CN112734632A - Image processing method, image processing device, electronic equipment and readable storage medium - Google Patents
Image processing method, image processing device, electronic equipment and readable storage medium Download PDFInfo
- Publication number
- CN112734632A CN112734632A CN202110009523.2A CN202110009523A CN112734632A CN 112734632 A CN112734632 A CN 112734632A CN 202110009523 A CN202110009523 A CN 202110009523A CN 112734632 A CN112734632 A CN 112734632A
- Authority
- CN
- China
- Prior art keywords
- image
- migrated
- initial
- key feature
- target
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 18
- 238000012545 processing Methods 0.000 title claims abstract description 9
- 239000011159 matrix material Substances 0.000 claims abstract description 55
- 230000005012 migration Effects 0.000 claims abstract description 49
- 238000013508 migration Methods 0.000 claims abstract description 49
- 238000000034 method Methods 0.000 claims description 31
- 230000007704 transition Effects 0.000 claims description 22
- 239000002131 composite material Substances 0.000 claims description 17
- 230000002194 synthesizing effect Effects 0.000 claims description 2
- 230000002123 temporal effect Effects 0.000 claims description 2
- 238000012549 training Methods 0.000 abstract description 15
- 230000036544 posture Effects 0.000 description 20
- 230000008569 process Effects 0.000 description 7
- 210000000323 shoulder joint Anatomy 0.000 description 7
- 238000010586 diagram Methods 0.000 description 5
- 210000002310 elbow joint Anatomy 0.000 description 5
- 210000000544 articulatio talocruralis Anatomy 0.000 description 4
- 210000000511 carpometacarpal joint Anatomy 0.000 description 4
- 210000004394 hip joint Anatomy 0.000 description 4
- 210000000629 knee joint Anatomy 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 210000003857 wrist joint Anatomy 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000004590 computer program Methods 0.000 description 2
- 210000001508 eye Anatomy 0.000 description 2
- 210000002414 leg Anatomy 0.000 description 2
- 230000033001 locomotion Effects 0.000 description 2
- 102000020897 Formins Human genes 0.000 description 1
- 108091022623 Formins Proteins 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 210000005069 ears Anatomy 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 210000003414 extremity Anatomy 0.000 description 1
- 210000004709 eyebrow Anatomy 0.000 description 1
- 210000003128 head Anatomy 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 210000001503 joint Anatomy 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 210000000214 mouth Anatomy 0.000 description 1
- 210000001331 nose Anatomy 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/04—Context-preserving transformations, e.g. by using an importance map
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Artificial Intelligence (AREA)
- Social Psychology (AREA)
- Human Computer Interaction (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Psychiatry (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Image Analysis (AREA)
- Processing Or Creating Images (AREA)
Abstract
The invention provides an image processing method and an image processing device, wherein the image processing method comprises the following steps: acquiring an image to be migrated and a reference image; the image to be migrated includes: a target object of which the posture is to be converted; the reference image includes: a reference object presenting a reference pose; acquiring a first key feature of a target object and a second key feature of a reference object; determining a posture migration matrix according to the first key feature and the second key feature; acquiring an initial image; and determining a target synthetic image according to the attitude migration matrix, the image to be migrated and the initial image. In the embodiment of the invention, a large number of training sample training models are not required to be acquired to obtain the target synthetic image, the complexity of image migration is reduced, the initial image is acquired, and the attitude migration matrix, the image to be migrated and the initial image are adopted to migrate the whole image to be migrated, so that the details of the image to be migrated can be ensured to be displayed in the target synthetic image, and the omission of the details is prevented.
Description
Technical Field
The present invention relates to the field of image processing technologies, and in particular, to an image processing method and apparatus, an electronic device, and a readable storage medium.
Background
The pose transition is to process one image a to make the person P in the image a have the pose of the person H in the other image B, and to obtain a composite image C.
At present, in order to implement pose migration, a plurality of images a, a plurality of images B, and a plurality of images C are used as training samples, an image migration model is trained, and then a new image a and a new image B are processed according to the image migration model to obtain a new composite image C.
According to the posture migration mode, when the image migration model is trained, a large number of training samples need to be prepared, and the training mode is complicated. Moreover, when the image migration model is used for image migration, when the difference between the clothes shapes of the people in the two images is large, the people in the composite image C cannot keep the details of the people P in the original image a, the shapes of the people in different composite images C at different viewing angles and postures are large, in addition, only part of the people in the people may be migrated, and other parts of the people need to be processed again to achieve the migration, which results in a complicated migration process.
