WO2022267653A1

WO2022267653A1 - Image processing method, electronic device, and computer readable storage medium

Info

Publication number: WO2022267653A1
Application number: PCT/CN2022/087744
Authority: WO
Inventors: 辛琪; 孙宇超; 魏文; 姚聪
Original assignee: 北京旷视科技有限公司; 北京迈格威科技有限公司
Priority date: 2021-06-23
Filing date: 2022-04-19
Publication date: 2022-12-29
Also published as: CN113591562A; CN113591562B

Abstract

An image processing method, an electronic device, and a computer readable storage medium, relating to the field of image processing. The method comprises: obtaining face key point information of an image to be processed (S110); according to the face key point information, determining a movement reference point when moving pixels in a face area (S120), and dividing the face area into different local areas (S130); and during face thinning, moving pixels included in the local areas toward the movement reference point according to pixel moving policy parameters corresponding to the local areas to which the pixels belong (S140). Therefore, the final face thinning effect can be natural as much as possible, thereby improving the visual effect after face thinning.

Description

Image processing method, electronic device, and computer-readable storage medium

Cross References to Related Applications

This disclosure claims the priority of the Chinese patent application with application number 202110701338X and titled "image processing method, device, electronic device, and computer-readable storage medium" submitted to the State Intellectual Property Office of China on June 23, 2021, all of which The contents are incorporated by reference in this disclosure.

technical field

The disclosure belongs to the field of image processing, and in particular relates to an image processing method, electronic equipment, and a computer-readable storage medium.

Background technique

With the development of pan-entertainment trends, face beautification and face-lifting technologies will be involved in application scenarios such as video conferences, live broadcasts, taking pictures, and photo processing.

In related face-beautifying and face-lifting techniques, a staff member pre-sets a facial contour template, and then moves the facial contour included in the acquired image to be processed by a fixed distance according to the fixed parameters provided by the facial contour template. Although this operation can make the image appear thinner, the effect is to optimize the face curve of the image to be processed into a uniform awl face shape, which will make the face appear unnatural visual effect, and even cause the face to be out of shape distortion.

Contents of the invention

In view of this, the present disclosure provides an image processing method, device, electronic equipment, and computer-readable storage medium, which can improve the visual effect of the image to be processed obtained after face-thinning processing.

Some embodiments of the present disclosure are implemented as follows:

An embodiment of the present disclosure provides an image processing method, the method may include: acquiring face key point information of an image to be processed, the image to be processed may include a face area; according to the face key point information, determine A moving reference point when moving the pixels in the face area; according to the key point information of the face, dividing the face area into a plurality of local areas; for each of the local areas, based on the The parameter of the pixel movement strategy corresponding to the local area is used to move each pixel in the local area toward the direction of the moving reference point.

With reference to the embodiments of the present disclosure, in a possible implementation manner, the pixel movement strategy parameter may include a pixel position adjustment ratio; correspondingly, based on the pixel movement strategy parameter corresponding to the local area, the local Each pixel in the area moves toward the direction of the moving reference point, which may include: for each pixel in the local area, determining a first distance between the pixel and the moving reference point; based on Adjusting the ratio of the first distance and the corresponding pixel position to determine a second distance between the pixel point and the moving reference point; moving the pixel point toward the direction of the moving reference point so that the The distance between the pixel point after moving and the moving reference point is equal to the second distance.

With reference to the embodiments of the present disclosure, in a possible implementation manner, the second distance between the pixel point and the moving reference point is determined based on the first distance and the corresponding pixel position adjustment ratio. , may include: determining a first product value of the first distance and a pixel position adjustment ratio corresponding to the local area to which the pixel point belongs; determining the first product value as the second distance.

With reference to the embodiments of the present disclosure, in a possible implementation manner, after moving each pixel point in the local area toward the direction of the moving reference point based on the pixel movement strategy parameter corresponding to the local area , the method may further include: in response to a face adjustment instruction triggered by the user, acquiring the position information of each pixel in the face area that has been moved before and after the move; wherein, the face The adjustment command carries a face-thinning intensity coefficient; based on the face-thinning intensity coefficient, the position information before the movement, and the position information after the movement, determine the target position corresponding to each pixel point in the face area that has been moved Information; moving each pixel point in the face area that has been moved to the target position information.

With reference to the embodiments of the present disclosure, in a possible implementation manner, the location information may be represented by coordinate information; based on the face-thinning intensity coefficient, the location information before the movement, and the location information after the movement, Determining the target position information corresponding to each moved pixel in the face area may include: determining the distance between the second coordinate information and the first coordinate information for each moved pixel in the face area The first coordinate difference; wherein, the position information before the movement can be represented by the first coordinate information, and the position information after the movement can be represented by the second coordinate information; calculate the thin face intensity coefficient and the A second product value between the first coordinate difference values; and a sum value of the first coordinate information and the second product value is determined as the target position information.

With reference to the embodiments of the present disclosure, in a possible implementation manner, the size of the face-thinning strength coefficient is adjustable.

With reference to the embodiments of the present disclosure, in a possible implementation manner, the determining the moving reference point when moving the pixel points in the human face area according to the key point information of the human face may include: determining the The human face key point corresponding to the central position of the eyes in the human face area; the key point of the human face corresponding to the central position of the eyes is determined as the mobile reference point;

Alternatively, a key point of the human face corresponding to the position of the tip of the nose in the face area may be determined, and the key point of the human face corresponding to the position of the tip of the nose may be determined as the moving reference point.

With reference to the embodiments of the present disclosure, in a possible implementation manner, the determining the moving reference point when moving the pixel points in the human face area according to the key point information of the human face may include: based on The human face key point on the center line in the vertical direction of the human face area determines a reference line; for each pixel point in the human face area, the distance between the reference line and the pixel point The minimum face key point is determined as the moving reference point corresponding to the pixel point; correspondingly, the moving each pixel point in the local area toward the direction of the moving reference point may include: moving the local area Each pixel point in the area moves towards the direction of the corresponding moving reference point.

With reference to the embodiments of the present disclosure, in a possible implementation manner, the face key point information may include identification information of the face key points, and according to the face key point information, the face area is divided into For a plurality of local areas, it may include: according to the corresponding relationship between the obtained identification information of the key points of the face and the facial organs, determine the set of key points of the face corresponding to each of the facial organs; The area surrounded by each key point of the face is determined as a local area.

With reference to the embodiments of the present disclosure, in a possible implementation manner, pixel position adjustment ratios corresponding to different local regions are different.

With reference to the embodiments of the present disclosure, in a possible implementation manner, the acquisition of face key point information of the image to be processed may include: inputting the image to be processed into a face key point detection model, through the human face key point detection model, The facial key point detection model performs human face key point detection on the image to be processed; and obtains the human face key point information output by the human face key point detection model.

