases rapidly while descending from the apex position of the nose down the y axis, and is recognized as the position where the pulling starts, and then descends again toward the lip by the change of the pixel value. The position of and the position of the jaw are recognized in sequence (S502-3).

Subsequently, a reference standard comparison pattern is generated at each of the recognized face portion shapes. First, the nose generates a reference standard comparison pattern in the center region at each apex, and a horizontal line connecting the centers of the two eyes. At 63 (see FIG. 6), a longitudinal center line (see FIG. 6) is connected to the same distance between the two eyes and the center of the nose, and the forehead, the middle ear, the mouth, and the chin representing the remaining face parts. The reference standard comparison pattern of is generated so as to be positioned above the longitudinal center line, and the horizontal line 62 (see FIG. 6) connecting the centers of both eyes is set parallel to the horizontal line (see Fig. 6). In step S502-4, a reference standard comparison pattern corresponding to both eyebrows is generated.

Subsequently, the projection 3D image editing unit 17 transmits the positional information (71 to 78, see FIG. 6) of the reference standard comparison patterns of the entire face shape generated as described above, and the projection 3D image retrieval engine 5. The positional information (41 to 48, see FIG. 4C) of the standard comparison pattern corresponding to the face portion shape among the three-dimensional partial face shapes thus obtained is combined by a method of overlapping and matching the image layer method (S503). .

Then, after the conversion unit 18 converts the produced three-dimensional face montage into three-dimensional polygon data, the three-dimensional polygon data memory unit 19 temporarily stores the converted three-dimensional polygon data, and three-dimensional The polygon data texture generation unit 21 adds color information to the three-dimensional face montage of the temporarily stored three-dimensional polygon data form (S307).

Subsequently, the inverse transform unit 23 reads the 3D face montage edited as described above from the 3D polygon data memory unit 19 and inversely converts the projected 3D image in the form of range data to the projected 3D image memory unit 24. Store in

In addition, the three-dimensional face montage stored in the projection 3D image memory unit 24 is transmitted to the client terminal 1 by the I / O unit 20 (S308), and the correction request data is then sent from the client terminal 1. Upon reception (YES in S309), the projection 3D image editing unit 17 reads out a 3D face montage that is a projection 3D image from the projection 3D image memory unit 24 (S310).

Subsequently, the projection 3D image editing unit 17 rotates, copies, reduces, enlarges, and positions the read 3D face montage on a 2D plane with a minimum unit of 3D partial face shape in response to the received correction request data. An editing operation such as moving or performing an editing operation is performed to replace an arbitrary three-dimensional partial face shape with another three-dimensional partial face shape in the combined three-dimensional face montage (S311).

The substitution editing operation performed by the projection 3D image editing unit 17 will be described in detail with reference to FIGS. 7A and 7B.

First, an arbitrary three-dimensional face portion ('eye', 'forehead', 'nose', 'mouth', etc. among three-dimensional face montages from the client terminal 1 may be used, but the present embodiment is 'mouth'). Suppose that the correction request data requesting to replace with another three-dimensional face partial shape is received, the projection 3D image retrieval engine 5 receives the three-dimensional partial face shape or 'mouth' corresponding to the previously received correction request data. (Refer to 81, FIG. 7A) is read from the DB 12 and transmitted to the projection 3D image editing unit 17, and the projection 3D image editing unit 17 prepares the projected mouth 81 to produce the projection 3D image. The 3D face montage in the form of a projected 3D image that has been temporarily stored in the memory unit 24 is replaced and edited as shown in FIG. 7B.

The three-dimensional face montage substituted and edited by the projection 3D image editing unit 17 as described above is stored in the three-dimensional polygon data memory unit 19 via the converting unit 18. For reference, the three-dimensional polygon data stored after the substitution editing is stored. The three-dimensional face montage of the shape is as shown in FIG.

In addition, the three-dimensional polygon data editing unit 22 may read and edit the three-dimensional face montage in the form of three-dimensional polygon data stored in the three-dimensional polygon data memory unit 19 in response to the received correction request data. Since the edges of the three-dimensional partial face shape combined by the 3D image editing unit 17 contain discontinuities, smoothly change these discontinuities into continuous lines, and fill in occlusion that cannot be represented by the projected 3D image. An interpolation editing operation can also be performed.