Disclosure of Invention
In view of this, the present invention provides an image processing method, which solves the problems of tedious migration process and incomplete migration to some extent.
A first aspect of an embodiment of the present invention provides an image processing method, where the method includes:
acquiring an image to be migrated and a reference image; the image to be migrated comprises: a target object of which the posture is to be converted; the reference image comprises: a reference object presenting a reference pose;
acquiring a first key feature of the target object and a second key feature of the reference object;
determining an attitude migration matrix according to the first key feature and the second key feature;
acquiring an initial image;
and determining a target synthetic image according to the attitude transition matrix, the image to be transitioned and the initial image.
A second aspect of embodiments of the present invention provides an image processing apparatus, including:
the first acquisition module is used for acquiring an image to be migrated and a reference image; the image to be migrated comprises: a target object of which the posture is to be converted; the reference image comprises: a reference object presenting a reference pose;
the second acquisition module is used for acquiring the first key feature of the target object and the second key feature of the reference object;
the first determining module is used for determining a posture migration matrix according to the first key feature and the second key feature;
the third acquisition module is used for acquiring an initial image;
and the second determining module is used for determining a target synthetic image according to the attitude transition matrix, the image to be transitioned and the initial image.
A third aspect of embodiments of the present invention provides an electronic device, which includes a processor, a memory, and a program or instructions stored on the memory and executable on the processor, and when executed by the processor, the program or instructions implement the steps of the method according to the first aspect.
A fourth aspect of embodiments of the present invention provides a readable storage medium, on which a program or instructions are stored, which when executed by a processor implement the steps of the method according to the first aspect.
In the embodiment of the invention, an image to be migrated and a reference image are obtained; the image to be migrated comprises: a target object of which the posture is to be converted; the reference image comprises: a reference object presenting a reference pose; acquiring a first key feature of the target object and a second key feature of the reference object; determining an attitude migration matrix according to the first key feature and the second key feature; acquiring an initial image; and determining a target synthetic image according to the attitude transition matrix, the image to be transitioned and the initial image. In the embodiment of the invention, a large number of training sample training models are not required to be acquired to obtain the target synthetic image, the complexity of image migration is reduced, the initial image is acquired, and the attitude migration matrix, the image to be migrated and the initial image are adopted to migrate the whole image to be migrated, so that the details of the image to be migrated can be ensured to be displayed in the target synthetic image, and the omission of the details is prevented.
Drawings
Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention. Also, like reference numerals are used to refer to like parts throughout the drawings. In the drawings:
FIG. 1 is a flowchart illustrating steps of an image processing method according to an embodiment of the present invention;
FIG. 2 is a diagram illustrating an image processing method according to an embodiment of the present invention;
FIG. 3 is a diagram illustrating another image processing method according to an embodiment of the present invention;
fig. 4 is a block diagram of an image processing apparatus according to an embodiment of the present invention;
fig. 5 is a block diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
Exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention can be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.
Referring to fig. 1, a flowchart illustrating steps of an image processing method according to an embodiment of the present invention is shown, where the image processing method specifically includes the following steps:
Wherein, the image to be migrated is m1*n1Image of x 3, wherein m1Is the width, n, of the image to be migrated1Is the height of the image to be migrated, 3 means that the image to be migrated is an RGB image. Reference picture is m2*n2Image of x 3, wherein m2Is the width of the reference picture, n2Is the high of the reference image, 3 means that the reference image is an RGB image.
In the embodiment of the present invention, the target object and the reference object generally refer to human objects in an image; referring to fig. 2, an image a is an image to be migrated, and an image B is a reference image; the image A to be migrated comprises a target object P, and the reference image B comprises: reference object H.
In the embodiment of the present invention, the user may select the image to be migrated and the reference image from the image memory according to the requirement, or may capture the image at any time, which is not limited to this.
In addition, a user can select a section of video as a reference video, each frame image in the reference video is used as a reference image, and then the image to be migrated is processed based on each frame of reference image.
