GB2330973A

GB2330973A - Lip image signal reconstruction

Info

Publication number: GB2330973A
Application number: GB9813898A
Authority: GB
Inventors: Chang-Buhm Lee
Original assignee: Daewoo Electronics Co Ltd
Current assignee: WiniaDaewoo Co Ltd
Priority date: 1997-10-29
Filing date: 1998-06-26
Publication date: 1999-05-05
Also published as: GB9813898D0; KR100278065B1; KR19990034391A

Abstract

An apparatus for use in a face image signal coding system to reconstruct a lip image signal(LIS) included in a face image signal includes, a LIS extraction circuit 20 and a fundamental LIS providing circuit 10 for providing a LIS having N number of fundamental edge points (FEP's) corresponding to the edge points (EP's) of a contour signal of a lip image. A feature point detection circuit 30 detects K number of feature points (FP's) and K number of corresponding fundamental feature points (FFP's) from the EP's and the FEP's, respectively and a modification parameter calculation circuit 40 calculates K number of modification parameters (MP's) based on the K number of FP's and the K number of FFP's wherein an MP represents the difference between corresponding FP's and FFP's. An edge point motion estimation circuit 50 performs motion estimation on the N number of EP's based on the K number of MP's and the N number of FEP's to obtain N number of reconstructed edge points(REP's) corresponding to the number of FEP's thereby supplying a reconstructed contour signal for the LIS having the N number of REP's.

Description

LIP IMAGE SIGNAL RECONSTRUCTION METHOD AND APPARATUS The present invention relates to a face image signal coding system; and, more particularly, to a lip image signal reconstruction method and apparatus for use in the face image signal coding system.

In recent years, the so called model based coding technique" is gaining wide acceptance in various fields such as video phone, video conference, virtual reality and computer graphics industry and the like.

Generally, the physical image of a speaker shown in the video phone or the video conference screen is the upper body of the speaker. Especially, in the video phone or the video conference system, the majority of video data therefor pertains to the face of the speaker. For this reason, therefore, other data than the face of the speaker, e.g., background image data can be ignored and only image data for the face is transmitted, thereby greatly reducing the amount of transmission data thereof.

Namely, in the video phone or the video conference system, the main part of the information transmitted and processed is that for the face image of the speaker, especially for the lip image of the speaker since the picture of lip movement during the speech is crucial to recognizing the characteristics of speech and emotional expressions of the speaker.

Therefore, in a video phone system, a lip image signal included in a face image signal and a corresponding voice signal are simultaneously coded and transmitted. Further, in the conventional lip image signal coding process, normally a composed or reconstructed lip image signal in lieu of a corresponding lip image signal is coded and then transmitted to further reduce the amount of transmission data thereof.

Referring to Fig. 1, there is shown a block diagram of a conventional lip image signal reconstruction apparatus 8 for use in a face image signal coding system. The apparatus 8 comprises a lip image signal extraction circuit 2, a fundamental lip image signal providing circuit 3 and an edge point motion detection circuit 5.

First, a face image signal including a lip image signal is fed to the lip image signal extraction circuit 2 from a face image signal supplying circuit in the face image signal coding system. The lip image signal has a contour signal containing N number of inter-connected edge points representing the position and shape of a lip image in a face image and a texture signal, wherein N is a predetermined positive integer. It should be noted here that the texture signal has luminance data with or without chrominance data for the lip image.

Meanwhile, the fundamental lip image signal providing circuit 3 provides a fundamental lip image signal to the edge point motion detection circuit 5 through a line L13. The fundamental lip image signal has a fundamental contour signal containing N number of inter-connected fundamental edge points thereon corresponding to the edge points of the contour signal, respectively.

The fundamental contour signal represents the position and shape of a fundamental lip image. It should be noted here that the fundamental lip image signal is usually stored in a memory within the fundamental lip image signal providing circuit 3 beforehand.

Referring to Fig. 2, there is represented a schematic diagram revealing an upper fundamental lip image having an outer fundamental contour 110 and an inner fundamental contour 115; and an outer contour 120 of a lip image for use in explaining a lip image signal reconstruction process performed by the conventional lip image signal reconstruction apparatus 8 shown in Fig. 1.