The three-dimensional face montage edited by the projection 3D image editing unit 17 or the 3D polygon data editing unit 22 is stored (S312) in the projection 3D image memory unit 24 via the inverse transform unit 23 and at the same time. It is transmitted to the client terminal 1 via the I / O unit 20 (S308).

Then, the display unit 2 of the client terminal 1 displays the three-dimensional face montage as a three-dimensional image and corrected request data through an input device (not shown) of the client terminal 1 by a user's manipulation. Is re-inputted, the I / O unit 3 of the client terminal 1 repeats the above-described operations of S309 to S312 by transmitting the correction request data to the server computer 25 again.

FIG. 2 is a configuration diagram of a second embodiment to which the present invention is applied, and in FIG. 1, a three-dimensional partial face shape (forehead, eyes and eyebrows, nose, human, mouth, chin, ear, etc.) stored in a plurality of DBs 8 to 16. While it is characterized by this independent division, the three-dimensional partial face shape stored in a plurality of DBs (126 to 130) for the purpose of facilitating the recombination of the face has the head and forehead, eyes and eyebrows, nose, mouth, and chin. There is a different feature to the shape of the face divided horizontally to include each.

That is, FIG. 2 shows a three-dimensional partial face shape (face part shape) having a designated size and position information, as shown in FIG. 9, with head, forehead 91, eyes and eyebrows 92, nose 93, and mouth. (94) and the jaw (95) are respectively stored a plurality of DB (126 ~ 130), the three-dimensional partial face shape (corresponding to the face portion shape) has the same size and location information, so as shown in Figure 1 Does not require an index DB.

The operation of FIG. 2 will be described in detail in parallel with FIG. 11 which is a flowchart of the second embodiment.

First, when a plurality of face feature information necessary for producing a three-dimensional face montage is transmitted from the client terminal 101 to the server computer 125 (S1101), the plurality of face feature information transmitted is the projected 3D image search engine in the server computer 125. 105 is entered.

Subsequently, the projection 3D image retrieval engine 105 has a full face DB 107 and a head and forehead DB 126 to second jaw DB 130 corresponding to the three-dimensional partial face shapes corresponding to the plurality of input facial feature information. Each of them is selected and read (S1102).

Even when the three-dimensional partial face shape corresponding to the plurality of received facial feature information is not found in the DBs 107 and 126 to 130, the projection 3D image search engine 5 returns an arbitrary three-dimensional partial face shape to the DB ( 107, 126 to 130 are randomly selected and read (S1103).

Then, the projection 3D image retrieval engine 5 transmits the selectively read three-dimensional partial face shape (full face shape and face partial shape) to the projection 3D image editing unit 17 (S1104).

The projected 3D image editing unit 117 combines the transmitted 3D partial face shapes (full face shape and face partial shape) to produce a 3D face montage. First, the 3D image editing unit 117 projects the 3D partial face. The shape of the face portion among the shapes is cut out to fit the left and right edges of the whole face shape (S1105).

That is, when the face portion shape is 93 of FIG. 9 including the nose and ears, the face portion shape of the face portion 93 cut out to fit the left and right edges of the entire face shape of FIG. 4A is FIG. 10, and FIG. 10A. Is a three-dimensional partial face shape (face part shape) is a projection 3D image, and FIG. 10B is a three-dimensional partial face shape (face part shape) is three-dimensional polygon data.

The projected 3D image editing unit 117 combines the five face portion shapes cut to fit the left and right edges of the whole face shape in the specified order (S1106) to produce a three-dimensional face montage (S1106). S1112) is performed in the same manner as in the first embodiment.

In addition, an embodiment in which the above-described first and second embodiments are simultaneously implemented may be proposed as the third embodiment, and in this case, the three-dimensional face montage may be more efficiently manufactured in that the above various characteristics are selectively used simultaneously. There is a characteristic.

The present invention is not limited to the above-described embodiments and may be variously modified and changed by those skilled in the art, which are included in the spirit and scope of the present invention as defined in the appended claims. Particularly, in the present embodiment, only the case where the projection surface of the projection 3D image having the form of range data is a plane and the projection 3D image is a front face shape can be replaced, but it is possible to substitute another spatial coordinate system, for example, the projection 3D image. 3D for the back of the human face by changing the Cartesian coordinate system expressing to the cylindrical coordinate system and performing all the above production steps in the cylindrical coordinate system applied to the production of the data and the three-dimensional face montage constituting the respective databases. Face montages can also be easily handled and side finishes can be improved.