With reference to the embodiments of the present disclosure, in a possible implementation manner, the face key point information may include the coordinate information of the face positioning frame and the coordinate information of the face key point; Point information, before determining the moving reference point when moving the pixel points in the human face area, the method may also include:

The coordinate information of the key points of the face is normalized based on the coordinate information of the face positioning frame, and the normalized coordinate information of the key points of the face is obtained.

The embodiment of the present application also provides an image processing device, which includes: an acquisition module, a determination module, a division module, and an adjustment module.

The acquiring module is used to acquire the face key point information of the image to be processed, and the image to be processed includes the face area; the determination module is used to determine the pixels in the face area according to the face key point information The moving reference point when the point moves; the division module is used to divide the face area into a plurality of local areas according to the key point information of the face; the adjustment module is used for each of the local areas based on The pixel movement strategy parameter corresponding to the local area moves each pixel in the local area toward the moving reference point.

An embodiment of the present disclosure also provides an electronic device, which may include: a memory and a processor, the memory is connected to the processor; the memory is used to store a program; the processor calls the program stored in the memory A program to execute the methods provided by the above-mentioned embodiments of the present disclosure and/or any possible implementation manners in combination with the embodiments of the present disclosure.

The embodiment of the present disclosure also provides a non-volatile computer-readable storage medium (hereinafter referred to as the computer-readable storage medium), on which a computer program is stored, and when the computer program is run by a computer, the above-mentioned embodiments of the present disclosure are executed. And/or in combination with the method provided in any possible implementation manner of the embodiments of the present disclosure.

Embodiments of the present disclosure also provide a computer program product including code instructions, which, when executed by a processor, cause the processor to execute any one of the above-mentioned embodiments of the present disclosure and/or in combination with the embodiments of the present disclosure. Possible implementations of the methods provided.

The embodiment of the present disclosure also provides a computer program, which executes the method provided by the embodiment of the present disclosure and/or in combination with any possible implementation manner of the embodiment of the present disclosure when the computer program is executed by a computer.

Additional features and advantages of the disclosure will be set forth in the description which follows, and in part will be apparent from the description, or may be learned by practice of the embodiments of the disclosure. The objectives and other advantages of the disclosure may be realized and attained by the structure particularly pointed out in the written description and appended drawings.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure or related technologies, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings in the following description are only some implementations of the present disclosure For example, those of ordinary skill in the art can also obtain other drawings based on these drawings on the premise of not paying creative efforts. The above and other objects, features and advantages of the present disclosure will be more clearly illustrated by the accompanying drawings. Like reference numerals designate like parts throughout the drawings. The drawings are not drawn to scale to actual size, emphasis is placed on illustrating the gist of the disclosure.

FIG. 1 shows a flow chart of an image processing method provided by an embodiment of the present disclosure.

Fig. 2 shows a structural block diagram of an image processing apparatus provided by an embodiment of the present disclosure.

Fig. 3 shows a schematic structural diagram of an electronic device provided by an embodiment of the present disclosure.

Reference numerals: 100—electronic device; 110—processor; 120—memory; 130—display screen; 400—image processing device; 410—obtaining module; 420—determining module; 430—dividing module;

detailed description

The technical solutions in the embodiments of the present disclosure will be described below with reference to the drawings in the embodiments of the present disclosure.

It should be noted that similar numbers and letters denote similar items in the following drawings, therefore, after an item is defined in one drawing, it does not require further definition and explanation in subsequent drawings . Meanwhile, relational terms such as "first", "second", etc. in the description of the present disclosure are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that these entities or operations, any such actual relationship or order exists. Moreover, the term "comprising" or any other variation thereof is intended to cover a non-exclusive inclusion such that a process, method, article or apparatus comprising a set of elements includes not only those elements but also other elements not expressly listed, Or also include elements inherent in such a process, method, article or apparatus. Without further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article or apparatus comprising said element.

Furthermore, the term "and/or" in the present disclosure is only an association relationship describing associated objects, indicating that there may be three relationships, for example, A and/or B may indicate that A exists alone, and A and B exist simultaneously. B, there are three situations of B alone.

In recent years, artificial intelligence-based computer vision, deep learning, machine learning, image processing, image recognition and other technologies have made important progress. Artificial Intelligence (AI) is an emerging science and technology that researches and develops theories, methods, technologies and application systems for simulating and extending human intelligence. The subject of artificial intelligence is a comprehensive subject that involves many technologies such as chips, big data, cloud computing, Internet of Things, distributed storage, deep learning, machine learning, and neural networks. As an important branch of artificial intelligence, computer vision is specifically to allow machines to recognize the world. Computer vision technology usually includes face recognition, liveness detection, fingerprint recognition and anti-counterfeiting verification, biometric recognition, face detection, pedestrian detection, target detection, pedestrian detection, etc. Recognition, image processing, image recognition, image semantic understanding, image retrieval, text recognition, video processing, video content recognition, 3D reconstruction, virtual reality, augmented reality, simultaneous localization and mapping (SLAM), computational photography, robot navigation and positioning and other technologies. With the research and progress of artificial intelligence technology, this technology has been applied in many fields, such as security prevention and control, urban management, traffic management, building management, park management, face access, face attendance, logistics management, warehouse management , robots, intelligent marketing, computational photography, mobile imaging, cloud services, smart home, wearable devices, unmanned driving, automatic driving, smart medical care, face payment, face unlock, fingerprint unlock, witness verification, smart screen, smart TV, video camera, mobile Internet, webcast, beauty, makeup, medical cosmetology, intelligent temperature measurement and other fields.

With the development of pan-entertainment trends, there are more and more demands for beauty and face-lifting in various fields. However, the current face-lifting technology has the problem of face distortion, which cannot meet the needs of users.

In order to solve the above problems, the embodiments of the present disclosure provide an image processing method, device, electronic equipment and computer-readable storage medium, which can improve the visual effect of the image to be processed obtained after face-thinning processing. In addition, the defects in the face-slimming technology in related technologies (presenting unnatural visual effects and even causing distortion of human faces) are the result of the applicant's practice and careful research. Therefore, the above-mentioned defects The discovery process and the solutions proposed by the embodiments of the present disclosure hereinafter for the above defects should be recognized as the applicant's contribution to the present disclosure.

This technology can be realized by using corresponding software, hardware and a combination of software and hardware. The embodiments of the present disclosure will be described in detail below.

First, an embodiment of the present disclosure provides an image processing method for performing face thinning processing on an image to be processed including a human face area. Referring to Figure 1, the method may include the following steps:

Step S110: Acquiring face key point information of an image to be processed, where the image to be processed includes a face area.

Step S120: According to the face key point information, determine a moving reference point when moving pixels in the face area;

Step S130: Divide the face area into a plurality of local areas according to the key point information of the face;

Step S140: For each local area, based on the pixel movement strategy parameter corresponding to the local area, move each pixel in the local area toward the moving reference point.

