CN107346408A - Age recognition methods based on face feature - Google Patents

Abstract

A method for age recognition based on facial features, comprising the steps of: acquiring a face image, and grabbing a face area; setting feature points on the captured face area image; Determining the age feature area; extracting the age feature of the age feature area, and obtaining a corresponding age value; and comparing the age value with a preset threshold to determine age.

Description

Age Recognition Method Based on Facial Features

technical field

The invention relates to an age recognition method based on facial features.

Background technique

Existing face age recognition is mostly performed by processing the entire face image. However, the extraction of the entire image features of the face is very complicated, the information contained is complex and huge, and the related image processing is cumbersome, resulting in a lengthy pre-processing process. , the judgment efficiency is low.

Contents of the invention

In view of the above, it is necessary to provide a method for efficiently identifying age.

A method for age recognition based on facial features, comprising the steps of:

Obtain a face image and capture the face area;

Set feature points on the captured face area image;

Determine the age feature area on the face area according to the coordinates of the feature points;

Extracting the age feature of the age feature area, and obtaining the corresponding age value; and

The age is determined by comparing the age value with a preset threshold.

Preferably, the step of setting feature points on the captured face region image includes obtaining the positions of eyebrows, eyes, nose and mouth in the face region image, and marking these organs with feature points.

Preferably, the left eyebrow tail, the left eyebrow head, the right eyebrow head, the right eyebrow tail, the outer corner of the left eye, the center of the left eye, the inner corner of the left eye, the inner corner of the right eye, the center of the right eye, the right eye Mark the outer corner of the eye, the left alar, the midpoint of the tip of the nose, the right alar, the left corner of the mouth, the midpoint of the upper part of the mouth, the right corner of the mouth and the midpoint of the lower part of the mouth.

Preferably, before determining the age feature area on the face area according to the coordinates of the feature points, the distance between the midpoint of the two eyes and the vertical distance between the midpoint of the eyes and the midpoint of the nose are obtained through the position coordinates of the center of the left eye, the center of the right eye, and the midpoint of the nose , where the distance between the midpoints of the two eyes is equal to the abscissa of the midpoint of the left eye minus the abscissa of the midpoint of the right eye, and the vertical distance between the midpoint of the two eyes and the midpoint of the nose is equal to the ordinate of the midpoint of the nose minus the ordinate of the midpoint of the two eyes. .

Preferably, the age characteristic areas determined on the face area according to the coordinates of the feature points include left eye tail area, right eye tail area, left under eye area, right under eye area, left nose to mouth corner area and right nose to mouth corner area.

Preferably, prior to the step of extracting the age features of the age feature area, age feature areas of the same type are adjusted to have the same size.

Preferably, the left eye tail area and the right eye tail area are the same type of age characteristic area, the left eye area and the right eye area are the same type of age feature area, and the left nose to mouth area and right nose to mouth area are the same type Age characteristic area.

Preferably, after adjusting the age characteristic regions of the same type into regions of the same size, the image of the age characteristic region is converted into an easily identifiable image through a statistical transformation algorithm.

Preferably, the preset threshold includes a first threshold and a second threshold, and if the age value is greater than the first threshold after comparing the age value with the first threshold, after comparing the age value with the second threshold Two thresholds are compared.

Compared with the prior art, in the above-mentioned age recognition method based on facial features, the age feature area is obtained by setting feature points on the face area image, and the age is judged by processing these age feature areas, and the data processing The smaller the amount, the higher the judgment efficiency.

Description of drawings

Fig. 1 is a flow chart of the age recognition method based on facial features of the present invention.

FIG. 2 is a schematic diagram of capturing face regions in the age recognition method in FIG. 1 .

Fig. 3 is a schematic diagram of setting feature points in the age recognition method of Fig. 1 .

FIG. 4 is a schematic diagram of the distance between the midpoint of the eyes and the vertical distance between the midpoint of the eyes and the midpoint of the tip of the nose obtained in the age identification method of FIG. 1 .

FIG. 5 is a schematic diagram of determining an age characteristic region of a face in the age recognition method of FIG. 1 .

FIG. 6 is a schematic diagram of determining eye tail regions in the age recognition method in FIG. 1 .

FIG. 7 is a schematic diagram of determining the under-eye area in the age identification method in FIG. 1 .

FIG. 8 is a schematic diagram of determining the region from the alar of the nose to the corner of the mouth in the age recognition method of FIG. 1 .

