CN110706169A - Star portrait optimization method and device and storage device - Google Patents
Star portrait optimization method and device and storage device Download PDFInfo
- Publication number
- CN110706169A CN110706169A CN201910914783.7A CN201910914783A CN110706169A CN 110706169 A CN110706169 A CN 110706169A CN 201910914783 A CN201910914783 A CN 201910914783A CN 110706169 A CN110706169 A CN 110706169A
- Authority
- CN
- China
- Prior art keywords
- star
- portrait
- optimized
- star portrait
- optimization
- 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.)
- Pending
Links
- 238000005457 optimization Methods 0.000 title claims abstract description 103
- 238000000034 method Methods 0.000 title claims abstract description 60
- 238000003860 storage Methods 0.000 title claims abstract description 27
- 238000013135 deep learning Methods 0.000 claims abstract description 22
- 238000004590 computer program Methods 0.000 claims description 28
- 238000010276 construction Methods 0.000 claims description 8
- 238000012549 training Methods 0.000 claims description 6
- 238000001514 detection method Methods 0.000 claims description 5
- 230000008569 process Effects 0.000 description 24
- 230000006870 function Effects 0.000 description 10
- 238000010586 diagram Methods 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 5
- 230000001413 cellular effect Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/77—Retouching; Inpainting; Scratch removal
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30196—Human being; Person
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention is suitable for the technical field of video optimization, and particularly relates to a star portrait optimization method, a star portrait optimization device and a storage device, wherein the portrait optimization method comprises the following steps: deep learning is carried out on the star pictures in advance, and a trained star portrait optimization model is generated; acquiring star portrait information to be optimized in an input video; and calculating the star portrait information to be optimized by adopting a trained star portrait optimization model to obtain the optimized star portrait information, and outputting a video according to the optimized star portrait information. The star portrait optimization method can greatly improve the definition of the portrait in the picture and correct and perfect the part with poor quality of the portrait, thereby greatly improving the video watching experience of users.
Description
Technical Field
The invention is suitable for the technical field of video optimization, and particularly relates to a star portrait optimization method, a star portrait optimization device and a star portrait optimization storage device.
Background
With the abundance of amateur activities and the gradual development of various terminals, many people choose to watch various videos on the terminals, such as watching TV plays on mobile phones. The existing television series is often only the original video provided by the producer, and the user watches the original video through a third-party watching path (such as APP playing and the like). A third party often performs Video coding based on a new generation high performance Video coding standard hevc (high Efficiency Video coding), and although the coding method can greatly reduce the bit rate, for the increasingly long demand of people, if the screen of the terminal is small, the defects of poor definition of image quality and poor definition of portrait still exist, and the viewing experience of a user is greatly reduced.
Disclosure of Invention
The embodiment of the invention provides a star portrait optimization method, a star portrait optimization device and a star portrait storage device, and aims to solve the problems that the definition of image quality of videos provided for users to watch is poor and the portrait is not fine enough.
The embodiment of the invention is realized in such a way, and provides a star portrait optimization method, which comprises the following steps: deep learning is carried out on the star pictures in advance, and a trained star portrait optimization model is generated; acquiring star portrait information to be optimized in an input video; and calculating the star portrait information to be optimized by adopting a trained star portrait optimization model to obtain the optimized star portrait information, and outputting a video according to the optimized star portrait information.
Further, the step of performing deep learning on the star picture in advance to generate a trained star portrait optimization model specifically includes: constructing a star picture database according to a plurality of star pictures; constructing a star portrait recognition model and a star portrait compensation model for deep learning; and training and learning the star pictures in the star picture database by adopting the star portrait recognition model and the star portrait compensation model to generate a trained star portrait optimization model.
Further, the step of collecting the star portrait information to be optimized in the input video specifically includes: detecting in an input video according to preset time to obtain a sequence image; and acquiring star portrait information to be optimized in the sequence images, wherein the star portrait information corresponds to the labels of the sequence images.
Further, the star picture is a corresponding picture of each lead actor in the video.