In the embodiment of the present disclosure, instead of adjusting all the images to be processed according to the fixed distance specified by the facial contour template, the face area is first divided into different local areas, and there are corresponding pixels in different local areas Move the strategy parameters, then when adjusting, the pixels included in each local region will move in the direction of the moving reference point according to the pixel movement strategy parameters corresponding to the local region to which it belongs, so that the final face-lifting effect can be obtained as much as possible Naturally, it can improve the visual effect after face-lifting.

Each step in FIG. 1 will be described in detail below.

Wherein, the image processing method provided by the embodiments of the present disclosure may perform real-time processing on the image to be processed, or may perform post-processing on the image to be processed, that is, non-real-time processing.

When real-time processing is performed on the image to be processed, the image processing method can be applied to application scenarios such as video live broadcast, video conference, and portrait photography. In this embodiment, the image to be processed can be determined according to the pictures and/or video streams collected by the camera in real time.

When performing post-processing on the image to be processed, the image processing method may be applicable to an image processing application scenario. In this embodiment, the image to be processed may be determined according to the pre-downloaded picture and/or video stream, or the picture and/or video stream taken by the camera in advance.

Certainly, the camera may be a built-in component of the electronic device that executes or invokes the image processing method, or may be an external component of the electronic device. In an optional implementation manner, in the above step S110, the process of acquiring face key point information may be from a third-party application program, software, face key point detection model or other The process in which the device obtains the face key point information for the image to be processed may also be the process of performing face key point detection on the image to be processed through the face key point model; that is, when executing the method provided by the embodiment of the present disclosure , the acquired original parameters may be face key point information, or an image to be processed including a face area, and its specific implementation process may be selected according to an actual application scenario, which is not limited in the embodiments of the present disclosure.

In an optional implementation manner, the image processing method provided by the embodiment of the present disclosure may also include the process of performing face key point detection on the image to be processed, that is to say, the image processing method obtained by the electronic device The original parameter is the image to be processed including the face area, and then the obtained image to be processed is input to a face key point detection model with a face key point detection function for detection to obtain face key point information.

In this embodiment, in order to make the human face key point detection model have the function of detecting human face key points, it needs to be trained in advance. Optionally, before using the method provided by the embodiment of the present disclosure to process the image , to train the above face key point detection model, the training process is as follows.

Obtain a large number of pictures containing face regions, and label each picture to form a training set S including multiple samples.

Wherein, for the i-th sample x _i in the training set S, assuming that the corresponding label is y _i , then y _i may include position information G of each face key point in x _i . Among them, G=[(a _i1 , b _i1 ), (a _i2 , b _i2 )...(a _in , bin _in )], n is the identification information of key points of the face, such as number, ID (Identity Document, identity identification ), etc., (a _in , bin _in ) represent the coordinate information of the face key point identified as n in the i-th sample x _i .

It is worth pointing out that in the embodiments of the present disclosure, the encoding rules of the identification information of the key points of the face in each sample are set in advance, so that the key points of the face with the same identification information in different samples are characterized by The meaning is the same, and the identification information of the key points of the face belonging to a specific local area or a specific facial organ in the human face is limited to the range of identification information corresponding to the specific local area or the specific facial organ.

For example, in some implementations, the identification information can be the number of the key points of the face. When it is necessary to mark 81 key points of the face for each face, the preset identification information encoding rule can be: the eyes including the eyes The above forehead is regarded as an area, and the number of face key points belonging to this area ranges from 1-20; the chin area below the mouth including the mouth is regarded as an area, and the number of face key points belonging to this area The range is 21-40; the left face is regarded as a region, and the number range of face key points belonging to this region is 41-55; the right face is regarded as a region, and the number range of face key points belonging to this region is 56-70; the hairline is regarded as a region, and the number range of the face key points belonging to this region is 71-81.

For another example, in some implementations, the identification information may be the number of the key points of the face. When it is necessary to mark 81 key points of the face for each face, the preset identification information encoding rule may be: The numbering range of the face key points of the face contour is 1-20, the numbering range of the face key points belonging to the mouth in the face organ is 21-40, and the numbering of the face key points belonging to the nose in the face organ The range is 41-55, the number range of the face key points belonging to the eyes in the face organs is 56-70, and the number range of the face key points belonging to the eyebrows in the face organs is 71-81.

Certainly, the above identification information encoding rule is only an example, and it can be understood that in other implementation manners, other similar schemes may be adopted for the identification information encoding rule.

After the labeling is completed, the deep learning model can be trained through the training set S. The training process can be as follows: input each sample picture in the training set S to the deep learning model, and obtain the corresponding output (the key points of the face of the sample picture and its Coordinate information), and let the deep learning model automatically learn the internal correlation between the sample picture and the output, so as to obtain the face key point detection model.

Generally speaking, in the labeling stage, it is necessary to label N key points of the face for each sample, and the subsequent training obtains the key point detection model of the face to detect the key points of the face for the input image to be processed, and the key points of the face outputted The information may include N key points of the face with identification information and coordinate information thereof. For example, in the labeling stage, 81 face key points are marked for each sample, and the face key point detection model performs face key point detection on the input image to be processed, and the output face key point information includes 81 face key points with labels The face key points and their coordinate information of the information.

In addition, in some optional implementation manners, in the tagging stage, the tag y _i may also include information K=(u _i , v _i , m, f) of the face positioning frame in the sample _xi , where (u _i , vi ₎ represent the coordinate information of the face positioning frame, generally the coordinate information of a vertex (such as the point at the lower left corner) of the face positioning frame, and (m, f) represent the width and height of the face positioning frame respectively.

It is worth pointing out that the coordinate information of the lower left corner of the face positioning frame and the coordinate information of each key point of the face belong to the same rectangular coordinate system (for the sake of distinction, it is called the first coordinate system). A vertex of x _i (such as the point where the lower left corner is located) is used as the origin, and the two edges connected to the vertex are used as the X axis and the Y axis respectively.

In this embodiment, after the face key point detection model obtained by training performs face key point detection on the image to be processed, the outputted face key point information may include the identification information included in the image to be processed In addition to each face key point and its coordinate information, it can also include the information of the face positioning frame included in the image to be processed, that is, the output face key point information includes G and K.

After obtaining the face key point detection model, you can input the obtained image to be processed into the face key point detection model, and perform face key point detection on the image to be processed through the face key point detection model, and then obtain the face key point The face key point information output by the point detection model.

Wherein, in some implementations, the acquired image to be processed may be a face image including only the face area, or a large image including the face area and other areas of the human body.

Generally speaking, when the image to be processed is a large image, the subsequent face thinning process can be directly based on the large image, targeting the face area of the large image as the processing object.