In the embodiment of the present invention, the image to be migrated is represented by a dimensional vector, and the dimensional vector of the image to be migrated is: x (i × m + n + j × m + k) ═ x (j, k, i); wherein i is more than or equal to 3 and more than or equal to 1, n is more than or equal to i and more than or equal to 1, and m is more than or equal to i and more than or equal to 1. The reference image is also expressed by a dimension vector, and the dimension vector of the reference image is y (i × m × n + j × m + k) ═ y (j, k, i); wherein i is more than or equal to 3 and more than or equal to 1, n is more than or equal to i and more than or equal to 1, and m is more than or equal to i and more than or equal to 1.
Specifically, each pixel point in the image to be migrated can represent the position thereof by a dimensional vector or a coordinate, and each pixel point of the reference image can also represent the position thereof by a dimensional vector or a coordinate. For example: an image has 10 rows by 10 columns of pixel points, and the coordinates of the pixel point p in the 5 th row and the 5 th column are expressed as (5, 5); this pixel point p is denoted p (45) by a one-dimensional vector.
In the embodiment of the invention, the first key feature refers to coordinates of a plurality of feature points capable of marking the posture of the target object; for example, the first key feature may be coordinates of various joints of the target object; each joint includes: shoulder joint, elbow joint, wrist joint, carpometacarpal joint, hip joint, knee joint, ankle joint, etc. Furthermore, the first key feature may also be coordinates of a main part of the human body, for example, a part characterizing the head pose, including: eyes, nose tip, temple and chin tip; a portion characterizing a pose of an arm, comprising: shoulder, elbow and carpometacarpal joints; a portion characterizing hand pose, comprising: the knuckles and fingertips of the individual fingers; a site characterizing leg pose comprising: hip joint, knee joint, ankle joint.
In an embodiment of the present invention, the first key feature is a preset key feature in the target object; the second key features correspond to the first key features one to one.
Specifically, when the first key feature is obtained, a second key feature may be obtained according to the first key feature, where the second key feature corresponds to the first key feature, for example, the first key feature includes: and coordinates of the shoulder joint, the elbow joint, the radial wrist joint, the carpometacarpal joint, the hip joint, the knee joint and the ankle joint of the target object in the image to be migrated, and the second key characteristics comprise the coordinates of the shoulder joint, the elbow joint, the radial wrist joint, the carpometacarpal joint, the hip joint, the knee joint and the ankle joint of the reference object in the reference image.
In the embodiment of the present invention, if the target image only includes the face image, that is, only the face pose is migrated, the first key feature is set as the coordinates of each feature point of the human face, for example: eyes, nose, eyebrows, ears, mouth, etc.
In the embodiment of the present invention, the user may select the part to be migrated as needed, for example, when only the face is migrated, only the first key feature of the face is selected, and when only the body is migrated, only the first key feature of the body is selected.
In this embodiment of the present invention, the step 103 includes: determining coordinate values of each first key feature and each second key feature;
and determining the attitude transition matrix according to the coordinate values of the first key features and the coordinate values of the second key features, wherein the attitude transition matrix is used for converting the coordinate values of the first key features into the coordinate values of the second key features corresponding to the first key features.
In the embodiment of the present invention, the gesture migration matrix refers to a gesture migration matrix required for transferring the coordinates of the first key feature to the coordinates of the second key feature. For example, the first key features include: coordinates of the temple (a, b), the shoulder joint (c, d), the temple (m, n) of the second key feature, the shoulder joint (o, p); the coordinates of the temple of the first key feature are stored as (m, n), the coordinates of the shoulder joint of the first key feature are stored as (o, p), the coordinates of the elbow joint of the first key feature are stored as (q, r), and so on. Wherein, according toThe posture transition matrix W is obtained, and when the first key feature includes a plurality of (3 or more), the posture transition matrix W may be obtained in this manner.
In the embodiment of the present invention, if the coordinates of each first key feature is Px and the coordinates of each second key feature is Py, W ═ W [ Px, Py ] is the posture transition matrix required for transforming from the first key feature to the second key feature.
After the attitude migration matrix W is determined, all the pixel points of the image to be migrated can be transferred by adopting the attitude migration matrix W.
In the embodiment of the present invention, the initial image is an initial image that needs to be input to complete a subsequent step of obtaining a target composite image in a preset manner in the embodiment of the present invention.
In an embodiment of the present invention, step 104 includes: and inputting the attitude migration matrix and the image to be migrated into an initial network model to obtain an initial image.