Description of main component symbols

Feature points 1-17 distance between eyes W Vertical distance between the midpoint of the eyes and the midpoint of the tip of the nose h end of eye area a and b a, b Right now areas c and d c, d Nose to mouth area e and f e, f endpoint a1-a4, b1-b4, c1-c4, d1-d4, e1-e4, f1-f4

The following specific embodiments will further illustrate the present invention in conjunction with the above-mentioned drawings.

detailed description

Fig. 1 is the flow chart of a kind of age identification method based on facial feature of the present invention, is used for identifying people's age, and described age identification method comprises the following steps:

Step 101, acquire a face image, and capture a face area.

Please refer to FIG. 2 , usually the acquired face image will include other parts of the human body, such as the neck, etc. In step 101, the image of the face area is captured through the pre-established human face model and the acquired face image , and remove other parts of the image.

Step 102, setting feature points on the captured face region image.

Please refer to Figure 3. By comparing the feature model with the face area image, the positions of the eyebrows, eyes, nose, and mouth in the face area image are obtained, and these organs are marked with feature points, as shown in the figure As shown in 3, use feature point 1 to mark the end of the left eyebrow (the definition of left and right is based on the person viewing the attached drawing), use feature point 2 to mark the head of the left eyebrow, and use feature point 3 to mark the right eyebrow Mark the head of the eyebrow, use feature point 4 to mark the end of the right eyebrow, use feature point 5 to mark the outer corner of the left eye, use feature point 6 to mark the center of the left eye, and use feature point 7 to mark the inner corner of the left eye , use feature point 8 to mark the inner corner of the right eye, use feature point 9 to mark the center of the right eye, use feature point 10 to mark the outer corner of the right eye, use feature point 11 to mark the left nose, and use feature point 12 Mark the midpoint of the nose, mark the right nose with feature point 13, mark the left mouth corner with feature point 14, mark the midpoint of the upper part of the mouth with feature point 15, mark the right mouth corner with feature point 16, use The feature point 17 marks the midpoint of the lower part of the mouth, and in this embodiment seventeen feature points are used to mark the face region image; in other embodiments, more feature points or less can also be used as required The feature points mark the four organs of eyebrows, eyes, nose, and mouth. According to the needs of judgment, the feature points of the ear organ can also be marked, or other features of the human face can be marked.

Step 103, calculate the distance W between the midpoint of the eyes and the vertical distance H between the midpoint of the eyes and the midpoint of the nose according to the position information of the feature points.

Please refer to Figure 4, the distance between the midpoint of the two eyes and the vertical distance H between the midpoint of the eyes and the midpoint of the nose are usually calculated based on the coordinates of these feature points. For example, the coordinates of feature point 6 are (X6, Y6), and the coordinates of feature point 9 is (X9, Y9), and the right side of feature point 12 is (X12, Y12), then the distance between the midpoint of the two eyes is W=X9-X6, and the vertical distance between the midpoint of the two eyes and the midpoint of the nose is H=Y12-(Y6+Y9)/ 2.

Step 104, determine the age characteristic area of the face according to the coordinates of each coordinate point and the distance W, H.

Please refer to Fig. 5, in this embodiment, select eye end area a and b, eye area c and d, nose wing to mouth corner area e and f as the age characteristic area of face, because the age of these areas of people's face The features are relatively obvious, for example, the wrinkles in these areas are quite different for people of different ages; in other embodiments, other facial areas can also be added for judgment according to different needs, such as the forehead area.

Please refer to Fig. 6, which is a schematic diagram of determining a left eye tail area a and a right eye tail area b, the left eye tail area a is a square area, which includes an upper left corner endpoint a1, an upper right corner endpoint a2, and a lower left corner endpoint a3 and the lower right corner endpoint a4, and calculate the X-axis coordinates and Y-axis coordinates of each corner endpoint through the formula listed in the table below.

endpoint name X-axis coordinate calculation method Y-axis coordinate calculation method End point a1 of the upper left corner of the left eye tail area X=X6–W*0.6 Y=(Y6+Y9)/2 End point a2 of the upper right corner of the left eye tail area X=X6–W*0.4 Y=(Y6+Y9)/2 End point a3 of the lower left corner of the left eye tail area X=X6–W*0.6 Y=(Y6+Y9)+H*0.8 End point a4 of the lower right corner of the left eye tail area X=X6–W*0.4 Y=(Y6+Y9)+H*0.8

Similarly, the right eye tail area b is a square area, which includes the upper left corner endpoint b1, the upper right corner endpoint b2, the lower left corner endpoint b3 and the lower right corner endpoint b4, and the X-axis coordinates of each corner endpoint are calculated by the formula listed in the following table and Y-axis coordinates.