In Fig. 2, the outer contour 110 and the inner contour 120 are expressed in solid lines having fundamental edge points represented by small black circles thereon, respectively and the outer contour 120 is expressed in a dotted line having edge points represented by small black circles thereon.

In Fig. 2, the hatched region represents a texture region of the upper fundamental lip image and the boundary of the hatched region represents the inner contour 115 and the outer contour 110 of the fundamental lip image.

The lip image signal extraction circuit 2, by employing a predetermined known lip image signal extraction method, extracts the lip image signal from the face image signal to thereby provide the contour signal and the texture signal of the lip image signal to the edge point motion detection circuit 5 via a line Lll.

The edge point motion detection circuit 5 detects motion vectors (MV's) for all of the edge points on the contour signal by using the N number of edge points on the contour of the lip image signal inputted thereto via the line Lll and the N number of corresponding fundamental edge points on the fundamental contour of the fundamental lip image signal inputted thereto via the line L13. It should be noted here that a MV represents a displacement between an edge point and a corresponding fundamental edge point.

For example, referring back to Fig. 2, the center point of a lip image including the upper lip image having the outer contour 120 is first matched with the center point of a fundamental lip image including the upper fundamental lip image having the outer fundamental contour 110 and the inner fundamental contour 115 as an origin of a two dimensional xy-plane as represented therein. As shown in Fig. 2, coordinates of a fundamental edge point A3 and an edge point B3 on the xy-plane are (a31a3') and (b3,b3'), respectively.

Hence, a corresponding MV3 for the edge point B3 is represented by (b3-a3,b3'-a3').

And then, the edge point motion detection circuit 5 provides a reconstructed contour signal for the lip image signal and the texture signal to a transmitter (not shown) for the transmission thereof, the reconstructed contour signal including N number of reconstructed edge points, wherein each of the reconstructed edge points is obtained by moving a corresponding fundamental edge point by a corresponding MV.

The conventional lip image signal reconstruction or composition apparatus, however, as N increases, needs considerable amount of memory capacity and calculation time to detect corresponding MV's for all edge points of a lip image signal, thereby rendering the real time processing of a face image signal together with a corresponding audio signal, i.e., a voice signal difficult.

It is, therefore, an object of the present invention to provide a lip image signal reconstruction method and apparatus for use in a face image signal coding system to enhance the reconstruction efficiency thereof.

According to the present invention, there is provided an apparatus for reconstructing a lip image signal included in a face image signal, the lip image signal having a contour signal containing N number of inter-connected edge points representing the position and shape of a lip image in a face image and a texture signal having pixels, wherein N is a predetermined positive integer, the apparatus comprising: a fundamental lip image signal providing circuit to provide a fundamental lip image signal having a fundamental contour signal containing N number of inter-connected fundamental edge points (FEP's) corresponding to the edge points of the contour signal, respectively, wherein the fundamental contour signal represents the position and shape of a fundamental lip image; a lip image signal extraction circuit for extracting the lip image signal from the face image signal to thereby provide the contour signal and the texture signal of the lip image signal by employing a predetermined lip image signal extraction method; a feature point detection circuit for detecting K number of feature points (FP's) from the edge points and K number of corresponding fundamental feature points (FFP's) from the fundamental edge points by using a predetermined feature point detection method with K being a predetermined positive integer less than N, wherein a FP and a corresponding FFP are points essential to define the motion between the lip image and the fundamental lip image, respectively; a modification parameter calculation circuit for calculating K number of modification parameters (MP's) based on the K number of FP's and the K number of FFP's, wherein each of the MP's represents a displacement between a FP and a corresponding FFP; and an edge point motion estimation circuit for performing motion estimation on the N number of the edge points of the contour signal based on the K number of MP's and the N number of fundamental edge points of the fundamental contour signal to obtain N number of reconstructed edge points (REP's) corresponding to the N number of FEP's, thereby supplying a reconstructed contour signal for the lip image signal having the N number of REP's.

