CN102799885B - Lip external outline extracting method - Google Patents

Abstract

The invention relates to a lip external outline extracting method, which can be used for automatically extracting a lip external outline according to distribution characteristic difference of areas of a lip and a face skin color. A lip image is divided by using a self-adaption threshold value method; noises can be removed by combining a communication area with morphological treatment; and a lip external outline line is determined by using a curve fitting method. Therefore, the lip external outline can be rapidly and effectively extracted, and the requirement on characteristic extraction in man-machine interaction can be met.

Description

Lip Outside contour extraction method

Technical field

The present invention relates to a kind of lip Outside contour extraction method, for video digital images analysis and understanding, belong to intelligent information processing technology field.

Background technology

Along with the fast development of computer technology; research meets the novel human-computer interaction technology Showed Very Brisk that the mankind exchange custom naturally; and human-computer interaction technology from progressively transferring to centered by computing machine, and focus be put on man; multimedia user interface has then greatly enriched the form of expression of computerized information, makes user can replace or utilize multiple sensory channel simultaneously.Wherein, computing machine automatic Understanding mankind lip is allowed to move the focus that language has become intelligent human-machine interaction research field.Complete labiomaney system comprises lip location, lip moves the link such as feature extraction, Lip Movement Recognition, and wherein lip outline extracts very important effectively, automatically.

When extracting lip outline, generally have following method: one is based on colour of skin method, the method is large by illumination effect, and robustness is low; Two is based on lip model method, and the method is subject to the impact of different speaker's shape of the mouth as one speaks, the especially interference of tongue when magnifying mouth, causes that gained lip outline result is undesirable and algorithm is complicated, real-time is low.

Summary of the invention

The object of the invention is to for existing lip outline extracting method, there is mouth deformationization large, magnify the interference of tongue during mouth, and the problem such as real-time effect is undesirable, a kind of lip Outside contour extraction method of improvement is provided.It is the distribution characteristics difference according to lip and face complexion area, Fast Segmentation lip, and extracts lip outline, to improve dirigibility and the simplicity of man-machine interaction.

For achieving the above object, design of the present invention is: adopt method for detecting human face, maintains the invariance, carry out lip region coarse positioning according to the ratio between each feature of face face; According to the distribution characteristics difference of lip and face complexion area, adopt adaptive threshold fuzziness lip; According to the geometric properties of lip region distribution, adopt method for marking connected region, remove non-lip region, and remove burr by mathematical mor-phology method and fill up hole, to obtain more complete lip region, adopt curve curve-fitting method, fast and effeciently determine lip outline.

According to foregoing invention design, the present invention adopts following technical proposals:

1. a lip Outside contour extraction method, it is characterized in that the distribution characteristics difference segmentation lip according to lip and face complexion area, and automatically extract lip outline, concrete steps are as follows:

1) man face image acquiring system is started: gather video image;

2) lip region coarse positioning: carry out Face datection, obtains human face region; Second region of face vertical area is tentatively defined as lip region;

3) lip region segmentation

4) lip outline is extracted: in step 3) in the lip region bianry image that extracts, obtain lip marginal point, adopt curve fitting method to extract the outline of lip.

Above-mentioned steps 3) lip segmentation concrete operation step as follows:

(1) color space conversion, calculates color-values s: red by RGB color space r, green gtwo components, calculate color-values s:

(2) histogram generates: according to determined color svalue, obtains its histogram;

(3) lip region coarse segmentation: according to step 2) determined initial lip region, obtain its area pixel number a, utilize step (2) gained simultaneously shistogram, adds up larger sthe pixel number of value m; When mequal n%× atime, nfor real lip region accounts for the number of pixels number percent of the lip region initially determined, corresponding svalue is threshold value t.The image-region meeting following formula condition is defined as possible lip region g

G=s> T

(4) lip region is determined: according to the geometric properties of the wide, high of lip and area etc., adopt connected component labeling method to remove non-lip region, and adopt mathematical mor-phology method remove burr and fill up hole.

Principle of the present invention is as follows:

In the inventive solutions, according to saturation degree in HSV scomputing formula and the distribution characteristics of lip region RGB, derive a kind of new color conversion method, and adopt Adaptive Thresholding to be partitioned into lip from area-of-interest.

Saturation degree scomputing formula as follows:

In lip region, due to max (R, G, B)=R, min (R, G, B)=B, above formula can be reduced to

According to lip, the colour of skin and tongue region r, g, bdistribution characteristics, builds r-Gwith r-Bdistribution plan is analyzed, and draws utilization (R-G)/Rthere is better segmentation effect.But when there are some dressings in face time, rvalue may trend towards 0, cause lip region extract difficulty.For addressing this problem, derive following formula, will svalue is limited between 0-1.

Determine image segmentation threshold t.Real lip region accounts for the number of pixels of the lip region initially determined a's n%, nfor real lip region accounts for the number of pixels number percent of the lip region initially determined, the number of namely real lip region pixel is n%× a.Utilize statistics with histogram larger sthe pixel number of value, when it equals n%× atime, corresponding svalue is threshold value t.Split image according to the following formula, lip region is set to 1, and non-lip region sets to 0.