The embodiment of the present invention further provides a star portrait optimization device, where the star portrait optimization device includes: the generation unit is used for carrying out deep learning on the star picture in advance to generate a trained star portrait optimization model; the acquisition unit is used for acquiring star portrait information to be optimized in an input video; and the computing unit is used for computing the star portrait information to be optimized by adopting a trained star portrait optimization model, acquiring the optimized star portrait information and outputting a video according to the optimized star portrait information.
Still further, the generating unit includes: the database construction unit is used for constructing a star picture database according to the plurality of star pictures; the model construction unit is used for constructing a star portrait recognition model and a star portrait compensation model for deep learning; and the learning unit is used for training and learning the star pictures in the star picture database by adopting the star portrait recognition model and the star portrait compensation model to generate a trained star portrait optimization model.
Still further, the acquisition unit includes: the detection unit is used for detecting in the input video according to preset time to obtain a sequence image; and the star portrait information acquisition unit is used for acquiring star portrait information to be optimized in the sequence image, and the star portrait information corresponds to the label of the sequence image.
Further, the star picture is a corresponding picture of each lead actor in the video.
The embodiment of the invention also provides a star portrait optimization device, which comprises a memory, a processor and a computer program which is stored in the memory and can run on the processor, wherein the processor implements the steps of the star portrait optimization method when executing the computer program.
Embodiments of the present invention further provide a storage device, where the storage device stores a computer program, where the computer program can be executed to implement the steps of the star portrait optimization method described above.
Compared with the prior art, the star portrait optimization method, the star portrait optimization device and the storage device have the advantages that the trained star portrait optimization model is generated in advance, the video to be optimized is obtained, the star portrait information to be optimized is collected from the video after the video is input, the optimized star portrait information to be optimized is input into the star portrait optimization model, the optimized star portrait information can be automatically generated, the optimized star portrait information is input into the video, the portrait optimization of the whole video is completed, the definition of the portrait in the picture can be greatly improved, and the part with poor portrait quality can be corrected and improved, so that the video watching experience of a user is greatly improved.
Drawings
Fig. 1 is a block flow diagram of a star portrait optimization method provided in an embodiment of the present invention;
FIG. 2 is a block diagram of a process for generating a trained star portrait optimization model according to an embodiment of the present invention;
FIG. 3 is a block diagram of a process for collecting star portrait information to be optimized according to an embodiment of the present invention;
fig. 4 is a block diagram of a star portrait optimization apparatus according to an embodiment of the present invention;
FIG. 5 is a block diagram of a generating unit according to an embodiment of the present invention;
fig. 6 is a block diagram of an acquisition unit according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
By designing a star portrait optimization method, a star portrait optimization device and a storage device, a trained star portrait optimization model is generated in advance, a video to be optimized is obtained, star portrait information to be optimized is collected from the video after the video is input, the optimized star portrait information to be optimized is input into the star portrait optimization model, the optimized star portrait information can be automatically generated, and finally the optimized star portrait information is input into the video to complete portrait optimization of the whole video, so that the definition of a portrait in a picture can be greatly improved, and the part with poor portrait quality can be corrected and perfected, thereby greatly improving the video watching experience of a user.
Example one
Referring to fig. 1, the present embodiment provides a star portrait optimization method, where the star portrait optimization method includes the following steps:
step 1, deep learning is carried out on a star picture in advance to generate a trained star portrait optimization model;
step 2, acquiring star portrait information to be optimized in an input video; and
and 3, calculating the star portrait information to be optimized by adopting a trained star portrait optimization model, acquiring the optimized star portrait information, and outputting a video according to the optimized star portrait information.