In addition, because the face thinning process mainly performs image processing on the face area, in some optional implementations, when the image to be processed is a large image, and the image to be processed is input to the face key point detection model, the obtained When the obtained face key point information includes the information of the face positioning frame, the face image corresponding to the face positioning frame can also be intercepted from the image to be processed (i.e. the large image) according to the information of the obtained face positioning frame , so that the face area of the face image can be used as the processing object of the subsequent face thinning process directly on the basis of the face image, without processing the rest of the large image.

It can be understood that the data volume of face images including the same face area is smaller than the data volume of the image to be processed. Therefore, when the face image is used as the processing object, it is beneficial to reduce the time delay generated in the face thinning process.

Of course, since the coordinate origin of the coordinate system (the first coordinate system) where the various coordinate information output by the face key point detection model is located is a vertex of the large image, when the face image is used as the processing object, the coordinate origin is large The probability is outside the face image.

In order to operate only on the pixels in the face area in the face image, in some implementations, the various coordinate information output by the face key point detection model can also be normalized, so that the face The coordinate information of the face key points output by the face key point detection model is converted into new coordinate information in the intercepted face area. In a specific implementation, before performing the above step S120 of determining the moving reference point when moving the pixel points in the face area according to the key point information of the face, the above normalization processing operation may be performed. Certainly, before performing the above step S130 of dividing the human face area into a plurality of local areas according to the key point information of the human face, the above normalization processing operation may be performed. The embodiment of the present disclosure does not limit the specific execution process of the above normalization processing operation.

During specific implementation, the coordinate information of the key points of the face can be normalized based on the coordinate information of the face positioning frame, and the normalized coordinate information of the key points of the face can be obtained.

In this way, the subsequent steps of the method provided by the embodiments of the present disclosure are executed based on the normalized coordinate information of key points of the human face. That is, the coordinates of the key points of the human face are converted into the coordinates in the human face area of the intercepted human face image, so that only the pixels in the human face area need to be operated when performing the face thinning operation, and the human face area The pixels outside remain unchanged. After the face thinning operation is completed, the face area can be redrawn according to the pixels after the operation, and the original face area can be replaced.

Wherein, the origin of the coordinate system (the first coordinate system) corresponding to the coordinate information before the normalization operation can be one of the vertices of the image to be processed, and the coordinate system (the second coordinate system) corresponding to the coordinate information after the normalization operation system) can be one of the vertices of the face positioning frame referenced for the normalization operation.

Correspondingly, when the face image is used as the processing object, processing can be performed based on the normalized key point information of the face.

The process of normalization conversion will be introduced below.

Optionally, the coordinate difference between the coordinate information of the face positioning frame and the coordinate information of the key points of the face can be calculated, and then the coordinate difference can be used to update and replace the original coordinate information of the key points of the face to obtain the returned The unified face key point information.

Optionally, suppose that in the i-th image to be processed, the information of the face positioning frame is (u _i , v _i , m, f), and the coordinate information of the key points of the face is ((a _i1 , b _i1 ), ( a _i2 , b _i2 )...(a _in , b _in )), then the coordinate information of the face positioning frame is (u _i , v _i ). After normalized transformation, the coordinate information of each face key point included in the face image can be ((a _i1 -u _i , b _i1 -v _i ), (a _i2 -u _i , b _i2 -v _i )...(a _in -u _i , bin _in -v _i )), at this time, the coordinate origin is the point where the coordinate information of the face positioning frame is located.

Of course, it can be understood that the above-mentioned embodiment corresponding to normalization can be combined with any of the preceding or following embodiments (the implementation of any embodiment needs not to conflict with the implementation of this embodiment), for example : This embodiment can be combined with the embodiment corresponding to moving each pixel point in the local area towards the direction of the moving reference point based on the pixel movement strategy parameter in the foregoing; this embodiment can be combined with the embodiment in the following, based on The pixel position adjustment ratio included in the pixel movement strategy parameter is combined with the embodiment corresponding to moving each pixel point in the local area toward the direction of the mobile reference point; The face-thinning strength coefficient included in the face adjustment instruction and the acquired position information of each pixel in the face area that has been moved before and after the movement are used to determine the corresponding position of each pixel in the face area that has been moved. Embodiments corresponding to the target location information are combined and the like.

Step S120: According to the key point information of the human face, determine a moving reference point when moving the pixel points in the human face area.

Wherein, the moving reference point is configured to play a guiding role in the subsequent face thinning process.

In the embodiment of the present disclosure, all pixel points in the entire face area may correspond to the same moving reference point; in addition, each pixel point in the entire face area may also correspond to different moving reference points.

As mentioned above, the meaning represented by the key points of the face corresponding to each identification information is determined in advance.

In some optional implementations, according to the identification information of each key point of the face, the key point of the face corresponding to a certain position (such as the center of the eyes or the position of the tip of the nose, etc.) can be determined from the face area, and the The face key point is determined as a reference point for movement.

In a specific implementation, the above step S120: according to the key point information of the human face, determine the moving reference point when moving the pixels in the human face area, which may include at least the following two implementations:

The key point of the face corresponding to the center position of the eyes in the face area can be determined, and the key point of the face corresponding to the center position of the eyes can be determined as the above-mentioned mobile reference point; or, the face corresponding to the position of the tip of the nose in the face area can be determined key point, the key point of the face corresponding to the position of the tip of the nose is determined as the above-mentioned moving reference point.

Of course, the above is just an example of the specific position being the center position of the eyes and the position of the tip of the nose. In addition, the specific position above can also be other, such as the midpoint of the line connecting the center position of the eyes and the position of the tip of the nose, etc.; Embodiments of the present disclosure will not be described in detail one by one.

For example, in an optional implementation, if the pre-defined number range of face key points belonging to the eye area is 56-70, the face key point numbered 63 is used to represent the center position of the eyes, Then the key point of the face numbered 63 can be determined as the moving reference point.

In this embodiment, the moving reference points corresponding to each pixel point in the face area are the same as the key points of the face used to represent the center positions of the eyes in the face area.

Of course, it can be understood that the above-mentioned embodiment of using a specific position as a moving reference point when moving pixels in the face area can be combined with any of the preceding or following embodiments (the realization of any embodiment needs to be consistent with this embodiment. The implementation of the example does not conflict), for example: this embodiment can be combined with the embodiment corresponding to moving each pixel point in the local area towards the direction of the moving reference point based on the pixel movement strategy parameter; This embodiment can be combined with the embodiment corresponding to moving each pixel point in the local area toward the direction of the moving reference point based on the pixel position adjustment ratio included in the pixel movement strategy parameter; this embodiment can also be As will appear later, the face is determined according to the face-thinning intensity coefficient included in the face adjustment command triggered by the user and the acquired position information of each pixel in the face area that has been moved before and after the move. The embodiments corresponding to the target position information corresponding to the moved pixels in the area are combined and the like.