According to the geometric properties of mouth shape, adopt connected component labeling method to remove nontarget area, remove burr with Morphological scale-space and fill up hole, obtaining more complete lip region.According to the distribution characteristics of upper lower lip, adopt four times respectively and lower lip on conic fitting, thus obtain complete lip outline.

The present invention compared with prior art, there is following apparent outstanding substantive distinguishing features and remarkable advantage: the present invention maintains the invariance according to the ratio between each feature of face face, carry out lip coarse positioning, and it is different in nature according to the color difference of lip region and human face region, adopt adaptive threshold fuzziness lip, connected component labeling method and mathematical mor-phology method is utilized to remove non-lip region, adopt curve-fitting method determination lip outline, simple operation, flexibly, easy realization, solve when extract real-time lip outline, lip form of diverse and when holding labile impact and magnify mouth tongue region interference, the deficiency of computing complexity, improve the robustness of lip Outside contour extraction, the lip Outside contour extraction under different scene can be adapted to.Method of the present invention realizes fast, effectively, easily.

Accompanying drawing explanation

Fig. 1 is the flowsheet of the inventive method.

Fig. 2 is the original facial image of one embodiment of the invention.

Fig. 3 is the Face datection image of one embodiment of the invention.

Fig. 4 is the lip coarse positioning image of one embodiment of the invention.

Fig. 5 is that the threshold value of one embodiment of the invention chooses histogram.

The lip bianry image that Fig. 6 is partitioned into based on color space.

Fig. 7 be one embodiment of the invention the lip bianry image after non-lip region and holes filling is removed to bianry image.

Fig. 8 is the image of the upper and lower contour edge point of extraction lip of one embodiment of the invention.

Fig. 9 is the lip Outside contour extraction image of one embodiment of the invention.

Embodiment

A preferred embodiment of the present invention is: running program as shown in Figure 1.As shown in Figure 2, Face datection image, as Fig. 3, according to color space transfer algorithm, carries out lip segmentation to the coloured image shown in Fig. 3 to the original facial image of this example.Concrete steps are as follows:

1) man face image acquiring system is started: gather video image;

2) lip region coarse positioning

Carry out Face datection, obtain human face region (Fig. 3).Second region of face vertical area is tentatively defined as lip region (Fig. 4).

3) lip region segmentation

Concrete operation step is as follows:

(1) color space conversion, calculates color-values s: red by RGB color space r, green gtwo components, calculate color-values s:

(2) histogram generates: according to determined color svalue, obtains its histogram (Fig. 5);

(3) lip region coarse segmentation: according to step 2) determined initial lip region, the number of pixels obtaining initial lip region is 47040, utilizes step (2) gained simultaneously shistogram, adds up larger sthe pixel number of value m; When mequal 5 %× 47040(gets nbe 5) time, corresponding svalue is 0.26, is threshold value t(Fig. 5).The image-region meeting following formula condition is defined as possible lip region g.Fig. 6 is the bianry image of lip region coarse segmentation.

G= s> T

(4) lip region is determined: according to the geometric properties of the wide, high of lip and area etc., adopt connected component labeling method to remove non-lip region, and adopt mathematical mor-phology method remove burr and fill up hole, result as shown in Figure 7.

4) lip outline is extracted

In step 3) in the lip region bianry image that extracts, obtain lip marginal point (as Fig. 8), adopt four times and twice curve fitting method respectively, determine the outline of lip, result is as shown in Figure 9.

Claims (1)

1. a lip Outside contour extraction method, it is characterized in that the distribution characteristics difference segmentation lip according to lip and face complexion area, and automatically extract lip outline, concrete steps are as follows:

1) man face image acquiring system is started: gather video image;

2) lip region coarse positioning: carry out Face datection, obtains human face region; Second region of face vertical area is tentatively defined as lip region;

3) lip region segmentation;

4) lip outline is extracted: in step 3) in the lip region bianry image that extracts, obtain lip marginal point, adopt curve fitting method to extract the outline of lip;

Described step 3) lip segmentation concrete operation step as follows:

(1) color space conversion, calculates color-values s: red by RGB color space r, green gtwo components, calculate color-values s:

(2) histogram generates: according to determined color svalue, obtains its histogram;

(3) lip region coarse segmentation: according to step 2) determined initial lip region, obtain its area pixel number a, utilize step (2) gained simultaneously shistogram, adds up larger sthe pixel number of value m; When mequal n%× atime, nfor real lip region accounts for the number of pixels number percent of the lip region initially determined, corresponding svalue is threshold value t; The image-region meeting following formula condition is defined as possible lip region g

G=s> T

(4) lip region is determined: according to the geometric properties of the wide, high of lip and area etc., adopt connected component labeling method to remove non-lip region, and adopt mathematical mor-phology method remove burr and fill up hole.