The method comprises the steps of obtaining a plurality of star pictures in advance, wherein the star pictures can be obtained from a full-network high-definition picture database, such as a Baidu database, a Google database and the like. And then, carrying out deep learning on the acquired star pictures, and continuously carrying out picture input and portrait output processes based on the deep learning process to generate a trained star portrait optimization model. When star portrait optimization is carried out, a video to be optimized is input, a plurality of star portrait information to be optimized are collected in the input video, and the star portrait information can be star portrait pictures. And then, inputting the acquired star portrait information to be optimized into a trained star portrait optimization model, and automatically acquiring the optimized star portrait information through calculation. After the optimized star portrait information is obtained, video coding is carried out according to the optimized star portrait information, and finally the video optimized for the star portrait can be output, so that the definition of the portrait in the picture is greatly improved, the part with poor portrait quality is corrected and perfected, and the video watching experience of a user is greatly improved. It is worth mentioning that the portrait includes a face or a torso. In practical applications, the portrait typically includes only faces, since it is considered that the user's focus is typically only on the faces of the stars.
It should be noted that, the star picture obtaining process and the star portrait optimization model generating process may be performed on an input video, for example, if the input video is an a tv series, and the star picture obtaining process and the star portrait optimization model generating process are performed on each star of the a tv series. Or, the star picture obtaining process and the star portrait optimization model generating process may be performed for all videos, that is, the star picture obtaining process and the star portrait optimization model generating process are performed for part or all of the existing stars, and then star portrait optimization of part or all of the videos may be achieved based on the star portrait optimization model.
The star portrait optimization method can be executed by a video producer or a third party who plays the video, and of course, can be executed by other parties as long as the star portrait optimization method can optimize the portrait in the video.
Example two
Referring to fig. 2, on the basis of the first embodiment, the step of performing deep learning on a star picture in advance to generate a trained star portrait optimization model in the second embodiment specifically includes:
step 11, constructing a star picture database according to a plurality of star pictures;
step 12, constructing a star portrait recognition model and a star portrait compensation model for deep learning;
and step 13, training and learning the star pictures in the star picture database by adopting the star portrait recognition model and the star portrait compensation model to generate a trained star portrait optimization model.
The method comprises the steps that a plurality of required star pictures are obtained from a full-network high-definition picture database in advance, and a star picture database is built according to the star pictures; alternatively, if the existing star picture database contains a plurality of desired star pictures, the existing star picture database may be used directly. And then, a star portrait recognition model and a star portrait compensation model for deep learning are established, the star portrait recognition model can recognize corresponding stars aiming at the star picture to form a recognition result, the star portrait compensation model can compensate and optimize the star picture according to the recognition result, and parts with poor definition and quality in the picture are corrected and finely rendered aiming at the star picture. And then, training and learning the star pictures in the star picture database by adopting the constructed star portrait recognition model and the star portrait compensation model, continuously inputting the star pictures to be optimized and continuously outputting the optimized star pictures, and generating a trained star portrait optimization model. And subsequently, through the trained star portrait optimization model, the optimized star picture can be automatically output only by inputting the star picture.
EXAMPLE III
Referring to fig. 3, on the basis of the first embodiment, the step of collecting the star portrait information to be optimized in the input video of the third embodiment specifically includes:
step 21, detecting in the input video according to preset time to obtain a sequence image;
and step 22, acquiring star portrait information to be optimized in the sequence images, wherein the star portrait information corresponds to the labels of the sequence images.
In the method, a preset time is set, and the preset time can be customized, for example, one frame, one second, and the like, where a frame refers to a picture, and generally in a video (e.g., a movie), 24 frames per second are set. Then, the input video is obtained, the input video is detected according to the preset time, all sequence images in the preset time are further obtained, and each sequence image can be labeled according to the sequence. Subsequently, star portrait information to be optimized is collected in all the acquired sequence images, and the star portrait information corresponds to the labels of the sequence images. For example, for a first frame, a sequence image in the first frame is obtained, the sequence image is numbered 1, and then star portrait information to be optimized is collected in the sequence image, the star portrait information is also 1. For the second frame, a sequence image in the second frame is obtained, the sequence image is numbered 2, and then star portrait information to be optimized is collected in the sequence image, and the star portrait information is also 2. And repeating the steps until the star portrait information of all the sequence images in the video is acquired. After the corresponding marks are carried out, the subsequent encoding and output of the video are facilitated, and the corresponding star portrait information of the corresponding frame number only needs to be replaced mutually.
