CN105574814A - Portrait paper-cut special effect generation method - Google Patents

Abstract

The invention discloses a method for generating portrait paper-cut special effects, which comprises the following steps: (a) separating the portrait from the background; (b) locating the feature points of the face; (c) performing brightness detection on the image; (d) ) to grayscale the image to obtain a grayscale image; (e) to invert the grayscale image; (f) to do Gaussian blur to the negative film, and smooth the image after inversion; (g) to do color lightening; ( h) scan the generated gray-and-white sketch image, set a threshold value for the image, set the pixel color value of the pixel value smaller than the threshold value to red, and set the color value of the pixel greater than the threshold value to white or transparent; (i) connect the face feature point lines , superimposing facial contour lines; (j) saving the processing results as pictures in two formats of .jpg and .png; (k) performing post-modification on the generated images. The invention has the advantages of low development cost and the like.

Description

A method for generating portrait paper-cut special effects

technical field

The invention relates to a digital art technology, in particular to a method for generating portrait paper-cut special effects.

Background technique

Paper-cutting is a kind of hollow-out art. The research on paper-cutting at home and abroad mainly lies in the generation of paper-cutting patterns, the simulation of folding paper-cutting and the generation of portrait paper-cutting.

In terms of kirigami generation, Jiexu et al. proposed a computer-based kirigami generation method. For an input grayscale image, the kirigami image with geometric connectivity is output. In binary image generation, this method uses a dynamic threshold algorithm to allow users to interact The dynamic threshold parameters are set separately on each layer; Qiao Feng uses the active shape model to calibrate and extract the paper-cut face features, and then uses the deformation method to generate the paper-cut style face portrait; proposes a method based on The digital portrait paper-cut generation method of AND or figure firstly extracts the information expressing the characteristics of the portrait through the bottom-up process, and then performs artistic constraints through the top-down process, and performs paper-cut post-processing for hair and clothes.

However, the existing methods have the following problems:

1. The generation of portrait paper-cut special effects needs to rely on the portrait paper-cut facial features template, and the development cost is high. Deformation of the facial features, especially the eyes, affects the similarity of portrait paper-cuts.

2. The portrait paper-cut generation method based on facial features template matching cannot process portrait accessories (such as glasses, hats, scarves, watches, etc.) and hairstyles into paper-cut styles.

Contents of the invention

The purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art, and provide a method for generating special effects of paper-cut portraits.

The object of the present invention is achieved through the following technical solutions: a method for generating portrait paper-cut special effects, comprising the following steps:

(a) Segment the portrait from the background with the Grab-cut algorithm;

(b) Locate the feature points of the face through the ASM algorithm to obtain the feature point positions of each facial features;

(c) Perform brightness detection on the image. If the image has the problem of underexposure, first increase the brightness of the image;

(d) performing grayscale conversion to the image obtained in step (c) to obtain a grayscale image;

(e) do anti-color processing to the gray scale image that step (d) obtains;

(f) Gaussian blur is done to the negative film that step (e) obtains, and the image after smooth inversion;

(g) do color lightening;

(h) scan the gray-and-white sketch image generated, set a threshold value for the image, set the pixel color value of the pixel value less than the threshold value to red, and set the mark greater than the threshold value to be white or transparent;

(i) Connect the feature points of the face shape with a Bezier curve, and superimpose the contour lines of the face;

(j) the processing result is saved as pictures in two formats of .jpg and .png;

(k) Do post-modification on the generated image to enhance the artistic features of portrait paper-cutting.

In the above step (c), the method for detecting whether the brightness is insufficient is to evaluate by calculating the mean value and variance of the grayscale image, and the brightness mean value point is set to 128.

In the above step (d), the gray value of the pixel is calculated using a weighted method.

In the above step (f), the radius value of the Gaussian blur is set to 18 by default. This method implements adjusting the radius value to flexibly adjust the thickness of the portrait outline.