In this embodiment of the present invention, the initial network model may be a model obtained by training according to data samples, where the data samples include: converting a plurality of image samples to be migrated into attitude migration matrix samples of the reference image samples, and converting the plurality of image samples to be migrated and a plurality of target synthetic image samples; training by adopting the data samples to obtain an initial network model; and then inputting the attitude migration matrix and the image to be migrated into an initial network model obtained by training to obtain an initial image, wherein the initial image obtained by adopting the mode is an initial synthetic image of a target object in the image to be migrated in the attitude of a reference object, but details of the initial synthetic image are omitted and all characteristics of the image to be migrated cannot be completely presented. After the subsequent steps are continuously executed, the details of the image to be migrated can be completely supplemented.
In addition, the operating principle of the initial network model may also be Z0W x; wherein Z is0Is the dimensional vector of the initial image, W is the pose migration matrix, and x is the dimensional vector of the image to be migrated. The initial image obtained in this way has some characteristics of the image to be migrated, but is not clear yet, and all pixel points in the image to be migrated are not migrated. The initial image obtained in the mode is used as the basis of subsequent calculation, and the migration quality of the image to be migrated can be improved.
Optionally, step 104, includes: and taking a preset image with a zero dimensional vector as the initial image. The preset image can be stored in a memory, and is called when the image to be migrated is processed.
In the embodiment of the present invention, the value of the dimensional vector corresponding to the initial image may also be assigned to zero, so as to perform subsequent calculation.
And 105, determining a target synthetic image according to the attitude transition matrix, the image to be transitioned and the initial image.
In an embodiment of the present invention, step 105 includes: obtaining an intermediate composite image according to a preset mode, the attitude migration matrix, the image to be migrated and the initial image; and taking the intermediate synthetic image as a new initial image, and circularly executing the posture migration matrix, the image to be migrated and the initial image for preset times according to a preset mode to obtain an intermediate synthetic image.
In the embodiment of the present invention, F (Z, Px, Py) is set to represent a dimensional vector of the target synthetic image in which the target object in the image to be migrated is migrated from the posture of the target object to the posture of the reference object. Then min ∑ iif [ z, P is requiredx,Py]When x | approaches 0, the details of the image to be migrated are all present in the target composite image. Wherein x is a dimensional vector of the image to be migrated. The following pairs of min ∑ iif [ z, Px,Py]-x | solving step as approaching 0;
2) Let A be (W [ P ]x,Py])TW[Px,Py],b=(W[Px,Py])Tx; then pairCarrying out inverse problem modeling to solve an equation set AZ ═ b;
3) if the solution accuracy e is set to 0.0000001, r is0=b-AZ0;p0=r0(ii) a If r is0Is greater than the value of (a) e,rk=rk-1+αk-1Apk-1;pk=rk+βk-1pk-1;wherein, let A be (W)TW;P0=r0;r0=b-AZ0;b=WTx;
4) Obtained by arranging the formula, Zk+1=f(b,A,Zk) Visible, the target composite image Zk+1Is dependent on the pose migration matrix W, the image x to be migrated, and the initial image ZkIn (1).
In the embodiment of the present invention, as can be seen from the foregoing steps 1) to 4), the preset manner in the embodiment of the present invention is specifically as follows:
Zk+1=Zk+αkPk;
wherein Z isk+1A first dimension vector of the intermediate composite image; zkA second-dimension vector in the initial image; wherein W is the attitude migration matrix;rk=rk-1+αk-1Apk-1;pk=rk+βk-1pk-1;wherein, let A be (W)TW;P0=r0;r0=b-AZ0;b=WTx; wherein x is a third-dimensional vector of the image to be migrated, and Z0And obtaining an initial fourth-dimensional vector of the initial image for the initial network model.
In the embodiment of the present invention, the preset manner is to adopt the above formula Zk+1=Zk+αkPkAnd obtaining a target composite image.