As described above, in the present invention, unlike the conventional method of producing a montage by synthesizing the photo, a montage is produced by combining three-dimensional partial face shapes corresponding to each important part of the face, thereby being very close to a human face. Has the effect of producing a three-dimensional face montage.

In addition, the present invention has an effect that can easily create a delicate three-dimensional face montage even if not expert expert in three-dimensional face editing while relying on the judgment of the person using the montage production system rather than the creative ability of the producer.

In addition, the present invention has the effect of maximizing the performance and storage efficiency of the montage production system for handling data by storing the three-dimensional partial face shape in the form of range data.

In addition, the present invention by implementing the three-dimensional partial face shape in the form of a face divided horizontally, there is an effect of improving the montage production speed.

Claims (54)

Face shape storage means for storing a plurality of three-dimensional partial face shapes separately; Selective reading means for selectively reading out a plurality of three-dimensional partial face shapes corresponding to a plurality of face feature information from said face shape storing means; And And a production means for producing a three-dimensional face montage by combining the plurality of selected three-dimensional partial face shapes. The method of claim 1, And index storage means for storing index information corresponding to the facial feature information. The selective reading means reads index information corresponding to a plurality of facial feature information from the index storing means and selectively reads out a three-dimensional partial face shape corresponding to the read index information from the face shape storing means. 3D face montage acquisition system. The method of claim 2, The selected three-dimensional partial face shape is composed of a whole face shape and a face part shape, wherein the index information is position information of a standard comparison pattern in the face part shape, The manufacturing means, A detection unit for detecting a whole face shape among the three-dimensional partial face shapes read out; A generation unit generating a reference standard comparison pattern corresponding to the positional information of the standard comparison pattern in the detected whole face shape; And And a combination unit to match the generated reference standard comparison pattern with the standard comparison pattern to combine the selected three-dimensional partial face shapes. The method according to any one of claims 1 to 3, The three-dimensional partial face shape is a three-dimensional face montage acquisition system, characterized in that the form of range data (range data). The method according to any one of claims 1 to 3, When inputting the correction request data for the produced 3D face montage, the produced 3D face montage is edited in response to the inputted correction request data, and the 3D partial face shape is rotated, copied, reduced, enlarged, And a first editing means for performing at least one editing operation among the position movements or performing an editing operation of replacing an arbitrary three-dimensional partial face shape in the three-dimensional face montage with another three-dimensional partial face shape. 3D face montage acquisition system. The method of claim 4, wherein When inputting the correction request data for the produced 3D face montage, the produced 3D face montage is edited in response to the inputted correction request data, and the 3D partial face shape is rotated, copied, reduced, enlarged, And a first editing means for performing at least one editing operation among the position movements or performing an editing operation of replacing an arbitrary three-dimensional partial face shape in the three-dimensional face montage with another three-dimensional partial face shape. 3D face montage acquisition system. The method according to any one of claims 1 to 3, Editing the produced three-dimensional face montage in response to the inputted correction request data when inputting correction request data for the produced three-dimensional face montage, wherein an edge of an arbitrary three-dimensional partial face shape among the produced three-dimensional face montages is obtained. And a second editing means for smoothing the discontinuous lines into the continuous lines and filling and editing the disjoint regions. The method of claim 4, wherein Editing the produced three-dimensional face montage in response to the inputted correction request data when inputting the correction request data for the produced three-dimensional face montage, wherein an edge of an arbitrary three-dimensional part of the three-dimensional face montage is formed. And a second editing means for smoothing the discontinuous lines into the continuous lines and filling and editing the disjoint regions. The method of claim 5, Editing the produced three-dimensional face montage in response to the inputted correction request data when inputting correction request data for the produced three-dimensional face montage, wherein an edge of an arbitrary three-dimensional partial face shape among the produced three-dimensional face montages is obtained. And a second editing means for smoothing the discontinuous lines into the continuous lines and filling and editing the disjoint regions. The method of claim 6, Editing the produced