According to a second aspect of the present invention there is provided a method for reconstructing a lip image signal included in a face image signal, the lip image signal having a contour signal containing N number of inter-connected edge points representing the position and shape of a lip image in a face image and a texture signal, wherein N is a predetermined positive integer, the method comprising the steps of: (a) providing a fundamental lip image signal having a fundamental contour signal containing N number of interconnected fundamental edge points (FEP's) corresponding to the edge points of the contour signal, respectively, wherein the fundamental contour signal represents the position and shape of a fundamental lip image; (b) extracting the lip image signal from the face image signal to thereby provide the contour signal and the texture signal of the lip image signal by employing a predetermined lip image signal extraction method; (c) detecting K number of feature points (FP's) from the edge points and K number of corresponding fundamental feature points (FFP's) from the fundamental edge points by using a predetermined feature point detection method with K being a predetermined positive integer less than N, wherein a FP and a corresponding FFP are points essential to define the motion between the lip image and the fundamental lip image, respectively; (d) calculating K number of modification parameters (MP's) based on the K number of FP's and the K number of FFP's, wherein each of the MP's represents a displacement between a FP and a corresponding FFP; and (e) performing motion estimation on the N number of the edge points of the contour signal based on the K number of MP's and the N number of fundamental edge points of the fundamental contour signal to obtain N number of respective reconstructed edge points (REP's) corresponding to the N number of FEP's, thereby supply a reconstructed contour signal for the lip image signal having the N number of The above and other objects and features of the present invention will become apparent from the following description of preferred embodiments given in conjunction with the accompanying drawings, in which: Fig. 1 shows a block diagram of a conventional lip image signal reconstruction apparatus for use in a face image signal coding system; Fig. 2 represents a schematic diagram revealing an upper fundamental lip image having an outer fundamental contour and an inner fundamental contour; and an outer contour of a lip image for use in explaining a lip image signal reconstruction process performed by the conventional lip image signal reconstruction apparatus shown in Fig. 1; Fig. 3 illustrates a block diagram of a lip image signal reconstruction apparatus for use in a face image signal coding system in accordance with the present invention; Figs. 4A and 4B present schematic diagrams revealing a fundamental lip image having a fundamental contour containing a plurality of fundamental edge points and a lip image having a contour containing a plurality of edge points, respectively; Figs. 5A to 5D depict graphs of functions used to obtain reconstructed edge points (REP's); and Fig. 6 exhibits a schematic diagram for use in explaining a method to obtain a REP in accordance with the present invention.

Referring to Fig. 3, there is illustrated a block diagram of a lip image signal reconstruction apparatus 200 for use in a face image signal coding system in accordance with the present invention.

The apparatus 200 comprises a lip image signal extraction circuit 20, a fundamental lip image signal providing circuit 10, a feature point detection circuit 30, a modification parameter calculation circuit 40, an edge point motion estimation circuit 50 and a data formatting circuit 60.

First, a face image signal including a lip image signal is fed to the lip image signal extraction circuit 20 from a face image signal supplying circuit (not shown) in the face image signal coding system. The lip image signal has a contour signal containing N number of inter-connected edge points representing the position and shape of a lip image in a face image and a texture signal, wherein N is a predetermined positive integer.

Meanwhile, the fundamental lip image signal providing circuit 10 provides a fundamental lip image signal to the feature point detection circuit 30 and the edge point motion estimation circuit 50 through a line L21. The fundamental lip image signal has a fundamental contour signal containing N number of inter-connected fundamental edge points corresponding to the edge points of the contour signal, respectively. The fundamental contour signal represents the position and shape of a fundamental lip image. It should be noted here that the fundamental lip image signal is stored in a memory (not shown) within the fundamental lip image signal providing circuit 10.

The lip image signal extraction circuit 20, by employing a predetermined lip image signal extraction method, extracts the lip image signal from the face image signal to thereby provide the contour signal of the lip image signal to the feature point detection circuit 30 via a line L22 and the texture signal of the lip image signal to the data formatting circuit 60 via a line L23.

The feature point detection circuit 30 detects K number of feature points (FP's) from the edge points within the contour signal inputted thereto via the line L22 and at the same time detects K number of corresponding fundamental feature points (FFP's) from the fundamental edge points within the fundamental contour signal of the fundamental lip image signal fed thereto through the line L21, wherein K is a predetermined positive integer less than N. And the feature point detection circuit 30 transmits the FP's and the FFP's to the modification parameter calculation circuit 40 through a line L24.

It should be noted that the feature point detection circuit 30 utilizes a predetermined feature point detection method in detecting FP's and FFP's, wherein a FP and a corresponding FFP are points essential to define the motion between the lip image and the fundamental lip image, respectively.