Example four
On the basis of the first embodiment, the star picture of the fourth embodiment is a corresponding picture of each lead actor in the video. For each video, a lead actor list is created prior to production. Because more main plays exist in the main actor list, the user can put the focus on each main actor, and therefore, when the star portrait optimization method is executed subsequently, only each main actor in the main actor list needs to be aimed at, so that the optimization efficiency can be greatly improved, and the calculation amount is reduced.
EXAMPLE five
Referring to fig. 4, in a fifth embodiment, a star portrait optimization apparatus is provided, where the star portrait optimization apparatus includes:
the generation unit 100 is used for carrying out deep learning on the star picture in advance to generate a trained star portrait optimization model;
the acquisition unit 200 is used for acquiring star portrait information to be optimized in an input video; and
and the calculating unit 300 is configured to calculate the star portrait information to be optimized by using the trained star portrait optimization model, obtain optimized star portrait information, and output a video according to the optimized star portrait information.
In which, through the generating unit 100, a plurality of star pictures are obtained in advance, and the star pictures can be obtained from a full-network high-definition picture database, such as a Baidu database, a Google database, and the like. And then, carrying out deep learning on the acquired star pictures, and continuously carrying out picture input and portrait output processes based on the deep learning process to generate a trained star portrait optimization model. Through the acquisition unit 200, when star portrait optimization is performed, a video to be optimized is input, and a plurality of star portrait information to be optimized is acquired in the input video, wherein the star portrait information can be star portrait pictures. Then, the acquired star portrait information to be optimized is input into the trained star portrait optimization model through the calculation unit 300, and the optimized star portrait information is automatically acquired through calculation. After the optimized star portrait information is obtained, video coding is carried out according to the optimized star portrait information, and finally the video optimized for the star portrait can be output, so that the definition of the portrait in the picture is greatly improved, the part with poor portrait quality is corrected and perfected, and the video watching experience of a user is greatly improved. It is worth mentioning that the portrait includes a face or a torso. In practical applications, the portrait typically includes only faces, since it is considered that the user's focus is typically only on the faces of the stars.
It should be noted that, the star picture obtaining process and the star portrait optimization model generating process may be performed on an input video, for example, if the input video is an a tv series, and the star picture obtaining process and the star portrait optimization model generating process are performed on each star of the a tv series. Or, the star picture obtaining process and the star portrait optimization model generating process may be performed for all videos, that is, the star picture obtaining process and the star portrait optimization model generating process are performed for part or all of the existing stars, and then star portrait optimization of part or all of the videos may be achieved based on the star portrait optimization model.
The star portrait optimization device can be executed by a video producer or a third party for playing the video, and of course, can be executed by other parties as long as the star portrait optimization device can optimize the portrait in the video.
EXAMPLE six
Referring to fig. 5, on the basis of the fifth embodiment, the generating unit of the sixth embodiment includes:
a database construction unit 110, configured to construct a star picture database according to a plurality of star pictures;
a model construction unit 120, configured to construct a star portrait recognition model and a star portrait compensation model for deep learning;
a learning unit 130, configured to train and learn the star images in the star image database by using the star image recognition model and the star image compensation model, and generate a trained star image optimization model.
A database construction unit 110 is used for acquiring a plurality of required star pictures from a full-network high-definition picture database in advance, and constructing a star picture database according to the star pictures; alternatively, if the existing star picture database contains a plurality of desired star pictures, the existing star picture database may be used directly. Then, a star portrait recognition model and a star portrait compensation model for deep learning are constructed by the model construction unit 120, the star portrait recognition model can recognize a corresponding star for a star picture to form a recognition result, the star portrait compensation model can compensate and optimize the star picture according to the recognition result, and parts with poor definition and quality in the picture are corrected and finely rendered for the star picture. Subsequently, through the learning unit 130, the constructed star portrait recognition model and the star portrait compensation model are used to train and learn the star pictures in the star picture database, so as to continuously input the star pictures to be optimized and continuously output the optimized star pictures, thereby generating a trained star portrait optimization model. And subsequently, through the trained star portrait optimization model, the optimized star picture can be automatically output only by inputting the star picture.