three-dimensional face montage in response to the inputted correction request data when inputting correction request data for the produced three-dimensional face montage, wherein an edge of an arbitrary three-dimensional partial face shape among the produced three-dimensional face montages is obtained. And a second editing means for smoothing the discontinuous lines into a continuous line and filling and editing the occlusion region. The method according to any one of claims 1 to 3, Wherein the three-dimensional partial face shape, the three-dimensional face montage acquisition system, characterized in that each part of the face including the forehead, eyes and eyebrows, nose, patience, mouth, chin, ears are independently divided. The method of claim 4, wherein Wherein the three-dimensional partial face shape, the three-dimensional face montage acquisition system, characterized in that each part of the face including the forehead, eyes and eyebrows, nose, patience, mouth, chin, ears are independently divided. The method of claim 5, Wherein the three-dimensional partial face shape, the three-dimensional face montage acquisition system, characterized in that each part of the face including the forehead, eyes and eyebrows, nose, patience, mouth, chin, ears are independently divided. The method of claim 6, The three-dimensional partial face shape, the three-dimensional face montage acquisition system, characterized in that each part of the face including the forehead, eyes and eyebrows, nose, patience, mouth, chin, ears are independently divided. The method of claim 7, wherein Wherein the three-dimensional partial face shape, the three-dimensional face montage acquisition system, characterized in that each part of the face including the forehead, eyes and eyebrows, nose, patience, mouth, chin, ears are independently divided. The method of claim 8, Wherein the three-dimensional partial face shape, the three-dimensional face montage acquisition system, characterized in that each part of the face including the forehead, eyes and eyebrows, nose, patience, mouth, chin, ears are independently divided. The method of claim 9, Wherein the three-dimensional partial face shape, the three-dimensional face montage acquisition system, characterized in that each part of the face including the forehead, eyes and eyebrows, nose, patience, mouth, chin, ears are independently divided. The method of claim 10, The three-dimensional partial face shape, the three-dimensional face montage acquisition system, characterized in that each part of the face including the forehead, eyes and eyebrows, nose, patience, mouth, chin, ears are independently divided. The method of claim 1, The three-dimensional face montage is a three-dimensional face montage acquisition system, characterized in that the shape for each part of the face divided horizontally so that the head and forehead, eyes and eyebrows, nose, mouth, chin, respectively. The method of claim 19, The selected three-dimensional partial face shape is composed of the entire face shape and the face part shape, The manufacturing means, A cutting unit which cuts the selected face portion shape to fit left and right edges of the entire face shape; And And a combination unit for combining the cut-out face portions in a specified order. The method of claim 19 or 20, The three-dimensional partial face shape is a three-dimensional face montage acquisition system, characterized in that the form of range data (range data). The method of claim 19 or 20, When inputting the correction request data for the produced 3D face montage, the produced 3D face montage is edited in response to the inputted correction request data, and the 3D partial face shape is rotated, copied, reduced, enlarged, And a first editing means for performing at least one editing operation among the position movements or performing an editing operation of replacing an arbitrary three-dimensional partial face shape in the three-dimensional face montage with another three-dimensional partial face shape. 3D face montage acquisition system. The method of claim 21, When inputting the correction request data for the produced 3D face montage, the produced 3D face montage is edited in response to the inputted correction request data, and the 3D partial face shape is rotated, copied, reduced, enlarged, And a first editing means for performing at least one editing operation among the position movements or performing an editing operation of replacing an arbitrary three-dimensional partial face shape in the three-dimensional face montage with another three-dimensional partial face shape. 3D face montage acquisition system. The method of claim 19 or 20, Editing the produced three-dimensional face montage in response to the inputted correction request data when inputting correction request data for the produced three-dimensional face montage, wherein an edge of an arbitrary three-dimensional partial face shape among the produced three-dimensional face montages is obtained. And a second editing means for smoothing the discontinuous lines into the continuous lines and filling and editing the disjoint regions. The method of claim 21, Editing the produced three-dimensional face montage in response to the inputted correction request data when inputting correction request data for the produced three-dimensional face montage, wherein an edge of an arbitrary three-dimensional partial face shape among the produced three-dimensional face montages is obtained. And a second editing means for smoothing the discontinuous lines into the continuous lines and filling and editing the disjoint regions. The method