In accordance with a preferred embodiment of the present invention, the K number of FFP's have 4 FFP's (FFP1, FFP2, FFP3 and FFP4) with FFP1 being a fundamental edge point located at a horizontal right-end position in the fundamental lip image; FFP2 being a fundamental edge point located at a horizontal left-end position of the fundamental lip image; FFP3 being a fundamental edge point located at an intersection of a vertical center line of the fundamental lip image and a lower contour of the upper fundamental lip image of the fundamental lip image; and FFP4 being a fundamental edge point located at an intersection of the vertical center line of the fundamental lip image and an upper contour of the lower fundamental lip image of the fundamental lip image.

And the K number of FP's have 4 FP's (FP1, FP2, FP3 and FP4) with FPl being an edge point located at a horizontal right-end position in the lip image; FP2 being an edge point located at a horizontal left-end position of the lip image; FP3 being an edge point located at an intersection of the vertical center line of the lip image and a lower contour of the upper lip image of the lip image; and FP4 being an edge point located at an intersection of the vertical center line of the lip image and an upper contour of the lower lip image of the lip image.

Referring to Fig. 4A, there is presented a schematic diagram revealing a fundamental lip image 100 having a fundamental contour containing a plurality of fundamental edge points thereon. And, in Fig. 4A, points A, B, C and D represent FFP1, FFP2, FFP3 and FFP 4 detected from the fundamental edge points thereon, respectively. The fundamental lip image 100 includes an fundamental upper lip image and a lower fundamental lip image thereof.

In Fig. 4A, the hatched region represents a texture region of the fundamental lip image 100 and the boundary of the hatched region represents a fundamental contour of the fundamental lip image 100. And coordinates of the FFP's A, B, C and D in the xy-plane in Fig. 4A are (a,0), (b,0), (0,c) and (0,d) , respectively.

Referring to Fig. 4B, there is presented a schematic diagram revealing a lip image 150 having a contour containing a plurality of edge points thereon. And, in Fig. 4B, points A', B', C' and D' represent FPl, FP2, FP3 and FP4 detected from the edge points thereon, respectively. The lip image 150 includes an upper lip image and a lower lip image thereof, And in Fig. 4B, the hatched region represents a texture region of the lip image 150 and the boundary of the hatched region represents a contour of the lip image 150. And coordinates of the FP's A', B', C' and D' in the xy-plane in Fig. 4B are (a',0), (b',0) , (O,c') and (0,d'), respectively.

It should be noted that the coordinates of FFP's of the fundamental lip image 100 and those of FP's of lip image 150 are represented under the assumption that the center point of the fundamental lip image 100 is matched with the center point of the lip image 150 as an origin of a two dimensional xyplane as shown in Fig 4B. From now on, referring to Figs.

4A and 4B, the operation of the apparatus 200 will be described in detail.

The modification parameter calculation circuit 40 calculates K number of modification parameters (MP's) based on the K number of FP's and the K number of FFP's, wherein each of the MP's represents a displacement between an FP and a corresponding FFP. And then, the K number of MP's are fed to the edge point motion estimation circuit 50 through a line L25.

In accordance with a preferred embodiment of the present invention, the MP's have 4 MP's (MP1, MP2, MP3 and MP4) with MPi being a displacement from FFPi to FPi, wherein i is a positive integer ranging from 1 to 4.

In detail, referring to Fig. 4B, the MP1 as a vector can be represented by (a' -a, 0) since the MP1 is a displacement from the FFP A to FP A'. And the MP2, MP3 and MP4 as vectors can be represented by (b'-b,0), (0,c'-c) and (0, d-d), respectively.

The edge point motion estimation circuit 50 performs motion estimation on the N number of edge points of the contour signal based on the K number of MP's fed thereto through the line L25 and the N number of fundamental edge points (FEP's) within the fundamental contour signal of the fundamental lip image signal inputted thereto via the line L21 to obtain N number of respective reconstructed edge points (REP's) corresponding to the N number of FEP's.

The edge point motion estimation circuit 50 obtains the REP's by moving the FEP's in accordance with respective predetermined corresponding equations, thereby supplying a reconstructed contour signal for the lip image signal having the N number of REP's to the data formatting circuit 60 through a line L26.