In some embodiments, a reference line can also be determined, and for different pixel points in the face area, the moving reference point corresponding to the pixel point can be determined from the reference line according to the set rules. Therefore, the above step S120: according to the key point information of the human face, determine the moving reference point when moving the pixels in the human face area, which can also be realized through the following process:

Determine the reference line based on the face key points located on the center line in the vertical direction of the face area; for each pixel in the face area, the face key with the minimum distance between the reference line and the pixel The point is determined as the moving reference point corresponding to the pixel point.

In an optional implementation manner, the key point corresponding to the center position of the eyes on the center line may be used as the first face key point, and the face key point corresponding to the center position of the chin on the center line may be used as the second face key point. Face key points, the line segment formed by the first key point and the second face key point is determined as the above-mentioned reference line.

For example, in an optional implementation manner, assuming that the number range of the face key points belonging to the eye area is 56-70, the number range of the face key points belonging to the face contour area is 1-20, and the number is The key point of the face numbered 63 is the center of the eyes, and the key point of the face numbered 10 is the center of the chin. At this time, the line connecting the face key point numbered 10 and the face key point numbered 63 is determined as the reference line.

Of course, in another optional implementation manner, the key point corresponding to the brow center position on the center line can also be used as the first face key point, and the face key point corresponding to the lip center position on the center line can be used as the second face key point. The face key point is determined as the reference line based on the line segment formed by the first face key point and the second face key point.

On the basis of obtaining the reference line, for each pixel point in the face area, the face key point with the smallest distance between the reference line and the pixel point can be determined as the moving reference point corresponding to the pixel point .

In this embodiment, the moving reference points corresponding to each pixel in the face area are different, that is, each pixel in the face area has a corresponding moving reference point.

Of course, it can be understood that the above-mentioned embodiment of using the key points of the face on the reference line as the moving reference point when moving the pixels in the face area can be the same as any of the preceding or following embodiments (this any embodiment) need to be combined with the implementation of this embodiment), for example: this embodiment can correspond to the movement of each pixel point in the local area towards the direction of the reference point based on the pixel movement strategy parameter that appears above This embodiment can be combined with the embodiment corresponding to moving each pixel point in the local area towards the direction of the moving reference point based on the pixel position adjustment ratio included in the pixel movement strategy parameter that will appear later. ; This embodiment can also appear in the following, according to the thin face strength coefficient included in the face adjustment command triggered by the user and the obtained position information of each pixel point in the face area that has been moved before moving and after moving Combination of the embodiments corresponding to the target position information corresponding to each pixel point in the human face area that has been determined by using the position information, etc.

Step S130: Divide the face area into a plurality of local areas according to the key point information of the face.

As mentioned above, the face key point information may include the identification information of the face key point. Correspondingly, the above step S130 divides the face area into a plurality of local areas according to the face key point information, and optionally including:

According to the corresponding relationship between the acquired identification information of the key points of the face and the facial organs, determine the set of key points of the face corresponding to each organ of the face; The region identified as a local region.

The above-mentioned facial organs can be eyes, nose, mouth, eyebrows, face, etc.

Certainly, in some other optional implementation manners, the division of local areas may not be performed according to the areas corresponding to the above-mentioned human face organs, but may also be divided in other ways, for example, the forehead above the eyes including the eyes is used as One area, the chin area below the mouth including the mouth as one area, the left face area (including a part of the nose) located between the forehead area and the chin area as one area, and the right face area located between the forehead area and the chin area The face area (including a part of the nose) is regarded as an area; of course, in specific implementation, the area division can also be performed in other ways, and the embodiments of the present disclosure will not repeat them one by one.

For ease of understanding, the following examples will be used for illustration.

For example, the face key points whose numbers belong to the number range 1-20 can be divided into one group, and it is determined that the face key points included in this group belong to the face contour area; the faces whose numbers belong to the number range 21-40 can be divided into The key points are divided into one group, and the face key points included in the group are determined to belong to the mouth area; the face key points whose numbers belong to the number range 41-55 can be divided into one group, and the faces included in the group are determined The key points belong to the nose area; the face key points whose number belongs to the number range 56-70 can be divided into a group, and the face key points included in this group can be determined to belong to the eye area; the number belongs to the number range 71-81 The face key points are divided into one group, and it is determined that the face key points included in the group belong to the eyebrow region.

After obtaining multiple groups, each group is a set of face key points. The area surrounded by the coordinate information of each face key point in each face key point set is a local area.

Of course, it can be understood that, according to the above-mentioned corresponding relationship between the acquired identification information of the key points of the face and the facial organs, the set of key points of the face corresponding to each organ of the face can be determined; The embodiment that the area surrounded by each face key point is determined as a local area can be combined with any of the preceding or following embodiments (the realization of any embodiment needs not to conflict with the realization of this embodiment) , for example: this embodiment can be combined with the embodiment corresponding to moving each pixel point in the local area toward the direction of the moving reference point based on the pixel movement strategy parameter; this embodiment can be combined with the embodiment that appears later , based on the pixel position adjustment ratio included in the pixel movement strategy parameters to move each pixel in the local area towards the direction of the reference point; this embodiment can also be combined with the following, according to user trigger The face-thinning intensity coefficient included in the face adjustment instruction and the obtained position information of each pixel point in the face area after moving and the position information after moving are used to determine the position of each pixel point in the face area after moving. The embodiments corresponding to the corresponding target location information are combined and the like.

Step S140: For each local area, based on the pixel movement strategy parameter corresponding to the local area, move each pixel in the local area toward the direction of the moving reference point.

Face thinning operations can be performed, that is, pixel adjustments are performed on each local area divided by the face area that needs to be thinned.

As mentioned above, each local region has a corresponding pixel movement strategy parameter, and the pixel movement strategy parameters corresponding to different local regions are different.

In some implementations, the pixel shift strategy parameter may be a pixel position scaling. In this embodiment, each pixel included in each local area can be moved and adjusted toward the direction of the moving reference point, so as to achieve the effect of thinning the face. As for the degree of adjustment, it is determined by the adjustment ratio of the pixel position corresponding to the local area where the pixel is located.

In this implementation manner, when moving each pixel point in each local area towards the direction of the moving reference point, for each pixel point in each local area, determine the distance between the pixel point and the moving reference point The first distance d; then adjust the ratio T based on the first distance d and the pixel position corresponding to the local area where the pixel point is located, and determine the second distance d' between the pixel point and the mobile reference point; then the pixel point moving towards the direction of the moving reference point, so that the distance between the pixel point and the moving reference point after the movement is equal to the second distance.

In some implementations, a first product value of the first distance d and the pixel position adjustment ratio T corresponding to the local area to which the pixel belongs may be calculated, and the first product value is determined as the second distance d′.