EXAMPLE seven
Referring to fig. 6, on the basis of the fifth embodiment, the acquisition unit of the seventh embodiment includes:
a detection unit 210, configured to perform detection in an input video according to a preset time to obtain a sequence image;
a star portrait information collecting unit 220, configured to collect star portrait information to be optimized in the sequence image, where the star portrait information corresponds to a label of the sequence image.
The detection unit 210 sets a preset time, which can be customized, for example, one frame, one second, etc., where a frame refers to a picture, and usually in a video (e.g., a movie), 24 frames per second. Then, the input video is obtained, the input video is detected according to the preset time, all sequence images in the preset time are further obtained, and each sequence image can be labeled according to the sequence. Subsequently, through the star portrait information acquisition unit 220, star portrait information to be optimized is acquired in all the acquired sequence images, and the star portrait information corresponds to the labels of the sequence images. For example, for a first frame, a sequence image in the first frame is obtained, the sequence image is numbered 1, and then star portrait information to be optimized is collected in the sequence image, the star portrait information is also 1. For the second frame, a sequence image in the second frame is obtained, the sequence image is numbered 2, and then star portrait information to be optimized is collected in the sequence image, and the star portrait information is also 2. And repeating the steps until the star portrait information of all the sequence images in the video is acquired. After the corresponding marks are carried out, the subsequent encoding and output of the video are facilitated, and the corresponding star portrait information of the corresponding frame number only needs to be replaced mutually.
Example eight
On the basis of the fifth embodiment, the star picture of the eighth embodiment is a corresponding picture of each lead actor in the video. For each video, a lead actor list is created prior to production. Because the plays of each lead actor in the lead actor list are more, the user can put the attention point on each lead actor, and therefore, when the star portrait optimization device is used subsequently, only each lead actor in the lead actor list needs to be aimed at, so that the optimization efficiency can be greatly improved, and the calculation amount is reduced.
Example nine
The ninth embodiment provides a star portrait optimization apparatus, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of the star portrait optimization method according to the first to fourth embodiments when executing the computer program.
Illustratively, a computer program can be partitioned into one or more modules, which are stored in memory and executed by a processor to implement the present invention. One or more of the modules may be a series of computer program instruction segments capable of performing certain functions, the instruction segments being used to describe the execution process of the computer program in the computer device, namely the star figure optimizing device. For example, the computer program may be divided into the steps of the star portrait optimization method provided by the various method embodiments described above.
Those skilled in the art will appreciate that the above description of a computer apparatus is by way of example only and is not intended to be limiting of computer apparatus, and that the apparatus may include more or less components than those described, or some of the components may be combined, or different components may be included, such as input output devices, network access devices, buses, etc.
The Processor may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may be any conventional processor or the like, which is the control center of the computer apparatus and connects the various parts of the overall user terminal using various interfaces and lines.
The memory may be used to store the computer programs and/or modules, and the processor may implement various functions of the computer device by running or executing the computer programs and/or modules stored in the memory and invoking data stored in the memory. The memory may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required by at least one function (such as a sound playing function, an image playing function, etc.), and the like; the storage data area may store data (such as audio data, a phonebook, etc.) created according to the use of the cellular phone, and the like. In addition, the memory may include high speed random access memory, and may also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), at least one magnetic disk storage device, a Flash memory device, or other volatile solid state storage device.
The modules/units integrated by the computer device may be stored in a computer-readable storage medium if they are implemented in the form of software functional units and sold or used as separate products. Based on such understanding, all or part of the flow of the method according to the embodiments of the present invention may also be implemented by a computer program, which may be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of the method embodiments may be implemented. Wherein the computer program comprises computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, usb disk, removable hard disk, magnetic disk, optical disk, computer Memory, Read-Only Memory (ROM), Random Access Memory (RAM), electrical carrier wave signals, telecommunications signals, software distribution medium, and the like.
Example ten
The present embodiment provides a storage device, which stores a computer program, where the computer program can be executed to implement the steps of the star portrait optimization method according to the first to the fourth embodiments.