In an embodiment of the invention, gesture migrationThe shift matrix W is W [ P ] as described abovex,Py]。
Specifically, for example, the initial image obtained as described above is taken as Z0(ii) a The attitude migration matrix W, the image x to be migrated and the initial image Z are migrated for the first time0Inputting a formula corresponding to the preset mode to obtain:wherein r is0=b-Az0;P0=r0(ii) a Then alpha is0=1/A=1/(W)TW;P0=b-Az0=WTx-(W)TW·Z0(ii) a Then Z1=Z0+(1/(W)TW)·(WTx-(W)TW·Z0)=x/W-Z0. Finally obtain Z1=x/W-Z0Wherein Z is1Firstly, a posture migration matrix W, an image x to be migrated and an initial image Z are migrated0Inputting a formula corresponding to the preset mode to obtain a dimensional vector of the intermediate synthetic image; z0Is the dimensional vector of the acquired initial image.
The intermediate composite image Z obtained above1As a new initial image, the shift matrix W, the image x to be shifted and the new initial image Z are shifted a second time1Inputting a formula corresponding to the preset mode to obtain: z2=Z1+α1P1(ii) a Wherein,r1=r0+α0Ap0;p1=r1+β0p0;wherein, let A be (W)TW;P0=r0;r0=b-AZ0;b=WTx; to obtain Z2。
In the embodiment of the present invention, the dimensional vector of the image may be a one-dimensional vector, a two-dimensional vector, or a three-dimensional vector, which is not limited herein.
In the embodiment of the invention, the preset times are more than or equal to 2; when the preset times is 2 times, the final target synthetic image is Z2. Wherein when the final target synthetic image Z is obtained2When the details are not clear enough, the intermediate synthetic image can be circularly executed as a new initial image, and the intermediate synthetic image can be obtained according to the preset mode, the posture migration matrix, the image to be migrated and the initial image for a preset number of times until the target synthetic image is an image satisfied by the user.
In the embodiment of the invention, the method comprises the following steps: a plurality of frames of reference images; the reference image comprises a temporal sequence; then after step 105, the method further includes: and arranging the plurality of frames of the target synthetic images according to the time sequence to obtain a target synthetic video.
In the embodiment of the invention, the method further comprises the step of inputting the target synthetic image into the completion model to obtain a final synthetic image. The completion model is used for completing the missing part in the target synthetic image. For example, when the image to be migrated is input by the user as an image lacking a human face or lacking a part of a limb, the completion model completes the missing part.
Specifically, the completion model can be obtained by training a large number of images as training samples; for example, a patch model is trained using a back-shot (no face shot), no leg shot, and no arm shot and corresponding full body photograph as training samples.
Wherein, a plurality of frames of reference images with time sequence form a reference video; the user can click and upload the reference video, and the reference video comprises: a plurality of frame reference images, wherein the plurality of frame reference images have corresponding time sequences; the server or the electronic device sequentially executes the steps 101 and 105 on each frame of image in the image to be migrated and the reference video, and finally obtains a multi-frame target synthetic image, and the multi-frame target synthetic image is arranged according to the time sequence, so that the final target synthetic video can be obtained.
In the embodiment of the present invention, the method further includes: identifying each frame of image in a reference video, and selecting an image comprising a human body object as a reference image; taking an image not containing a human body object as a transition image; and finally arranging the multi-frame target synthetic image and the transition image according to the time sequence to obtain the target synthetic video.
The reference video includes a dance motion or other motions, which is not limited herein.
In an embodiment of the present invention, the step 105 includes: extracting a target object in the image to be migrated; determining a synthetic object according to the attitude transition matrix, the target object and the initial image; and synthesizing the background of the reference image and the synthetic object to obtain the target synthetic image.
In the embodiment of the present invention, only the target object in the image to be migrated is migrated, but the background of the target object in the image to be migrated is not migrated, the human body object obtained after the target object in the image to be migrated is a synthetic object, and then the background of the reference image is synthesized with the synthetic object, referring to fig. 3, after the target object corresponding to the image a to be migrated is migrated, the background of the reference image B is used to obtain the target synthetic image C.
In the embodiment of the present invention, referring to fig. 2, the whole image a to be migrated may also be migrated to obtain a target composite image C.
In the embodiment of the invention, an image to be migrated and a reference image are obtained; the image to be migrated comprises: a target object of which the posture is to be converted; the reference image comprises: a reference object presenting a reference pose; acquiring a first key feature of the target object and a second key feature of the reference object; determining an attitude migration matrix according to the first key feature and the second key feature; acquiring an initial image; and determining a target synthetic image according to the attitude transition matrix, the image to be transitioned and the initial image. In the embodiment of the invention, a large number of training sample training models are not required to be acquired to obtain the target synthetic image, the complexity of image migration is reduced, the initial image is acquired, and the attitude migration matrix, the image to be migrated and the initial image are adopted to migrate the whole image to be migrated, so that the details of the image to be migrated can be ensured to be displayed in the target synthetic image, and the omission of the details is prevented.