For example, the above process can be realized based on the formula d'=d×T, T∈[0,1].

Taking the moving reference point of each pixel in the face area as the same, which is the face key point 63 at the center of both eyes as an example, assume that the pixel position adjustment ratio corresponding to the face contour area is T1, and the pixel position corresponding to the mouth area The pixel position adjustment ratio is T2, then when performing pixel movement, each pixel point included in the face contour area is moved in the direction where the moving reference point represented by the key point 63 of the face is located, and the moving distance is the pixel The product value of the distance d and T1 of point distance people's face key point 63; Every pixel point included in the mouth contour area is all moved towards the direction where the mobile reference point represented by people's face key point 63 is located, and the moving distance is this The product value of the distance d from the pixel point to the face key point 63 and T2.

Of course, it can be understood that in this specific example, other local areas not enumerated and their corresponding pixel position adjustment ratios may also be included, and the moving reference points represented by the face key points 63 will be moved in a similar manner to the above. Move in the direction in which to move.

In addition, it is worth pointing out that when the moving reference point corresponding to each pixel point in the face area is different, during the above pixel moving process, when each pixel point moves toward the direction where the moving reference point is located, it can be respectively Move towards the direction of the movement reference point corresponding to itself.

In addition, it should be noted that, in the embodiment of the present disclosure, the above-mentioned pixel shift strategy parameter may also be the ratio between the distance between the position after pixel shift and the pixel reference point and the distance between the position before pixel shift and the pixel reference point, In this way, when pixel movement is performed, the target position information of the pixel can be determined according to the ratio.

Certainly, in some optional implementation manners, the above-mentioned pixel strategy movement parameter may also be the distance of pixel movement, etc.;

For the method of determining the moving reference point corresponding to each pixel point, please refer to the relevant description above, which will not be repeated here.

Of course, the pixel position adjustment ratios corresponding to the respective local regions are independent of each other, and may be partly the same or completely different.

In the embodiment of the present disclosure, in order to ensure a natural effect after the face is thinned, background staff configure corresponding pixel position adjustment ratios for each local area of the face in advance. Among them, the background staff can take different pixel position adjustment ratios for each local area to test, and observe the corresponding visual effect, so as to determine the optimal pixel position adjustment ratio corresponding to each local area, and save it. During actual use, the user cannot directly adjust the pixel position adjustment ratio of each local area separately.

Of course, in some implementations, after the above-mentioned pixel shifting operation, the effect may not be satisfactory to the user. In this case, the user can also trigger custom adjustments through virtual buttons or physical buttons, and then trigger face recognition. Adjust instructions.

Wherein, the face-thinning intensity coefficient k may be included in the face adjustment instruction, and the size of the face-thinning intensity coefficient k can be adjusted by the user, so that the user can adjust the current face-thinning degree based on the face-thinning intensity coefficient k.

In the embodiment of the present disclosure, when the electronic device running the image processing method acquires and responds to the face adjustment instruction (wherein the face adjustment instruction carries the face-thinning intensity coefficient), it can acquire the processed image (which may be the image to be processed, It may also be the position information of each pixel in the face area cut out from the image to be processed before the movement and the position information after the movement. Then, based on the face-thinning intensity coefficient k, the position information before the movement and the position information after the movement, determine the face area of the processed image (which can be the image to be processed, or the face image intercepted from the image to be processed). The target position information corresponding to each moved pixel point, and move each moved pixel point in the face area to the target position corresponding to the target position information.

In an optional implementation manner, the above position information can be characterized by coordinate information; correspondingly, based on the face thinning intensity coefficient k, the position information before the movement and the position information after the movement, determine the face area of the processed image that has been moved The target position information corresponding to each pixel of , can be realized through the following process:

For each pixel point that has been moved in the face area, determine the first coordinate difference between the second coordinate information (x′ _i , y′ _i ) and the first coordinate information ( _xi , y _i ), and then calculate the thin face The second product value between the intensity coefficient k and the first coordinate difference, and the sum of the first coordinate information ( _xi , y _i ) and the second product value is determined as the target position information

Wherein, the location information before the movement can be represented by the first coordinate information, and the location information after the movement can be represented by the second coordinate information.

For example the above procedure could be based on the formula

To achieve, k∈[0,1].

Of course, it is worth pointing out that when the processing object is an image to be processed when performing face-lifting processing, the coordinate information involved above may be coordinate information belonging to the first coordinate system; When the human face image is displayed, the above-mentioned coordinate information may be coordinate information belonging to the second coordinate system.

Optionally, in some embodiments, it is also possible to determine the processed image (which may be the image to be processed, or the The movement information of each pixel on the X-axis and the Y-axis (the movement information includes the moving distance and the moving direction) in the human face area of the human face image), and each pixel in the human face area is moved based on the above-mentioned movement information.

Optionally, in an optional implementation manner, the above position information may be represented by coordinate information, and correspondingly, the movement information corresponding to each of the above pixel points may be determined through the following process:

Determine the first coordinate difference between the second coordinate information (x′ _i , y′ _i ) and the first coordinate information ( _xi , y _i ) for each pixel point that has been moved in the face area; then calculate the design The difference between the fixed value and the face-lifting intensity coefficient k, calculate the product between the difference and the first coordinate difference, and determine the product value as the above-mentioned moving information, and the above-mentioned product value can be positive or negative Value, the positive or negative of the above product value indicates the direction of movement.

Of course, the embodiments of the present disclosure only exemplify possible implementations of face thinning based on the face thinning intensity coefficient. In addition, face thinning can also be performed in other ways based on the face thinning intensity coefficient, which will not be repeated here.

In addition, in some implementations, when the processing object of the face thinning process is a human face image cut out from the image to be processed, after the face thinning process is performed on the face image, it is also necessary to use the face after the face thinning process The image replaces the original face image included in the image to be processed, so that the image to be processed presents a face-thinning effect.

In addition, referring to FIG. 2 , an embodiment of the present disclosure further provides an image processing apparatus 400 , which may include: an acquisition module 410 , a determination module 420 , a division module 430 and an adjustment module 440 .

The obtaining module 410 is configured to obtain face key point information of an image to be processed, and the image to be processed may include a face area;

The determination module 420 is configured to determine a moving reference point when moving pixels in the human face area according to the key point information of the human face;

The dividing module 430 is configured to divide the human face area into a plurality of local areas according to the key point information of the human face;

The adjustment module 440 is configured to, for each of the local areas, move each pixel in the local area toward the moving reference point based on the pixel movement strategy parameter corresponding to the local area.