For example, K number of MP's, e.g., the MP1 to MP4 are fed to the edge point motion estimation circuit 50 via the line L25. And all of the FEP's within the fundamental contour signal of the fundamental lip image signal 100 are fed to the edge point motion estimation circuit 50 via the line L21.

In accordance with a preferred embodiment of the present invention, the edge point motion estimation circuit 50 obtains respective REP's by moving respective corresponding FEP's in accordance with the predetermined respective equations, e.g., the following respective equations (Eq. 1, Eq. 2, Eq. 3, Eq. 4): x' = x + #MP1#z(x) ....... (Eq. 1) (condition: x#0), x' = x + #MP2#f(x) ....... (Eq. 2) (condition: x#0), y' = y + #MP3#h(y) ....... (Eq. 3) (condition: y#0), y' = y + #MP4#g(y) ....... (Eq. 4) (condition: y#0).

It should be noted here that (x',y') and (x,y) in the equations Eq. 1 to Eq. 4 represent xy-coordinates of a REP and a corresponding FEP, respectively, under the condition that the center point of the lip image is matched with the center point of the fundamental lip image as an origin of a two dimensional xy-plane.

And it also should be noted that functions y=z(x), y=f(x), y=h (x) and y=g (x) satisfy the following boundary conditions on the xy-plane: z(0)=0, z(a)=1, f(0)=0, f(b)=l, h(d)=0 and g(c)=0 under the assumption that xy-coordinates of FFP1, FFP2, FFP3 and FFP4 are (a,0), (b, 0), (c,0) and (d,0), respectively.

In accordance with a preferred embodiment of the present invention, the functions z (x), f (x), g (x) and h (x) are defined by following equations: x(x) = x, a 1 f(x) = x, b

(condi tion : ysd), (condi tion : yzc), wherein none of a, b, c and d are zero's.

As can be easily understood in the above description, z(x) and f(x) represent linear functions, respectively. And h(y) and g(y) represent hyperbolic functions, respectively.

Referring to Figs. SA to 5D, there are depicted graphs of the functions described above used to obtain the REP's.

In accordance with a preferred embodiment of the present invention, a can be regarded as equal to b. This is due to the fact that in general, a lip image in a human face image of a speaker is symmetrical.

Fig. 6 exhibits a schematic diagram for use in explaining a method to obtain a REP in accordance with the present invention. In Fig. 6, there are exhibited a fundamental edge point J on an outer fundamental contour 610 of an upper fundamental lip image and a reconstructed edge point R on an outer contour 620 of a reconstructed upper lip image.

From now on, referring to Fig. 6, an exemplary process of obtaining a REP in accordance with a preferred embodiment of the present invention will be described.

For example, if the coordinates of J and R are (jl,j2) and (rl,r2), respectively, and MPl=3.0, MP2=3.0, MP3=0.3, MP4=0.4, jl=1.0, j2=4.0, a=10, b=10, c=0.5 and d=0.5, then rl and r2 are obtained as followings: r1 = 1.0 + #3.0####1.0 = 1.3,

The data formatting circuit 60 combines the reconstructed contour signal inputted thereto through the line L26 with the texture signal inputted thereto through the line L23, thereby supplying a reconstructed lip image signal to a transmitter (not shown) for the transmission thereof.

As described above, in accordance with the lip image signal reconstruction method and apparatus of the present invention, the lip image signal is effectively reconstructed by employing a modification parameter calculation method and an edge point motion estimation technique such as described above, thereby decreasing the reconstruction time of the lip image signal and enhancing the reconstruction efficiency thereof.

While the present invention has been shown and described with reference to the particular embodiments, it will be apparent to those skilled in the art that many changes and modifications may be made without departing from the scope of the invention as defined in the appended claims