Illustratively, a computer program can be partitioned into one or more modules, which are stored in memory and executed by a processor to implement the present invention. One or more of the modules may be a series of computer program instruction segments capable of performing certain functions, the instruction segments being used to describe the execution process of the computer program in the computer device, namely the star figure optimizing device. For example, the computer program may be divided into the steps of the star portrait optimization method provided by the various method embodiments described above.
The storage device may be used to store the computer programs and/or modules, and the processor may implement various functions of the computer device by running or executing the computer programs and/or modules stored in the storage device and invoking data stored in the storage device. The storage device may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program (such as a sound playing function, an image playing function, etc.) required by at least one function, and the like; the storage data area may store data (such as audio data, a phonebook, etc.) created according to the use of the cellular phone, and the like. In addition, the storage device may include a high speed random access memory device, and may also include a non-volatile storage device, such as a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), at least one magnetic disk storage device, a Flash memory device, or other volatile solid state storage device.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.
Claims (10)
1. A star portrait optimization method is characterized by comprising the following steps:
deep learning is carried out on the star pictures in advance, and a trained star portrait optimization model is generated;
acquiring star portrait information to be optimized in an input video; and
calculating the star portrait information to be optimized by adopting a trained star portrait optimization model, acquiring the optimized star portrait information, and outputting a video according to the optimized star portrait information.
2. The star portrait optimization method according to claim 1, wherein the step of performing deep learning on the star picture in advance to generate the trained star portrait optimization model specifically includes:
constructing a star picture database according to a plurality of star pictures;
constructing a star portrait recognition model and a star portrait compensation model for deep learning;
and training and learning the star pictures in the star picture database by adopting the star portrait recognition model and the star portrait compensation model to generate a trained star portrait optimization model.
3. A star portrait optimization method according to claim 1, wherein the step of collecting star portrait information to be optimized in the input video specifically comprises:
detecting in an input video according to preset time to obtain a sequence image;
and acquiring star portrait information to be optimized in the sequence images, wherein the star portrait information corresponds to the labels of the sequence images.
4. A star portrait optimization method as claimed in claim 1, wherein the star picture is a corresponding picture of each lead actor in the video.
5. A star figure optimizing device, comprising:
the generation unit is used for carrying out deep learning on the star picture in advance to generate a trained star portrait optimization model;
the acquisition unit is used for acquiring star portrait information to be optimized in an input video; and
and the computing unit is used for computing the star portrait information to be optimized by adopting a trained star portrait optimization model, acquiring the optimized star portrait information and outputting a video according to the optimized star portrait information.
6. A star portrait optimization apparatus as claimed in claim 5, wherein the generation unit comprises:
the database construction unit is used for constructing a star picture database according to the plurality of star pictures;
the model construction unit is used for constructing a star portrait recognition model and a star portrait compensation model for deep learning;
and the learning unit is used for training and learning the star pictures in the star picture database by adopting the star portrait recognition model and the star portrait compensation model to generate a trained star portrait optimization model.
7. A star portrait optimization device as claimed in claim 5, wherein the acquisition unit comprises:
the detection unit is used for detecting in the input video according to preset time to obtain a sequence image;
and the star portrait information acquisition unit is used for acquiring star portrait information to be optimized in the sequence image, and the star portrait information corresponds to the label of the sequence image.
8. A star portrait optimization device as claimed in claim 5, wherein the star picture is a corresponding picture of each lead actor in the video.
9. A celebrity image optimization apparatus comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the steps of the method of optimizing a celebrity image according to any one of claims 1 to 4 when executing the computer program.