endpoint name X-axis coordinate calculation method Y-axis coordinate calculation method End point b1 of the upper left corner of the right eye tail area X=X9+W*0.4 Y=(Y6+Y9)/2 End point b2 of the upper right corner of the right eye tail area X=X9+W*0.6 Y=(Y6+Y9)/2 End point b3 of the lower left corner of the right eye tail area X=X9+W*0.4 Y=(Y6+Y9)+H*0.8 End point b4 of the lower right corner of the right eye tail area X=X9+W*0.6 Y=(Y6+Y9)+H*0.8

Please refer to Figure 7, the area c under the left eye is a square area, which includes the upper left corner endpoint c1, the upper right corner endpoint c2, the lower left corner endpoint c3 and the lower right corner endpoint c4, and the X of each corner endpoint is calculated by the formula listed in the following table Axis coordinates and Y-axis coordinates.

endpoint name X coordinate calculation method Y coordinate calculation method End point c1 of the upper left corner of the area under the left eye X=X6–W*0.4 Y=(Y6+Y9)/2+H*0.35 End point c2 of the upper right corner of the area under the left eye X=X6+W*0.3 Y=(Y6+Y9)/2+H*0.35 End point c3 of the lower left corner of the area under the left eye X=X6–W*0.4 Y=(Y6+Y9)/2+H*0.8 End point c4 of the lower right corner of the area under the left eye X=X6+W*0.3 Y=(Y6+Y9)/2+H*0.8

Similarly, the area d under the right eye is a square area, which includes the endpoint d1 of the upper left corner, the endpoint d2 of the upper right corner, the endpoint d3 of the lower left corner, and the endpoint d4 of the lower right corner, and calculate the X-axis coordinates of each corner endpoint and Y-axis coordinates.

endpoint name X coordinate calculation method Y coordinate calculation method End point d1 of the upper left corner of the area under the right eye X=X9–W*0.4 Y=(Y6+Y9)/2+H*0.35 End point d2 in the upper right corner of the area under the right eye X=X9+W*0.3 Y=(Y6+Y9)/2+H*0.35 End point d3 of the lower left corner of the area under the right eye X=X9–W*0.4 Y=(Y6+Y9)+H*0.8 End point d4 of the lower right corner of the area under the right eye X=X9+W*0.3 Y=(Y6+Y9)+H*0.8

Please refer to Figure 8, the area e from the left nose wing to the corner of the mouth is a square area, which includes the upper left corner endpoint e1, the upper right corner endpoint e2, the lower left corner endpoint e3 and the lower right corner endpoint e4, and the corner endpoints are calculated by the formula listed in the following table The X-axis coordinates and Y-axis coordinates.

endpoint name X coordinate calculation method Y coordinate calculation method Point e1 from the left nose to the upper left corner of the mouth area X=X6–W*0.32 Y=Y12 End point e2 from the left nose wing to the upper right corner of the corner of the mouth X=X6+W*0.05 Y=Y12 End point e3 from the left nose wing to the lower left corner of the corner of the mouth X=X6–W*0.32 Y=(Y14+Y16)/2 From the left nose to the bottom right corner of the mouth area e4 X=X6+W*0.05 Y=(Y14+Y16)/2

Similarly, the area f from the right nose wing to the corner of the mouth is a square area, which includes the upper left corner endpoint f1, the upper right corner endpoint f2, the lower left corner endpoint f3 and the lower right corner endpoint f4, and the X-axis of each corner endpoint is calculated by the formula listed in the following table coordinates and Y-axis coordinates.

endpoint name X coordinate calculation method Y coordinate calculation method From the right nose wing to the upper left corner endpoint e1 of the mouth corner area X=X9+W*0.05 Y=Y12 From the right nose to the upper right endpoint e2 of the corner of the mouth X=X9+W*0.32 Y=Y12 From the right nose to the bottom left corner of the mouth area e3 X=X9+W*0.05 Y=(Y14+Y16)/2 From the right nostril to the lower right endpoint of the corner of the mouth e4 X=X9+W*0.32 Y=(Y14+Y16)/2

Therefore, the positions and sizes of the age feature regions a-f are set by the above method.

Step 105, adjusting the age feature areas of the same type to areas of the same size.