Fig. 4 is a block diagram of an image processing apparatus according to an embodiment of the present invention, and as shown in the drawing, the apparatus may include:
the first acquisition module is used for acquiring an image to be migrated and a reference image; the image to be migrated comprises: a target object of which the posture is to be converted; the reference image comprises: a reference object presenting a reference pose;
the second acquisition module is used for acquiring the first key feature of the target object and the second key feature of the reference object;
the first determining module is used for determining a posture migration matrix according to the first key feature and the second key feature;
the third acquisition module is used for acquiring an initial image;
and the second determining module is used for determining a target synthetic image according to the attitude transition matrix, the image to be transitioned and the initial image.
The image processing device provided by the embodiment of the invention is provided with the corresponding functional module for executing the image processing method, can execute the image processing method provided by the embodiment of the invention and can achieve the same beneficial effects.
In another embodiment provided by the present invention, there is also provided an electronic device, which may include: the processor, the memory and the computer program stored in the memory and capable of running on the processor, when the processor executes the program, the processes of the image processing method embodiment are realized, the same technical effect can be achieved, and the details are not repeated here to avoid repetition. For example, as shown in fig. 5, the electronic device may specifically include: a processor 301, a storage device 302, a display screen 303 with touch functionality, an input device 304, an output device 305, and a communication device 306. The number of the processors 301 in the electronic device may be one or more, and one processor 301 is taken as an example in fig. 5. The processor 301, the storage means 302, the display 303, the input means 304, the output means 305 and the communication means 306 of the electronic device may be connected by a bus or other means.
In yet another embodiment of the present invention, a computer-readable storage medium is further provided, which has instructions stored therein, and when the instructions are executed on a computer, the instructions cause the computer to execute the image processing method described in any of the above embodiments.
In yet another embodiment, the present invention further provides a computer program product containing instructions which, when run on a computer, cause the computer to perform the image processing method described in any of the above embodiments.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
All the embodiments in the present specification are described in a related manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment.
The above description is only for the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention shall fall within the protection scope of the present invention.
Claims (12)
1. An image processing method, characterized in that the method comprises:
acquiring an image to be migrated and a reference image; the image to be migrated comprises: a target object of which the posture is to be converted; the reference image comprises: a reference object presenting a reference pose;
acquiring a first key feature of the target object and a second key feature of the reference object;
determining an attitude migration matrix according to the first key feature and the second key feature;
acquiring an initial image;
and determining a target synthetic image according to the attitude transition matrix, the image to be transitioned and the initial image.
2. The method of claim 1, wherein said acquiring an initial image comprises:
and inputting the attitude migration matrix and the image to be migrated into an initial network model to obtain an initial image.
3. The method of claim 1, wherein said acquiring an initial image comprises:
and taking a preset image with a zero dimensional vector as the initial image.
4. The method of claim 1, wherein determining a target composite image from the pose migration matrix, the image to be migrated, and the initial image comprises:
obtaining an intermediate composite image according to a preset mode, the attitude migration matrix, the image to be migrated and the initial image;
and taking the intermediate synthetic image as a new initial image, and circularly executing the step of obtaining an intermediate synthetic image according to a preset mode for a preset number of times, the attitude transition matrix, the image to be transitioned and the initial image to obtain the target synthetic image.
5. The method according to claim 4, wherein the predetermined manner is as follows:
Zk+1=Zk+αkPk;
wherein Z isk+1A first dimension vector of the intermediate composite image; zkA second-dimension vector in the initial image; wherein W is the attitude migration matrix;rk=rk-1+αk-1Apk-1;pk=rk+βk-1pk-1;wherein, let A be (W)TW;P0=r0;r0=b-AZ0;b=WTx; wherein x is a third-dimensional vector of the image to be migrated, and Z0And obtaining an initial fourth-dimensional vector of the initial image for the initial network model.
6. The method of claim 1, wherein the first key feature is a preset key feature in the target object; the second key features correspond to the first key features one to one.