In a possible implementation manner, the pixel moving strategy may include a pixel position adjustment ratio; the adjustment module 440 is configured to determine the relationship between the pixel point and the pixel point for each pixel point in the local area. moving a first distance between reference points; adjusting a ratio based on the first distance and a corresponding pixel position, determining a second distance between the pixel point and the moving reference point; moving the pixel point toward the The direction of the moving reference point is moved, so that the distance between the pixel point and the moving reference point after the movement is equal to the second distance.

In a possible implementation manner, the adjustment module 440 is configured to determine a first product value of the first distance and a pixel position adjustment ratio corresponding to the local area to which the pixel point belongs; A product value is determined as the second distance.

In a possible implementation manner, the adjustment module 440 is further configured to, in response to a user-triggered face adjustment instruction, obtain the position information and movement information of each pixel in the face area that has been moved before the movement. post-movement position information; wherein, the face adjustment instruction may carry a face-thinning intensity coefficient; based on the face-thinning intensity coefficient, the position information before the movement, and the position information after the movement, determine the face area the target position information corresponding to each pixel point that has been moved; and move each pixel point that has been moved in the face area to the target position corresponding to the target position information.

In a possible implementation manner, the position information is represented by coordinate information; the adjustment module 440 is configured to determine the second coordinate information and the first coordinate information for each pixel point in the face area that has been moved The first coordinate difference between the information; wherein, the position information before the movement can be represented by the first coordinate information, and the position information after the movement can be represented by the second coordinate information; calculate the thin face A second product value between the intensity coefficient and the first coordinate difference; determining the sum of the first coordinate information and the second product value as the target position information.

In a possible implementation manner, the determination module 420 is configured to determine the key points of the human face corresponding to the center position of the eyes in the face area, and determine the key point of the human face corresponding to the center position of the eyes is said mobile reference point;

Or, it is configured to determine the key point of the human face corresponding to the position of the tip of the nose in the human face area, and determine the key point of the human face corresponding to the position of the tip of the nose as the moving reference point.

In a possible implementation manner, the determination module 420 is configured to determine a reference line based on the key points of the face located on the center line in the vertical direction of the face area; For each pixel point, the face key point with the minimum distance between the reference line and the pixel point is determined as the moving reference point corresponding to the pixel point;

Correspondingly, the adjustment module 440 is configured to move each pixel point in the local area toward a direction corresponding to the moving reference point.

In a possible implementation manner, the key point information of the face may include identification information of the key points of the face, and the division module 430 is configured to The corresponding relationship of each face organ is determined to determine the face key point set corresponding to each face key point set; the area surrounded by each face key point in each face key point set is determined as a local area.

In a possible implementation manner, pixel position adjustment ratios corresponding to different local regions are different.

In a possible implementation manner, the acquisition module 410 is configured to input the image to be processed into a human face key point detection model, and perform artificial intelligence on the image to be processed through the human face key point detection model. Facial key point detection; obtaining the human face key point information output by the human face key point detection model.

In a possible implementation manner, the face key point information may include the coordinate information of the face positioning frame and the coordinate information of the face key point; the device may also include a normalization module configured to The coordinate information of the key points of the face is normalized based on the coordinate information of the face positioning frame, and the normalized coordinate information of the key points of the face is obtained.

The image processing device 400 provided by the embodiment of the present disclosure has the same realization principle and technical effect as the aforementioned method embodiment. For a brief description, for the part not mentioned in the device embodiment, please refer to the corresponding content in the aforementioned method embodiment. .

In addition, an embodiment of the present disclosure also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is run by a computer, the steps included in the above-mentioned image processing method are executed.

In addition, referring to FIG. 3 , an embodiment of the present disclosure further provides an electronic device 100 for implementing an image processing method and apparatus.

The electronic device 100 may be a mobile phone, a smart camera, a tablet computer, a personal computer (Personal computer, PC) and other devices. Users can use the electronic device 100 to perform activities such as taking pictures, live video broadcasting, and image processing.

Wherein, the electronic device 100 may include: a processor 110 , a memory 120 , and a display screen 130 .

It should be noted that the components and structure of the electronic device 100 shown in FIG. 3 are only exemplary rather than limiting, and the electronic device 100 may also have other components and structures as required. For example, in some cases, the electronic device 100 may further include a camera configured to capture images to be processed in real time.

The processor 110 , the memory 120 , the display screen 130 and other components that may appear in the electronic device 100 are electrically connected to each other directly or indirectly to realize data transmission or interaction. For example, the processor 110, the memory 120, the display screen 130 and other possible components may be electrically connected to each other through one or more communication buses or signal lines.

The memory 120 is used to store a program, for example, a program corresponding to the above-mentioned image processing method or the above-mentioned image processing device is stored. Optionally, when the image processing device is stored in the memory 120, the image processing device may include at least one software function module that may be stored in the memory 120 in the form of software or firmware (firmware).

Optionally, the software function modules included in the image processing apparatus may also be solidified in an operating system (operating system, OS) of the electronic device 100 .

The processor 110 is configured to execute executable modules stored in the memory 120, such as software function modules or computer programs included in the image processing apparatus. After the processor 110 receives the execution instruction, it can execute the computer program, for example, execute: acquire the face key point information of the image to be processed, the image to be processed includes the face area; according to the face key point information, determine A moving reference point when moving the pixels in the face area; according to the key point information of the face, dividing the face area into a plurality of local areas; for each of the local areas, based on the The parameter of the pixel movement strategy corresponding to the local area is used to move each pixel in the local area toward the direction of the moving reference point.

Certainly, the method disclosed in any embodiment of the present disclosure may be applied to the processor 110 or implemented by the processor 110 .

The embodiments of the present disclosure also provide a computer program, which executes the method provided by the embodiments of the present disclosure and/or in combination with any possible implementation manners of the embodiments of the present disclosure when the computer program is run by a computer.

To sum up, the image processing method, device, electronic device, and computer-readable storage medium proposed by the embodiments of the present disclosure, when the image to be processed needs to be thinned, can first obtain the face key point information of the image to be processed, and then According to the face key point information, determine the mobile reference point and divide the face area into different local areas, then when performing face thinning processing, move the pixels included in each local area according to the pixels corresponding to the local area to which it belongs The strategy moves in the direction of moving the reference point, so that the final face-lifting effect can be as natural as possible, and the visual effect after the face-lifting can be improved.

It should be noted that each embodiment in this specification is described in a progressive manner, each embodiment describes the difference from other embodiments, and the same and similar parts of each embodiment can be referred to each other .

In the several embodiments provided in the present disclosure, it should be understood that the disclosed devices and methods may also be implemented in other ways. The device embodiments described above are only illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the architecture, functions and possible implementations of devices, methods and computer program products according to multiple embodiments of the present disclosure. operate. In this regard, each block in a flowchart or block diagram may represent a module, program segment, or portion of code that contains one or more logical functions configured to implement the specified executable instructions. It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks in succession may, in fact, be executed substantially concurrently, or they may sometimes be executed in the reverse order, depending upon the functionality involved. It should also be noted that each block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or flowchart illustrations, can be implemented by a dedicated hardware-based system that performs the specified function or action , or may be implemented by a combination of dedicated hardware and computer instructions.