of claim 22, Editing the produced three-dimensional face montage in response to the inputted correction request data when inputting correction request data for the produced three-dimensional face montage, wherein an edge of an arbitrary three-dimensional partial face shape among the produced three-dimensional face montages is obtained. And a second editing means for smoothing the discontinuous lines into the continuous lines and filling and editing the disjoint regions. The method of claim 23, wherein Editing the produced three-dimensional face montage in response to the inputted correction request data when inputting correction request data for the produced three-dimensional face montage, wherein an edge of an arbitrary three-dimensional partial face shape among the produced three-dimensional face montages is obtained. And a second editing means for smoothing the discontinuous lines into the continuous lines and filling and editing the disjoint regions. A selective reading step of selectively reading out a plurality of three-dimensional partial face shapes corresponding to the plurality of face feature information from a face shape database; And 3. The method of claim 3, further comprising a step of manufacturing a three-dimensional face montage by combining the plurality of selected three-dimensional partial face shapes. The method of claim 28, In the selective reading step, the index information corresponding to the plurality of facial feature information is read from an index database, and the three-dimensional partial face shape corresponding to the read index information is selectively read from the face shape database. How to obtain a dimensional face montage. The method of claim 29, The selected three-dimensional partial face shape is composed of a whole face shape and a face part shape, wherein the index information is position information of a standard comparison pattern in the face part shape, The production step, A detection step of detecting a whole face shape among the three-dimensional partial face shapes read out; Generating a reference standard comparison pattern corresponding to the positional information of the standard comparison pattern in the detected whole face shape; And And combining the selected read-out three-dimensional partial face shapes by matching the generated reference standard comparison pattern with the standard comparison pattern. 3. The method according to any one of claims 28 to 30, The three-dimensional partial face shape is a three-dimensional face montage acquisition method characterized in that the form of range data (range data). The method according to any one of claims 28 to 30, When inputting the correction request data for the produced 3D face montage, at least one or more editing operations of rotating, copying, reducing, enlarging, and moving the 3D partial face shape as a minimum unit are performed, or any of the 3D face montages. And a first editing step of performing an editing operation of replacing the three-dimensional partial face shape with another three-dimensional partial face shape. The method of claim 31, wherein When inputting the correction request data for the produced 3D face montage, at least one or more editing operations of rotating, copying, reducing, enlarging, and moving the 3D partial face shape as a minimum unit are performed, or any of the 3D face montages. And a first editing step of performing an editing operation of replacing the three-dimensional partial face shape with another three-dimensional partial face shape. The method according to any one of claims 28 to 30, Upon inputting the correction request data for the produced three-dimensional face montage, the discontinuity line of the edge of an arbitrary three-dimensional partial face shape among the produced three-dimensional face montages is smoothly changed into a continuous line and an occlusion region is included. Method of obtaining a three-dimensional face montage, characterized in that further comprises a second editing step of filling and editing. The method of claim 31, wherein Upon inputting the correction request data for the produced three-dimensional face montage, the discontinuity line of the edge of an arbitrary three-dimensional partial face shape among the produced three-dimensional face montages is smoothly changed into a continuous line and an occlusion region is included. Method of obtaining a three-dimensional face montage, characterized in that further comprises a second editing step of filling and editing. The method of claim 32, Upon inputting the correction request data for the produced three-dimensional face montage, the discontinuity line of the edge of an arbitrary three-dimensional partial face shape among the produced three-dimensional face montages is smoothly changed into a continuous line and an occlusion region is included. Method of obtaining a three-dimensional face montage, characterized in that further comprises a second editing step of filling and editing. The method of claim 33, Upon inputting the correction request data for the produced three-dimensional face montage, the discontinuity line of the edge of an arbitrary three-dimensional partial face shape among the produced three-dimensional face montages is smoothly changed into a continuous line and an occlusion region is included. Method of obtaining a three-dimensional face montage, characterized in that further comprises a second editing step of filling and editing. The method according to any one of claims 28 to 30, The three-dimensional partial face shape is a three-dimensional face montage acquisition method, characterized in that