7. The method of claim 6, wherein determining a pose migration matrix from the first and second key features comprises:
determining coordinate values of each first key feature and each second key feature;
and determining the attitude transition matrix according to the coordinate values of the first key features and the coordinate values of the second key features, wherein the attitude transition matrix is used for converting the coordinate values of the first key features into the coordinate values of the second key features corresponding to the first key features.
8. The method of claim 7, comprising: a plurality of frames of reference images; the reference image comprises a temporal sequence;
after determining the target composite image according to the attitude transition matrix, the image to be transitioned, and the initial image, the method further includes:
and arranging the plurality of frames of the target synthetic images according to the time sequence to obtain a target synthetic video.
9. The method of claim 1, wherein determining a target composite image from the pose migration matrix, the image to be migrated, and the initial image comprises:
extracting a target object in the image to be migrated;
determining a synthetic object according to the attitude transition matrix, the target object and the initial image;
and synthesizing the background of the reference image and the synthetic object to obtain the target synthetic image.
10. An image processing apparatus, characterized in that the apparatus comprises:
the first acquisition module is used for acquiring an image to be migrated and a reference image; the image to be migrated comprises: a target object of which the posture is to be converted; the reference image comprises: a reference object presenting a reference pose;
the second acquisition module is used for acquiring the first key feature of the target object and the second key feature of the reference object;
the first determining module is used for determining a posture migration matrix according to the first key feature and the second key feature;
the third acquisition module is used for acquiring an initial image;
and the second determining module is used for determining a target synthetic image according to the attitude transition matrix, the image to be transitioned and the initial image.
11. An electronic device, comprising a processor, a memory, and a program or instructions stored on the memory and executable on the processor, which when executed by the processor, implement the steps of the method of any one of claims 1-9.
12. A readable storage medium, characterized in that it stores thereon a program or instructions which, when executed by a processor, implement the steps of the method according to any one of claims 1-9.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110009523.2A CN112734632B (en) | 2021-01-05 | 2021-01-05 | Image processing method, device, electronic equipment and readable storage medium |
PCT/CN2022/070336 WO2022148379A1 (en) | 2021-01-05 | 2022-01-05 | Image processing method and apparatus, electronic device, and readable storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110009523.2A CN112734632B (en) | 2021-01-05 | 2021-01-05 | Image processing method, device, electronic equipment and readable storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112734632A true CN112734632A (en) | 2021-04-30 |
CN112734632B CN112734632B (en) | 2024-02-27 |
Family
ID=75591231
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110009523.2A Active CN112734632B (en) | 2021-01-05 | 2021-01-05 | Image processing method, device, electronic equipment and readable storage medium |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN112734632B (en) |
WO (1) | WO2022148379A1 (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2022148379A1 (en) * | 2021-01-05 | 2022-07-14 | 百果园技术(新加坡)有限公司 | Image processing method and apparatus, electronic device, and readable storage medium |
CN115713582A (en) * | 2022-12-02 | 2023-02-24 | 北京百度网讯科技有限公司 | Virtual image generation method, device, electronic equipment and medium |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115423752B (en) * | 2022-08-03 | 2023-07-07 | 荣耀终端有限公司 | Image processing method, electronic equipment and readable storage medium |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109670474A (en) * | 2018-12-28 | 2019-04-23 | 广东工业大学 | A kind of estimation method of human posture based on video, device and equipment |
CN110705625A (en) * | 2019-09-26 | 2020-01-17 | 北京奇艺世纪科技有限公司 | Image processing method and device, electronic equipment and storage medium |
CN111626218A (en) * | 2020-05-28 | 2020-09-04 | 腾讯科技(深圳)有限公司 | Image generation method, device and equipment based on artificial intelligence and storage medium |
US20200380720A1 (en) * | 2019-06-03 | 2020-12-03 | Microsoft Technology Licensing, Llc | Novel pose synthesis |