In the above step (g), the method of lightening the color is: the grayscale image of the original image and the result image obtained in step (f) are calculated and synthesized using the color lightening formula:

RC=BC+(MC*BC)/(255-MC),

Among them, RC represents the result color, BC represents the base color, and MC represents the mixed color;

In the above step (h), the threshold is set to 240.

In the above step (i), the method of overlaying the contour line of the face is as follows: connect the 13 facial feature points among the feature points of the face obtained in step (b) with a red Bezier curve in order to repair the contour line of the face Incomplete and thin problems.

In the above step (k), in order to meet the needs of users for interactive modification, the function of later modifying the portrait line drawing is realized, including manually drawing red lines to ensure connectivity and erasing isolated points. The realization of this step also enables the user to freely create on the basis of the generated line drawing, further enhancing the artistry of portrait paper-cutting.

The working principle of the present invention: the generation method of portrait paper-cut special effect provided by the present invention first uses the Grab-cut algorithm to segment the portrait from the background, and then uses the ASM algorithm to locate the feature points of the face to obtain the feature point positions of each facial features, Then perform brightness detection on the image and increase the brightness of the underexposed image, then grayscale the image, then invert the grayscale image, then do Gaussian blur on the negative film, then do color lightening, and then gray and white sketch The image is thresholded and set as a red-white binary image, then the facial contour lines are superimposed, and then the processing results are saved as .jpg and .png files, and finally the image is interactively modified later.

Compared with the prior art, the present invention has the following advantages and effects:

1. The generation of paper-cut special effects does not require a large number of facial features paper-cut template materials, which reduces the development cost.

2. The execution speed of the portrait paper-cut special effect algorithm is fast, and the generated portrait paper-cut special effect has a higher degree of similarity to the original portrait picture.

3. This portrait paper-cut special effect method can process portrait pictures with accessories such as glasses, hats, scarves, etc., and can process various hairstyles into paper-cut styles.

4. Support users to adjust the effect of paper-cut special effects according to their needs, and support users to freely modify and create interactively, ensuring the artistry of portrait paper-cut special effects.

5. Compared with the method of generating portrait kirigami based on deforming the face kirigami image, the portrait kirigami special effect generated by the present invention has a higher degree of resemblance to the original image; compared with the portrait kirigami generation method based on facial features template matching, the present invention generates Portrait paper-cut special effects can achieve stylized processing of portrait hairstyles, accessories, etc., and the development cost is low. In addition, the present invention supports post-repair by users, which guarantees the artistry of portrait paper-cutting to a certain extent.

Description of drawings

FIG. 1 is a flowchart of a method for generating portrait paper-cut special effects according to the present invention.

detailed description

The present invention will be further described in detail below in conjunction with the embodiments and the accompanying drawings, but the embodiments of the present invention are not limited thereto.

Example

As shown in Figure 1, the main process of the generation method of portrait paper-cut special effects includes the following steps:

(a) portrait image preprocessing;

(b) generating a portrait line drawing;

(c) Subsequent revisions.

Step (a) comprises the following steps:

(a-1) Receive the input image, frame the portrait target and specify some pixels to belong to the portrait target, and then use the Grab-cut algorithm to segment the portrait from the background.

(a-2) Locate the feature points of the face through the ASM algorithm, obtain the feature point positions of the facial features in the output image of step (a-1), and store the feature point coordinates in an array.

(a-3) Luminance detection is performed on the image. If the image is underexposed, first increase the brightness of the image. If the image is not underexposed, no brightness adjustment is necessary.

Step (b) comprises the following steps:

(b-1) Perform grayscale processing on the image. The present invention uses a weighted method to calculate the gray value of the pixel point, i.e. f(i,j)=0.30R(i,j)+0.59G(i,j)+0.11B(i,j);

(b-2) Perform inverse color processing on the grayscale image obtained in step (b-1).