Claims

What is claimed is: 1. An apparatus for reconstructing a lip image signal included in a face image signal, the lip image signal having a contour signal containing N number of inter-connected edge points representing the position and shape of a lip image in a face image and a texture signal, wherein N is a predetermined positive integer, the apparatus comprising: means for providing a fundamental lip image signal having a fundamental contour signal containing N number of interconnected fundamental edge points (FEP's) corresponding to the edge points of the contour signal, respectively, wherein the fundamental contour signal represents the position and shape of a fundamental lip image; means for extracting the lip image signal from the face image signal to thereby provide the contour signal and the texture signal of the lip image signal by employing a predetermined lip image signal extraction method; means for detecting K number of feature points (FP's) from the edge points and K number of corresponding fundamental feature points (FFP's) from the fundamental edge points by using a predetermined feature point detection method with K being a predetermined positive integer less than N, wherein a FP and a corresponding FFP are points essential to define the motion between the lip image and the fundamental lip image, respectively; means for calculating K number of modification parameters (MP's) based on the K number of FP's and the K number of FFP's, wherein each of the MP's represents a displacement between a FP and a corresponding FFP; and means for performing motion estimation on the N number of the edge points of the contour signal based on the K number of MP's and the N number of fundamental edge points of the fundamental contour signal to obtain N number of respective reconstructed edge points (REP's) corresponding to the N number of FEP's, thereby supplying a reconstructed contour signal for the lip image signal having the N number of REP's.
2. A apparatus according to claim 1, wherein said K number of FFP's have 4 FFP's (FFP1, FFP2, FFP3 and FFP4) with FFP1 being a fundamental edge point located at a horizontal right-end position in the fundamental lip image; FFP2 being a fundamental edge point located at a horizontal left-end position of the fundamental lip image; FFP3 being a fundamental edge point located at an intersection of a vertical center line of the fundamental lip image and a lower contour of an upper fundamental lip image of the fundamental lip image; and FFP4 being a fundamental edge point located at an intersection of a vertical center line of the fundamental lip image and an upper contour of a lower fundamental lip image of the fundamental lip image.
3. Apparatus according to claim 1 or 2, , wherein said K number of FP's have 4 FP's (FPl, FP2, FP3 and FP4) with FP1 being an edge point located at a horizontal right-end position in the lip image; FP2 being an edge point located at a horizontal left-end position of the lip image; FP3 being an edge point located at an intersection of a vertical center line of the lip image and a lower contour of an upper lip image of the lip image; and FP4 being an edge point located at an intersection of a vertical center line of the lip image and an upper contour of the lower lip image of the lip image.
4. Apparatus according to any of claims 1 tc3,wherein said K number of MP's have 4 MP's (MP1, MP2, MP3 and MP4) with MPi being a displacement from FFPi to FPi, wherein i is a positive integer ranging from 1 to 4.
5. Apparatus according to any one of clairnslto4,wherein said motion estimation means obtains respective REP's by moving respective corresponding FEP's in accordance with the following respective equations (Eq. 1 to Eq. 4) x' = x + #MP1#z(x) ....... (Eq. 1) (condition: x#0), x' = x + #MP2#f(x) ....... (Eq. 2) (condition: x#0), y' = y + #MP3#h(y) ....... (Eq. 3) (condition: y#0), y' = y + #MP4#g(y) ....... (Eq. 4) (condition: y#0) wherein (x',y') and (x,y) in the equations Eq. 1 to Eq. 4 represent xy-coordinates of a REP and a corresponding FEP, respectively, under the condition that the center point of the lip image is matched with the center point of the fundamental lip image as an origin of a two dimensional xy-plane; and functions y=z (x), y=f(x), y=h (x) and y=g (x) satisfy the following boundary conditions on the xy-plane: z(0)=0, z(a)=1, f(o)=o, f(b)=l, h(d)=0 and g(c)=0 under the assumption that xy-coordinates of FFP1, FFP2, FFP3 and FFP4 are (a,0), (b,0), (c,0) and (d,0), respectively.
6. The apparatus according to claim 5, wherein the functions z(x), f(x), g (x) and h(x) are expressed as following equations: z(x) = x, a 1 f(x) = x,