10. A storage device, characterized in that it stores a computer program executable to implement the steps of the star portrait optimization method according to any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910914783.7A CN110706169A (en) | 2019-09-26 | 2019-09-26 | Star portrait optimization method and device and storage device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910914783.7A CN110706169A (en) | 2019-09-26 | 2019-09-26 | Star portrait optimization method and device and storage device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110706169A true CN110706169A (en) | 2020-01-17 |
Family
ID=69198073
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910914783.7A Pending CN110706169A (en) | 2019-09-26 | 2019-09-26 | Star portrait optimization method and device and storage device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110706169A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112351336A (en) * | 2020-10-29 | 2021-02-09 | 南京创维信息技术研究院有限公司 | Method, device, terminal and medium for optimizing television image quality based on video image segmentation |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108615219A (en) * | 2018-04-28 | 2018-10-02 | 东莞市华睿电子科技有限公司 | A kind of image landscaping treatment method of intelligent terminal |
| CN109271911A (en) * | 2018-08-24 | 2019-01-25 | 太平洋未来科技（深圳）有限公司 | Three-dimensional face optimization method, device and electronic equipment based on light |
| CN109302628A (en) * | 2018-10-24 | 2019-02-01 | 广州虎牙科技有限公司 | A kind of face processing method based on live streaming, device, equipment and storage medium |
| WO2019127273A1 (en) * | 2017-12-28 | 2019-07-04 | 深圳市锐明技术股份有限公司 | Multi-person face detection method, apparatus, server, system, and storage medium |
-
2019
- 2019-09-26 CN CN201910914783.7A patent/CN110706169A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019127273A1 (en) * | 2017-12-28 | 2019-07-04 | 深圳市锐明技术股份有限公司 | Multi-person face detection method, apparatus, server, system, and storage medium |
| CN108615219A (en) * | 2018-04-28 | 2018-10-02 | 东莞市华睿电子科技有限公司 | A kind of image landscaping treatment method of intelligent terminal |
| CN109271911A (en) * | 2018-08-24 | 2019-01-25 | 太平洋未来科技（深圳）有限公司 | Three-dimensional face optimization method, device and electronic equipment based on light |
| CN109302628A (en) * | 2018-10-24 | 2019-02-01 | 广州虎牙科技有限公司 | A kind of face processing method based on live streaming, device, equipment and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 赵小川: "《MATLAB图像处理 程序实现与模块化仿真》", 北京航空航天大学出版社, pages: 43 - 47 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112351336A (en) * | 2020-10-29 | 2021-02-09 | 南京创维信息技术研究院有限公司 | Method, device, terminal and medium for optimizing television image quality based on video image segmentation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10796685B2 (en) | Method and device for image recognition | |
| CN110557678B (en) | Video processing method, device and equipment | |
| CN109784304B (en) | Method and apparatus for labeling dental images | |
| TW202027033A (en) | Image processing method and apparatus, electronic device and storage medium | |
| CN109168026A (en) | Instant video display methods, device, terminal device and storage medium | |
| CN114331820A (en) | Image processing method, image processing device, electronic equipment and storage medium | |
| CN109815448B (en) | Slide generation method and device | |
| CN112927144A (en) | Image enhancement method, image enhancement device, medium, and electronic apparatus | |
| WO2020118503A1 (en) | Method and apparatus for determining image focusing region | |
| CN114330236A (en) | Character generation method and device, electronic equipment and storage medium | |
| CN110717421A (en) | Video content understanding method and device based on generation countermeasure network | |
| CN117111879A (en) | Dial generation method and device, intelligent wearable device and storage medium | |
| CN110647508B (en) | Data compression method, data decompression method, device and electronic equipment | |
| CN110097004B (en) | Facial expression recognition method and device | |
| CN112785669B (en) | Virtual image synthesis method, device, equipment and storage medium | |
| CN110706169A (en) | Star portrait optimization method and device and storage device | |
| CN114004905A (en) | Method, device and equipment for generating character style image and storage medium | |
| US11051080B2 (en) | Method for improving video resolution and video quality, encoder, and decoder | |
| CN111626922B (en) | Picture generation method and device, electronic equipment and computer readable storage medium | |
| CN111049735B (en) | Group head portrait display method, device, equipment and storage medium | |
| CN109327736B (en) | Program recommendation method based on program viewing expression and user information | |
| CN111784726A (en) | Image matting method and device | |
| CN116012418A (en) | Multi-target tracking method and device | |
| US11637953B2 (en) | Method, apparatus, electronic device, storage medium and system for vision task execution | |
| CN111383289A (en) | Image processing method, image processing device, terminal equipment and computer 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 |