(b-3) Perform Gaussian blur on the negative film obtained in step (b-2), and smooth the inverted image. The default radius value is 18. The invention can adjust the radius according to the smoothed image. If the outline of the image after the first smoothing is too thin, increase the sampling radius; if the outline of the image after the first smoothing is too thick, decrease the sampling radius.

(b-4) Do color dodge. The grayscale image of the original image and the result image obtained in step (b-3) are calculated and synthesized using the color lightening formula:

RC=BC+(MC*BC)/(255-MC),

Among them, RC represents the result color, BC represents the base color, and MC represents the mixed color;

(b-5) Connect the 13 facial feature points sequentially with Bezier curves, and superimpose the facial contour lines.

(b-6) Scan the generated gray-and-white sketch image, set a threshold value for the image, set the pixel color value of the pixel value smaller than the threshold value to red, and set the pixel value greater than the threshold value to white or transparent. Here the threshold is set to 240. At this point, the portrait paper-cut special effect is generated.

(b-7) Save the processing results as pictures in .jpg and .png formats.

Step (c) realizes the functions of drawing and modifying strokes by the user. The unique artistic features of paper-cutting, such as the fine lines at the corners of the eyes and the fine lines on the nose, are difficult to generate with computer technology and need to be artificially added. It is difficult to ensure the connectivity of the lines in the portrait line drawing generated by the above method. Use the "Eraser" tool to eliminate the redundant red part, and use the "Brush" tool to draw a red Bezier curve.

The above-mentioned embodiment is a preferred embodiment of the present invention, but the embodiment of the present invention is not limited by the above-mentioned embodiment, and any other changes, modifications, substitutions, combinations, Simplifications should be equivalent replacement methods, and all are included in the protection scope of the present invention.

Claims (7)

1. a generation method for portrait paper-cut making special efficacy, is characterized in that, comprise the following steps:

A portrait splits with Grab-cut algorithm by () from background;

B () is positioned human face characteristic point by ASM algorithm, obtain the characteristic point position of each face;

C () carries out brightness detection to image; If image exists under-exposed problem, first improve the brightness of image;

D () is carried out gray processing to the image obtained in step (c) and is obtained gray-scale map;

E () does inverse process to the gray-scale map that step (d) obtains;

F () does Gaussian Blur to the negative film that step (e) obtains, the image after smooth anti-phase;

G () does color dodge process;

H greyish white sketch image that () scanning generates, to image setting threshold value, pixel color value pixel value being less than threshold value is set to redness, is greater than being labeled as of threshold value and is set to white or transparent;

I shape of face unique point Bezier couples together by (), superposition face mask lines;

J result is saved as the picture of .jpg and .png two kinds of forms by ();

K () does later stage modify, to ensure the feature of portrait paper-cut making art to the image generated.

2. the generation method of portrait paper-cut making special efficacy according to claim 1, is characterized in that: whether step (c) is that the average and variance by calculating gray-scale map carrys out evaluate image under-exposed, and brightness average point is set to 128.

3. the generation method of portrait paper-cut making special efficacy according to claim 1, is characterized in that: step (d) is the gray-scale value adopting the method for weighting to calculate pixel.

4. the generation method of portrait paper-cut making special efficacy according to claim 1, is characterized in that: the radius value default setting of step (f) is 18; This method realizes regulating radius value to adjust the thickness of portrait profile.

5. the generation method of portrait paper-cut making special efficacy according to claim 1, is characterized in that: it is the gray-scale map of former figure and result images obtained in the previous step are used following color dodge formulae discovery synthesis that step (g) does color dodge:

RC＝BC+(MC*BC)/(255-MC)，

Wherein, RC represents result look, and BC represents primary colours, and MC represents color mixture.

6. the generation method of portrait paper-cut making special efficacy according to claim 1, is characterized in that: threshold value is set to 240 by step (h).

7. the generation method of portrait paper-cut making special efficacy according to claim 1, is characterized in that: in step (k), and the described image to generating does later stage modify and comprises the red lines of hand drawn and erasing isolated point.