(condition:y#d), (condition:y#c), wherein none of a, b, c and d are zero's.
7. Apparatus according to any one of the preceding claims, further comprising means for combining the reconstructed contour signal with the texture signal to thereby supply a reconstructed lip image signal.
8. A method for reconstructing a lip image signal included in a face image signal, the lip image signal having a contour signal containing N number of inter-connected edge points representing the position and shape of a lip image in a face image and a texture signal, wherein N is a predetermined positive integer, the method comprising the steps of: (a) providing a fundamental lip image signal having a fundamental contour signal containing N number of inter-connected fundamental edge points (FEP's) corresponding to the edge points of the contour signal, respectively, wherein the fundamental contour signal represents the position and shape of a fundamental lip image; (b) extracting the lip image signal from the face image signal to thereby provide the contour signal and the texture signal of the lip image signal by employing a predetermined lip image signal extraction method; (c) detecting K number of feature points (FP's) from the edge points and K number of corresponding fundamental feature points (FFP's) from the fundamental edge points by using a predetermined feature point detection method with K being a predetermined positive integer less than N, wherein a FP and a corresponding FFP are points essential to define the motion between the lip image and the fundamental lip image, respectively; (d) calculating K number of modification parameters (MP's) based on the K number of FP's and the K number of FFP's, wherein each of the MP's represents a displacement between a FP and a corresponding FFP; and (e) performing motion estimation on the N number of the edge points of the contour signal based on the K number of MP's and the N number of fundamental edge points of the fundamental contour signal to obtain N number of respective reconstructed edge points (REP's) corresponding to the N number of FEP's, thereby supplying a reconstructed contour signal for the lip image signal having the N number of REP's.
9. A method according to claim 8, wherein said K number of FFP's have 4 FFP's (FFP1, FFP2, FFP3 and FFP4) with FFP1 being a fundamental edge point located at a horizontal right-end position in the fundamental lip image; FFP2 being a fundamental edge point located at a horizontal left-end position of the fundamental lip image; FFP3 being a fundamental edge point located at an intersection of a vertical center line of the fundamental lip image and a lower contour of an upper fundamental lip image of the fundamental lip image; and FFP4 being a fundamental edge point located at an intersection of a vertical center line of the fundamental lip image and an upper contour of a lower fundamental lip image of the fundamental lip image.
10. A method according to claim 8 or 9, wherein said K number of FP's have 4 FP's (FP1, FP2, FP3 and FP4) with FP1 being an edge point located at a horizontal right-end position in the lip image; FP2 being an edge point located at a horizontal left-end position of the lip image; FP3 being an edge point located at an intersection of a vertical center line of the lip image and a lower contour of an upper lip image of the lip image; and FP4 being an edge point located at an intersection of a vertical center line of the lip image and an upper contour of the lower lip image of the lip image.
11. A method according to claims 8, 9 or 10, wherein said K number of MP's have 4 MP's (MP1, MP2, MP3 and MP4) with MPi being a displacement from FFPi to FPi, wherein i is a positive integer ranging from 1 to 4.
12. A method according to any ce of clalls 8 wherein in said step (e) , respective REP's are obtained by moving respective corresponding FEP's in accordance with the following respective equations (Eq. 1, Eq. 2, Eq. 3 and Eq. 4): x' = x + #MP1#z(x) ....... (Eq. 1) (condition: x#0), x' = x + #MP2#f(x) ....... (Eq. 2) (condition: x#0), y' = y + #MP3#h(y) ....... (Eq. 3) (condition: y#0), y' = y + #MP4#g(y) ....... (Eq. 4) (condition: y#0), wherein (x',y') and (x, y) in the equations Eq. 1 to Eq. 4 represent xy-coordinates of a REP and a corresponding FEP, respectively, under the condition that the center point of the lip image is matched with the center point of the fundamental lip image as an origin of a two dimensional xy-plane; and functions y=z(x), y=f(x), y=h (x) and y=g(x) satisfy the following boundary conditions on the xy-plane: z(o)=o, z(a)=1, f(0)=0, f(b)=l, h(d)=0 and g(c)=0 under the assumption that xy-coordinates of FFP1, FFP2, FFP3 and FFP4 are (a,0) , (b, o), (c,0) and (d,0) , respectively.
13. The method according to claim 12, wherein the functions z(x), f(x), g(x) and h (x) are expressed as following equations: z(x) = x, a 1 f(x) = x, b

(condition:ysd), (condition:yc), wherein none of a, b, c and d are zero's.
14. A method according to any one of claims 8 to 13, further comprising the step of: (f) combining the reconstructed contour signal with the texture signal to thereby supply a reconstructed lip image signal.
15. An apparatus for reconstructing a lip image signal constructed and arranged substantially as herein described with reference to or as shown in Figure. 3 of the accompanying drawings.
16. A method of reconstructing a lip image signal constructed and arranged substantially as herein described with reference to or as shown in Figures. 4A to 4B and Figures. 5A to 5D and Figure. 6 of the accompanying drawings.