In the above age characteristic regions a, b, c, d, e and f, the left eye tail region a and the right eye tail region b are age characteristic regions of the same type, and the left eye lower region c and right eye lower region d are of the same type The age feature area, the left nose to mouth corner area e and the right nose to mouth corner area f are the same type of age feature area, for example, if the length of the left eye tail area a is 9, the height is 6, and the length of the right eye tail area b is 6 , the width is 8, then by shortening the length of the left eye tail area a from 9 to 6, stretching the height of the left eye tail area a from 6 to 8, and making the left eye tail area a and the right eye tail area b are the same size.

In step 106, the image of each age feature area is converted into an image that is easy to recognize by a statistical transformation algorithm.

The statistical transformation algorithm divides every 3*3 pixels of the image into an area, then calculates the pixel average of all pixels in the area, and then compares the pixel value of each pixel in the area with the pixel average For comparison, if the pixel value of a certain point is greater than the average value of the pixel, the pixel value of the point is changed to 1; if the pixel value of a certain point is smaller than the average value of the pixel, the pixel value of the point is changed to 0.

Step 107, extract the age feature of the converted image, and obtain the corresponding age value, compare the age value with a first threshold, if the age value is greater than or equal to the first threshold, then go to step 108; if the age value is less than the first threshold, go to step 109.

Step 108, judge as middle-aged and elderly people, and compare the age value with a second threshold; if the age value is greater than or equal to the second threshold value, go to step 110; if the age value is greater than or equal to the second threshold value , to step 111.

Step 109, judging as a young person.

Step 110, judging as an elderly person.

Step 111, judging as a middle-aged person.

In the above-mentioned age recognition method based on facial features, the age feature area is obtained by setting feature points on the face area image, and the age is judged by processing and judging these age feature areas. The amount of data processing is small, and the judgment Higher efficiency.

Claims (9)

1. A method for age recognition based on facial features, comprising the steps of: Obtain a face image and capture the face area; Set feature points on the captured face area image; Determine the age feature area on the face area according to the coordinates of the feature points; Extracting the age feature of the age feature area, and obtaining the corresponding age value; and The age is determined by comparing the age value with a preset threshold. 2. The age recognition method as claimed in claim 1, wherein: the step of setting feature points on the captured face region image comprises obtaining the four points of eyebrows, eyes, nose and mouth in the face region image. The location of each organ, and mark these organs with feature points. 3. The age recognition method as claimed in claim 2, characterized in that: use feature points to align left eyebrow tail, left eyebrow head, right eyebrow eyebrow head, right eyebrow eyebrow tail, left eye outer corner, left eye center, left eye Mark the inner corner of the eye, the inner corner of the right eye, the center of the right eye, the outer corner of the right eye, the left nose, the midpoint of the tip of the nose, the right nose, the left corner of the mouth, the midpoint of the upper part of the mouth, the right corner of the mouth and the midpoint of the lower part of the mouth. 4. The age recognition method as claimed in claim 3, characterized in that: before determining the age feature region on the face region according to the coordinates of the feature points, two points are obtained by the position coordinates of the center of the left eye, the center of the right eye and the midpoint of the tip of the nose. The distance between the midpoint of the eyes and the vertical distance between the midpoint of the two eyes and the midpoint of the tip of the nose. The distance between the midpoint of the two eyes is equal to the abscissa of the midpoint of the left eye minus the abscissa of the midpoint of the right eye. The vertical distance between the midpoint of the two eyes and the midpoint of the tip of the nose is equal to the midpoint of the tip of the nose The ordinate of , minus the average value of the ordinate of the midpoint of the two eyes. 5. The age recognition method as claimed in claim 4, characterized in that: the age characteristic regions determined on the face region according to the coordinates of the feature points include the left eye tail region, the right eye tail region, the left eye region, and the right eye region , left nose to mouth area and right nose to mouth area. 6. The age identification method according to claim 5, characterized in that: before the step of extracting the age characteristics of the age characteristic region, the age characteristic regions of the same type are adjusted to the same size region. 7. The age recognition method as claimed in claim 6, characterized in that: the left eye tail area and the right eye tail area are the same type of age characteristic area, the left eye lower area and the right eye lower area are the same type of age characteristic area, the left eye end area is the same type of age characteristic area, The region from the alar to the corner of the mouth and the region from the right alar to the corner of the mouth are the same type of age characteristic regions. 8. The age recognition method as claimed in claim 6, characterized in that: after adjusting the age characteristic regions of the same type into regions of the same size, the image of the age characteristic region is converted into an easily identifiable image through a statistical transformation algorithm. 9. The age identification method according to claim 1, wherein the preset threshold includes a first threshold and a second threshold, and if the age value is greater than the first threshold after comparing with the first threshold A first threshold, comparing the age value with the second threshold.