In addition, each functional module in each embodiment of the present disclosure may be integrated together to form an independent part, each module may exist independently, or two or more modules may be integrated to form an independent part.

If the functions are realized in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the essence of the technical solution of the present disclosure or the part that contributes to the related technology or the part of the technical solution can be embodied in the form of a software product, and the computer software product can be stored in a storage medium, including Several instructions are used to make a computer device (which may be a personal computer, a notebook computer, a server, or a network device, etc.) execute all or part of the steps of the methods described in various embodiments of the present disclosure. The aforementioned storage medium can include: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disc, etc., which can store program codes. medium.

The above are only optional implementations of the present disclosure, but the scope of protection of the present disclosure is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope of the present disclosure. , these changes or substitutions should all fall within the protection scope of the present disclosure.

Industrial Applicability

The disclosure provides an image processing method, device, electronic equipment and computer-readable storage medium. When the image to be processed needs to be thinned, the key point information of the face of the image to be processed is obtained first, and then according to the key point information of the face, the reference point for moving the pixels in the face area is determined and the face is moved. The area is divided into different local areas. When performing face thinning processing, the pixels included in each local area are moved in the direction of the moving reference point according to the pixel movement strategy parameters corresponding to the local areas to which they belong, so that the final obtained The effect of face-lifting is as natural as possible, which can improve the visual effect after face-lifting.

In addition, it can be understood that the image processing method, device, electronic device, and computer-readable storage medium of the present disclosure are reproducible, and can be used in various industrial applications, for example, in the field of image processing.

Claims

An image processing method, characterized in that the method comprises:

Acquiring face key point information of an image to be processed, the image to be processed includes a face area;

According to the key point information of the human face, determine a moving reference point when moving the pixels in the human face area;

According to the key point information of the human face, the human face area is divided into a plurality of local areas;

For each local area, each pixel in the local area is moved toward the moving reference point based on the pixel movement strategy parameter corresponding to the local area.
The method according to claim 1, wherein the pixel movement strategy parameters include a pixel position adjustment ratio;

Correspondingly, the moving each pixel in the local area toward the moving reference point based on the pixel movement strategy parameter corresponding to the local area includes:

For each pixel in the local area, determine a first distance between the pixel and the moving reference point;

determining a second distance between the pixel point and the moving reference point based on the first distance and a corresponding pixel position adjustment ratio;

Moving the pixel point toward the direction of the moving reference point, so that the distance between the pixel point and the moving reference point after moving is equal to the second distance.
The method according to claim 2, wherein said determining a second distance between said pixel point and said moving reference point based on said first distance and said corresponding pixel position adjustment ratio comprises :

determining a first product value of the first distance and a pixel position adjustment ratio corresponding to the local area to which the pixel belongs;

The first product value is determined as the second distance.
The method according to claim 1, characterized in that, after moving each pixel in the local area toward the moving reference point based on the pixel movement parameters corresponding to the local area, the Methods also include:

In response to the face adjustment command triggered by the user, the position information of each pixel point in the face area that has been moved before the movement and the position information after the movement are obtained; wherein, the face adjustment command carries a face-thinning intensity coefficient ;

Based on the face-thinning intensity coefficient, the position information before the movement, and the position information after the movement, determine the target position information corresponding to each pixel point in the face area that has been moved;

Moving the moved pixels in the face area to the target position corresponding to the target position information.
The method according to claim 4, wherein the location information is represented by coordinate information;

The determining the target position information corresponding to each pixel point in the face area that has been moved based on the face-thinning intensity coefficient, the position information before the movement, and the position information after the movement includes:

For each pixel in the face area that has been moved, determine a first coordinate difference between the second coordinate information and the first coordinate information; wherein, the position information before the movement is represented by the first coordinate information , the moved position information is represented by the second coordinate information;

calculating a second product value between the face-thinning strength coefficient and the first coordinate difference;

A sum of the first coordinate information and the second product value is determined as the target position information.
The method according to claim 4 or 5, characterized in that the size of the face-thinning intensity coefficient is adjustable.
The method according to any one of claims 1 to 6, wherein, according to the key point information of the human face, determining the moving reference point when moving the pixel points in the human face area includes:

Determining the key point of the face corresponding to the center position of the eyes in the face area, and determining the key point of the face corresponding to the center position of the eyes as the moving reference point;

or,

Determining a human face key point corresponding to a nose tip position in the human face area, and determining the human face key point corresponding to the nose tip position as the moving reference point.
The method according to any one of claims 1 to 6, wherein, according to the key point information of the human face, determining the moving reference point when moving the pixel points in the human face area includes:

Determining a reference line based on the key points of the face located on the center line in the vertical direction of the face area;

For each pixel point in the human face area, determine the human face key point with the smallest distance between the reference line and the pixel point as the moving reference point corresponding to the pixel point;

Correspondingly, the moving each pixel point in the local area toward the direction of the moving reference point includes:

Each pixel point in the local area is moved towards the direction corresponding to the moving reference point.
The method according to any one of claims 1 to 6, wherein the facial key point information includes identification information of human face key points, and according to the human face key point information, the The face region is divided into multiple local regions, including:

According to the corresponding relationship between the acquired identification information of the key points of the face and the facial organs, determine the set of key points of the face corresponding to each of the facial organs;

The area surrounded by each face key point in each face key point set is determined as a local area.
The method according to claim 2 or 3, characterized in that the pixel position adjustment ratios corresponding to different local areas are different.
The method according to any one of claims 1 to 6, wherein said acquisition of the face key point information of the image to be processed comprises:

The image to be processed is input to a human face key point detection model, and the human face key point detection is carried out to the image to be processed by the human face key point detection model;

The human face key point information output by the human face key point detection model is acquired.
The method according to any one of claims 1 to 6, wherein the face key point information includes the coordinate information of the face positioning frame and the coordinate information of the face key point; Face key point information, before determining the moving reference point when moving the pixels in the human face area, the method also includes:

The coordinate information of the key points of the face is normalized based on the coordinate information of the face positioning frame, and the normalized coordinate information of the key points of the face is obtained.
An electronic device, characterized in that it includes: a memory and a processor, the memory is connected to the processor;

The memory is used to store programs;

The processor invokes a program stored in the memory to perform the method according to any one of claims 1 to 12.
A computer-readable storage medium, characterized in that a computer program is stored thereon, and when the computer program is run by a computer, the method according to any one of claims 1 to 12 is executed.
A computer program product comprising code instructions which, when executed by a processor, cause the processor to perform the method according to any one of claims 1 to 12.