each part of the face including the forehead, eyes and eyebrows, nose, patience, mouth, chin, ears are independently divided. The method of claim 31, wherein The three-dimensional partial face shape is a three-dimensional face montage acquisition method, characterized in that each part of the face including the forehead, eyes and eyebrows, nose, patience, mouth, chin, ears are independently divided. The method of claim 32, The three-dimensional partial face shape is a three-dimensional face montage acquisition method, characterized in that each part of the face including the forehead, eyes and eyebrows, nose, patience, mouth, chin, ears are independently divided. The method of claim 33, The three-dimensional partial face shape is a three-dimensional face montage acquisition method, characterized in that each part of the face including the forehead, eyes and eyebrows, nose, patience, mouth, chin, ears are independently divided. The method of claim 34, The three-dimensional partial face shape is a three-dimensional face montage acquisition method, characterized in that each part of the face including the forehead, eyes and eyebrows, nose, patience, mouth, chin, ears are independently divided. The method of claim 35, wherein The three-dimensional partial face shape is a three-dimensional face montage acquisition method, characterized in that each part of the face including the forehead, eyes and eyebrows, nose, patience, mouth, chin, ears are independently divided. The method of claim 36, The three-dimensional partial face shape is a three-dimensional face montage acquisition method, characterized in that each part of the face including the forehead, eyes and eyebrows, nose, patience, mouth, chin, ears are independently divided. The method of claim 37, wherein The three-dimensional partial face shape is a three-dimensional face montage acquisition method, characterized in that each part of the face including the forehead, eyes and eyebrows, nose, patience, mouth, chin, ears are independently divided. The method of claim 28, The three-dimensional partial face shape is a three-dimensional face montage acquisition method characterized in that the shape for each part divided horizontally to the head and forehead, eyes and eyebrows, nose, mouth, jaw respectively included. The method of claim 46, The selected three-dimensional partial face shape is composed of the entire face shape and the face part shape, The production step, Cutting the selected out-of-face part shape to fit left and right edges of the entire face shape; And And a combining step of combining the cut face portions in a specified order. 48. The method of claim 46 or 47, The three-dimensional partial face shape is a three-dimensional face montage acquisition method characterized in that the form of range data (range data). 48. The method of claim 46 or 47, When inputting the correction request data for the produced 3D face montage, at least one or more editing operations of rotating, copying, reducing, enlarging, and moving the 3D partial face shape as a minimum unit are performed, or any of the 3D face montages. And a first editing step of performing an editing operation of replacing the three-dimensional partial face shape with another three-dimensional partial face shape. The method of claim 48, When inputting the correction request data for the produced 3D face montage, at least one or more editing operations of rotating, copying, reducing, enlarging, and moving the 3D partial face shape as a minimum unit are performed, or any of the 3D face montages. And a first editing step of performing an editing operation of replacing the three-dimensional partial face shape with another three-dimensional partial face shape. 48. The method of claim 46 or 47, Upon inputting the correction request data for the produced three-dimensional face montage, the discontinuity line of the edge of an arbitrary three-dimensional partial face shape among the produced three-dimensional face montages is smoothly changed into a continuous line and an occlusion region is included. Method of obtaining a three-dimensional face montage, characterized in that it further comprises a second editing step of filling and editing. The method of claim 48, Upon inputting the correction request data for the produced three-dimensional face montage, the discontinuity line of the edge of an arbitrary three-dimensional partial face shape among the produced three-dimensional face montages is smoothly changed into a continuous line and an occlusion region is included. Method of obtaining a three-dimensional face montage, characterized in that it further comprises a second editing step of filling and editing. The method of claim 49, Upon inputting the correction request data for the produced three-dimensional face montage, the discontinuity line of the edge of an arbitrary three-dimensional partial face shape among the produced three-dimensional face montages is smoothly changed into a continuous line and an occlusion region is included. Method of obtaining a three-dimensional face montage, characterized in that it further comprises a second editing step of filling and editing. 51. The method of claim 50, Upon inputting the correction request data for the produced three-dimensional face montage, the discontinuity line of the edge of an arbitrary three-dimensional partial face shape among the produced three-dimensional face montages is smoothly changed into a continuous line and an occlusion region is included. Method of obtaining a three-dimensional face montage, characterized in that it further comprises a second editing step of filling and editing.