CN112115783A (en) * | 2020-08-12 | 2020-12-22 | 中国科学院大学 | Human face characteristic point detection method, device and equipment based on deep knowledge migration |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109960986A (en) * | 2017-12-25 | 2019-07-02 | 北京市商汤科技开发有限公司 | Human face posture analysis method, device, equipment, storage medium and program |
CN111027438B (en) * | 2019-12-03 | 2023-06-02 | Oppo广东移动通信有限公司 | Human body posture migration method, mobile terminal and computer storage medium |
CN111652798B (en) * | 2020-05-26 | 2023-09-29 | 浙江大华技术股份有限公司 | Face pose migration method and computer storage medium |
CN112734632B (en) * | 2021-01-05 | 2024-02-27 | 百果园技术(新加坡)有限公司 | Image processing method, device, electronic equipment and readable storage medium |
-
2021
- 2021-01-05 CN CN202110009523.2A patent/CN112734632B/en active Active
-
2022
- 2022-01-05 WO PCT/CN2022/070336 patent/WO2022148379A1/en active Application Filing
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109670474A (en) * | 2018-12-28 | 2019-04-23 | 广东工业大学 | A kind of estimation method of human posture based on video, device and equipment |
US20200380720A1 (en) * | 2019-06-03 | 2020-12-03 | Microsoft Technology Licensing, Llc | Novel pose synthesis |
CN110705625A (en) * | 2019-09-26 | 2020-01-17 | 北京奇艺世纪科技有限公司 | Image processing method and device, electronic equipment and storage medium |
CN111626218A (en) * | 2020-05-28 | 2020-09-04 | 腾讯科技(深圳)有限公司 | Image generation method, device and equipment based on artificial intelligence and storage medium |
CN112115783A (en) * | 2020-08-12 | 2020-12-22 | 中国科学院大学 | Human face characteristic point detection method, device and equipment based on deep knowledge migration |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2022148379A1 (en) * | 2021-01-05 | 2022-07-14 | 百果园技术(新加坡)有限公司 | Image processing method and apparatus, electronic device, and readable storage medium |
CN115713582A (en) * | 2022-12-02 | 2023-02-24 | 北京百度网讯科技有限公司 | Virtual image generation method, device, electronic equipment and medium |
CN115713582B (en) * | 2022-12-02 | 2023-10-27 | 北京百度网讯科技有限公司 | Avatar generation method, device, electronic equipment and medium |
Also Published As
Publication number | Publication date |
---|---|
WO2022148379A1 (en) | 2022-07-14 |
CN112734632B (en) | 2024-02-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11468612B2 (en) | Controlling display of a model based on captured images and determined information | |
CN112734632B (en) | Image processing method, device, electronic equipment and readable storage medium | |
CN109960962B (en) | Image recognition method and device, electronic equipment and readable storage medium | |
CN111861872B (en) | Image face changing method, video face changing method, device, equipment and storage medium | |
CN109815776B (en) | Action prompting method and device, storage medium and electronic device | |
JP2019096113A (en) | Processing device, method and program relating to keypoint data | |
CN113706699B (en) | Data processing method and device, electronic equipment and computer readable storage medium | |
US10970849B2 (en) | Pose estimation and body tracking using an artificial neural network | |
JP2017037424A (en) | Learning device, recognition device, learning program and recognition program | |
JP2022501732A (en) | Image processing methods and devices, image devices and storage media | |
CN114022645A (en) | Action driving method, device, equipment and storage medium of virtual teacher system | |
WO2020147791A1 (en) | Image processing method and device, image apparatus, and storage medium | |
WO2022197024A1 (en) | Point-based modeling of human clothing | |
JP2010211732A (en) | Object recognition device and method | |
CN112102451B (en) | Wearable virtual live broadcast method and equipment based on common camera | |
CN111416938B (en) | Augmented reality close-shooting method and device and computer readable storage medium | |
CN111783497A (en) | Method, device and computer-readable storage medium for determining characteristics of target in video | |
JP7067709B1 (en) | Programs, devices and methods for statistically analyzing skeleton-based body length from skin models | |
US11508121B2 (en) | Method for annotating points on a hand image to create training dataset for machine learning | |
JP2022092528A (en) | Three-dimensional person attitude estimation apparatus, method, and program | |
CN114638744B (en) | Human body posture migration method and device | |
US12033281B2 (en) | Automatic blending of human facial expression and full-body poses for dynamic digital human model creation using integrated photo-video volumetric capture system and mesh-tracking | |
CN114973396B (en) | Image processing method, image processing device, terminal equipment and computer readable storage medium | |
WO2024149294A1 (en) | Virtual fitting method and apparatus, fitting model training method and apparatus, and device | |
CN118230359A (en) | Human body posture estimation method, device and system and readable storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |