CN110517187A - Advertisement generation method, apparatus and system - Google Patents
Advertisement generation method, apparatus and system Download PDFInfo
- Publication number
- CN110517187A CN110517187A CN201910820682.3A CN201910820682A CN110517187A CN 110517187 A CN110517187 A CN 110517187A CN 201910820682 A CN201910820682 A CN 201910820682A CN 110517187 A CN110517187 A CN 110517187A
- Authority
- CN
- China
- Prior art keywords
- advertisement
- replaced
- frame
- image
- default
- 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
- 238000000034 method Methods 0.000 title claims abstract description 58
- 230000001815 facial effect Effects 0.000 claims abstract description 86
- 238000012545 processing Methods 0.000 claims description 32
- 230000008569 process Effects 0.000 claims description 19
- 238000004891 communication Methods 0.000 claims description 11
- 238000001914 filtration Methods 0.000 claims description 9
- 230000011218 segmentation Effects 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 3
- 206010034719 Personality change Diseases 0.000 claims 1
- 230000006870 function Effects 0.000 description 15
- 238000010586 diagram Methods 0.000 description 9
- 241000251468 Actinopterygii Species 0.000 description 7
- 230000008859 change Effects 0.000 description 6
- 230000004913 activation Effects 0.000 description 5
- 230000008901 benefit Effects 0.000 description 4
- 230000014509 gene expression Effects 0.000 description 4
- 210000000056 organ Anatomy 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 230000002087 whitening effect Effects 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 3
- 230000002146 bilateral effect Effects 0.000 description 2
- 238000004590 computer program Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000002708 enhancing effect Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000011551 log transformation method Methods 0.000 description 2
- 239000003595 mist Substances 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 238000002834 transmittance Methods 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000000877 morphologic effect Effects 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- 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
- G06T2207/30201—Face
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
Abstract
The present embodiment is related to advertisement promotion technical field, provides a kind of advertisement generation method, apparatus and system.The described method includes: obtaining the reference picture comprising target person of camera acquisition；Human face region in reference picture is split, the corresponding target facial image of target person is obtained；Target facial image is fused to the position to be replaced that each frame in default advertisement video presets advertisement frames, obtains targeted advertisements.Compared with prior art, the interest of advertisement can be enhanced in the embodiment of the present invention, improves spectators to the interest of advertisement.
Description
Technical field
The present invention relates to advertisement promotion technical fields, in particular to a kind of advertisement generation method, apparatus and system.
Background technique
In traditional advertisement, advertising is carried out generally by the mode of poster, video, still, since advertisement is pure
Essence is to publicize, and the interest of itself is not high, causes spectators not high to the interest of advertisement.
Summary of the invention
The purpose of the present invention is to provide a kind of advertisement generation method, apparatus and system, with improve advertisement for purely
Publicity, the interest of itself is not high, the status for causing interest of the spectators to advertisement not high.
The embodiment of the present invention this can be implemented so that
In a first aspect, the embodiment of the present invention provides a kind of advertisement generation method, applied to the processor of advertisement generating system,
The advertisement generating system further includes the camera with the processor communication, which comprises obtains the camera and adopts
The reference picture comprising target person of collection；Human face region in the reference picture is split, the target person is obtained
The corresponding target facial image of object；The target facial image is fused to each frame in default advertisement video and presets advertisement frames
Position to be replaced, obtains targeted advertisements.
Second aspect, the embodiment of the present invention provide a kind of advertisement generating means, applied to the processor of advertisement generating system,
The advertisement generating system further includes the camera with the processor communication, and the advertisement generating means include: acquisition module,
For obtaining the reference picture comprising target person of the camera acquisition；Processing module, for in the reference picture
Human face region be split, obtain the corresponding target facial image of the target person；The target facial image is merged
Into default advertisement video, each frame presets the position to be replaced of advertisement frames, obtains targeted advertisements.
The third aspect, the embodiment of the present invention provide a kind of advertisement generating system, and the advertisement generating system includes processor,
And the camera with the processor communication；The camera, the reference picture comprising target person for acquisition are concurrent
It send to the processor；The processor obtains the target for being split to the human face region in the reference picture
The corresponding target facial image of personage；The target facial image is fused to each frame in default advertisement video and presets advertisement frames
Position to be replaced, obtain targeted advertisements.
In terms of existing technologies, the application has the advantages that
The embodiment of the present invention provides a kind of advertisement generation method, apparatus and system, and being obtained by camera includes target person
The reference picture of object, and the human face region of reference picture is split, the corresponding target facial image of target person is obtained, most
Target facial image is fused to the position to be replaced that each frame in default advertisement video presets advertisement frames afterwards, it is wide to obtain target
It accuses.The position to be replaced that the facial image of spectators is preset to advertisement frames with each frame being fused in default advertisement video, generates
Targeted advertisements comprising spectators' facial image increase interacting for advertisement and spectators, enhance the interest of advertisement, so that spectators
Biggish interest can be generated to the targeted advertisements.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below will be to needed in the embodiment attached
Figure is briefly described, it should be understood that the following drawings illustrates only certain embodiments of the present invention, therefore is not construed as pair
The restriction of range for those of ordinary skill in the art without creative efforts, can also be according to this
A little attached drawings obtain other relevant attached drawings.
Fig. 1 shows the partial structure diagram of advertisement generating system provided by the embodiment of the present invention；
Fig. 2 shows the flow charts of advertisement generation method provided by the embodiment of the present invention；
Fig. 3 shows the first sub-step flow chart of step S3 provided by the embodiment of the present invention；
Fig. 4 shows the second sub-step flow chart of step S3 provided by the embodiment of the present invention；
Fig. 5 shows the block diagram of advertisement generating means provided by the embodiment of the present invention.
Icon: 100- advertisement generating system；101- processor；102- memory；103- bus；104- communication connects
Mouthful；105- display screen；106- camera；200- advertisement generating means；201- obtains module；202- processing module.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the present invention, instead of all the embodiments.The present invention being usually described and illustrated herein in the accompanying drawings is implemented
The component of example can be arranged and be designed with a variety of different configurations.
Therefore, the detailed description of the embodiment of the present invention provided in the accompanying drawings is not intended to limit below claimed
The scope of the present invention, but be merely representative of selected embodiment of the invention.Based on the embodiments of the present invention, this field is common
Technical staff's every other embodiment obtained without creative efforts belongs to the model that the present invention protects
It encloses.
It should also be noted that similar label and letter indicate similar terms in following attached drawing, therefore, once a certain Xiang Yi
It is defined in a attached drawing, does not then need that it is further defined and explained in subsequent attached drawing.
In the description of the present invention, it should be noted that if there is the side of the instructions such as term " on ", "lower", "inner", "outside"
Position or positional relationship be based on the orientation or positional relationship shown in the drawings or the invention product using when the side usually put
Position or positional relationship, are merely for convenience of description of the present invention and simplification of the description, rather than the device or member of indication or suggestion meaning
Part must have a particular orientation, be constructed and operated in a specific orientation, therefore be not considered as limiting the invention.
In addition, being only used for distinguishing description if there is term " first ", " second " etc., it is not understood to indicate or imply phase
To importance.
It should be noted that in the absence of conflict, the feature in the embodiment of the present invention can be combined with each other.
Advertisement generation method provided in an embodiment of the present invention is applied to advertisement generating system, referring to Fig. 1, Fig. 1 is shown
The block diagram of advertisement generating system provided in an embodiment of the present invention.Advertisement generating system 100 includes processor 101, storage
Device 102, bus 103, communication interface 104 and display screen 105 and camera 106.Processor 101, memory 102, communication interface
104, display screen 105 and camera 106 are connected by bus 103, and processor 101 is communicated with camera 106, and processor 101 is used
The executable module stored in execution memory 102, such as computer program.
Processor 101 may be a kind of IC chip, the processing capacity with signal.During realization, advertisement
Each step of generation method can be complete by the integrated logic circuit of the hardware in processor 101 or the instruction of software form
At.Above-mentioned processor 101 can be general processor, including central processing unit (Central Processing Unit, letter
Claim CPU), network processing unit (Network Processor, abbreviation NP) etc.；It can also be digital signal processor (Digital
Signal Processor, abbreviation DSP), specific integrated circuit (Application Specific Integrated
Circuit, abbreviation ASIC), ready-made programmable gate array (Field-Programmable Gate Array, abbreviation FPGA) or
Person other programmable logic device, discrete gate or transistor logic, discrete hardware components.
Memory 102 may include high-speed random access memory (RAM:Random Access Memory), it is also possible to
It further include non-labile memory (non-volatile memory), for example, at least a magnetic disk storage.
Bus 103 can be ISA (Industry Standard Architecture) bus, PCI (Peripheral
Component Interconnect) bus or EISA (Extended Industry Standard Architecture) be total
Line etc..It is only indicated with a four-headed arrow in Fig. 1, it is not intended that an only bus or a type of bus.
Advertisement generating system 100 is realized and outside by least one communication interface 104 (can be wired or wireless)
Other equipment between communication connection.Memory 102 is for storing program, such as advertisement generating means 200.Advertisement generates dress
Setting 200 includes that at least one can be stored in the memory 102 or be solidificated in the form of software or firmware (firmware)
Software function module in the operating system (operating system, OS) of advertisement generating system 100.The processor 101
After receiving and executing instruction, described program is executed to realize advertisement generation method.
For display screen 105 for showing to image, the content of display can be some processing results of processor 101.
Display screen 105 can be touch display screen, display screen of no interactions function etc..
Camera 106 is for acquiring/shooting the reference picture comprising target person and be sent to processor by bus 103
101 are handled.
It should be understood that structure shown in FIG. 1 is only the structure application schematic diagram of advertisement generating system 100, advertisement is raw
It may also include than shown in Fig. 1 more perhaps less component at system 100 or with the configuration different from shown in Fig. 1.Figure
Each component shown in 1 can be realized using hardware, software, or its combination.
It should be noted that advertisement generating system 100 can be an integrated electronic equipment, camera 106 and its
His functional module can be discrete and set, for example, in a scenario, in a tunnel, camera shooting is set when entrance
First 106, the reference picture taken is sent to processor 101 and handled by camera 106, and then processor 101 will be handled
As a result it when pedestrian passes through, is shown by display screen 105；Certainly, camera 106 can also be integrated with other functional modules
Together, when camera 106 collects reference picture, it is sent to processor 101, so that processor 101 uses advertisement generation side
Method handles the reference picture, generates targeted advertisements, and show by display screen 105.
Based on above-mentioned advertisement generating system 100, a kind of possible implementation of advertisement generation method, the party is given below
The executing subject of method can be the processor 101 of above-mentioned advertisement generating system 100, referring to Fig. 2, Fig. 2 shows realities of the present invention
A kind of flow chart of advertisement generation method of example offer is provided.Advertisement generation method the following steps are included:
S1 obtains the reference picture comprising target person of camera acquisition.
In embodiments of the present invention, target person can occur from the personage in the shooting area of camera, with reference to figure
As can be camera shooting/acquisition and include the image of target person.Obtain the ginseng comprising target person of camera acquisition
The step of examining image, it can be understood as, camera is mounted on to the region that can take pedestrian face, then, to by taking the photograph
As the pedestrian in head captured in real-time region shoots, the reference picture comprising target person is obtained.
It should be noted that target person can be all people's object in shooting area, place in shooting area can be
One or more personage in fixed position (for example, middle), the embodiment of the present invention are not limited thereto.
S2 is split the human face region in reference picture, obtains the corresponding target facial image of target person.
In embodiments of the present invention, target facial image can be target person in reference picture human face region it is corresponding
Image.The step of human face region in reference picture is split, target person corresponding facial image is obtained, it is possible to understand that
For firstly, then the position where detecting human face region in reference picture according to the position where human face region, will join
The human face region examined in image is split, and obtains the corresponding target facial image of target person.
It as an implementation, using method pixel-based or can be method or base based on boundary
The segmentation of human face region is carried out in the method in region.Wherein, the face based on the colour of skin point is contained in method pixel-based
It cuts, specific mode are as follows: first, reference picture is transformed into YCbCr, and (Y- luminance component, Cb- chroma blue component, Cr- are red
Chroma color component) space；Second, Cr component therein is subjected to OTSU Threshold segmentation；Third seeks the profile of segmentation result；
4th, the profile is filled, face's exposure mask figure is obtained, and according to face's exposure mask figure, obtains target human face region.
It, can be using deep in order to which reliable recognition goes out the face inside reference picture in other embodiments of the invention
Learning model is spent to identify face, for example, YOLO model, since the frame of YOLO model is not high to hardware requirement, in terms of speed
Compared to Faster R-CNN (Region Convolutional Neural Network), the networks such as Mask R-CNN are more
With advantage.Method particularly includes:
First, being marked out face with picture marking software on the basis of existing database can be with come title when, mark
It is determined according to businessman's demand.Such as the face of male and the face of women can be divided into according to gender, or be divided into according to the age
The face of child and the face of adult.Training data is made according to certain format data set after the completion of mark, to train YOLO mould
Type.
Second, weight file is obtained in YOLO model from training, and the weight file is based on, in intelligent measurement reference picture
Face.
Before step S2, may be used also in other embodiments of the present invention to improve the precision of human face region segmentation
To include being pre-processed to reference picture, so that reference picture is more clear.
As an implementation, reference picture is pre-processed, it can be understood as, image is carried out to reference picture and is gone
It makes an uproar, and/or image defogging, and/or image enhancement etc..
Image denoising is carried out to reference picture, it is ensured that under the premise of reference picture accuracy, remove reference picture
In most noise.The mode of image denoising may be, but not limited to, bilateral filtering, morphologic filtering, based on region growing
Filtering, gaussian filtering, median filtering, mean filter etc..
Image defogging is carried out to reference picture, interference of the foggy weather to reference picture can be excluded, avoid the big greasy weather
Gas bring influences.The main flow of image defogging is: first, seek the dark of reference picture；Second, it is calculated using dark
Transmissivity obtains transmittance figure；Third, the transmittance figure fining based on Steerable filter；4th, atmosphere is estimated using dark
Light；5th, it is based on mist graph model (that is, mist figure formula) defogging.
Image enhancement is carried out to reference picture, information useful in reference picture can be enhanced, may include based on histogram
Scheme equalization image enhancement, the image enhancement based on gamma transformation, based on object log transformation image enhancement, based on draw it is general
The image enhancement etc. of Laplacian operater.Wherein, the image enhancement based on histogram equalization is mainly used for enhancing the contrast of image；
The too high or too low picture of gray scale can be modified by the image enhancement based on gamma transformation, enhance contrast；Based on object
The image enhancement of log transformation can extend the low gray portion of image, and high gray scale is partially compressed, so that it is low to reach enhancing image
The purpose of gray portion；The image enhancement of base Laplace operator is mainly used for sharpening image.
Target facial image is fused to the position to be replaced that each frame in default advertisement video presets advertisement frames, obtained by S3
To targeted advertisements.
In embodiments of the present invention, default advertisement video can be needs that advertiser is stored in advertisement generating system into
The advertisement video that row is promoted, default advertisement video may include multiple default advertisement frames, and all advertisement video frames collectively constitute
Default advertisement video contains area image to be replaced, area image to be replaced, that is, target to be replaced institute in default advertisement frames
Area image, swum in the sea for example, the content in default advertisement frames can be small fish big, target to be replaced is the fish of small fish
Head, area image to be replaced can be the image of the fish head part of small fish, and every frame presets area image institute to be replaced in advertisement frames
The position i.e. position to be replaced of the default advertisement frames, it should be noted that since default advertisement video is one dynamic
The position to be replaced of process, different default advertisement frames may be different.
Referring to Fig. 3, step S3 may include following sub-step:
S31, the region to be replaced for presetting advertisement frames to each frame in default advertisement video are split, and obtain each frame
The default corresponding area image to be replaced of advertisement frames.
In embodiments of the present invention, region to be replaced can be the region where target to be replaced, and target to be replaced can be with
It is the object in default advertisement video or personage.For example, basketball, small fish, personage etc..Target to be replaced can be extensively
Accuse quotient storage advertisement be it is selected, can be what ad system detected automatically, the embodiment of the present invention is not limited thereto.
It should be noted that the quantity of target to be replaced can be one, it is also possible to multiple, the embodiment of the present invention is herein
It is not construed as limiting.For example, when the content of default advertisement frames is that 3 football players play soccer on football pitch, then this is default wide
The target to be replaced accused in frame can be football, can also be with one or more of which football player, can be with existing foot
Ball, and have football player.
The region to be replaced for presetting advertisement frames to each frame in default advertisement video is split, and it is default to obtain each frame
The step of advertisement frames corresponding area image to be replaced, it can be understood as, firstly, determining each frame based on above-mentioned mode
Target to be replaced in default advertisement frames, then, it is determined that target to be replaced is in the position where presetting in advertisement frames out, most
Afterwards, according to the region where target to be replaced, i.e., it is pre- to be partitioned into this in the position for presetting place in advertisement frames for region to be replaced
If the area image to be replaced in advertisement frames.
Each frame in default advertisement video is preset the corresponding area image to be replaced of advertisement frames and is updated to target by S32
Facial image obtains targeted advertisements.
In embodiments of the present invention, each frame in default advertisement video is preset into the corresponding administrative division map to be replaced of advertisement frames
As the step of being updated to target facial image, obtain targeted advertisements, it can be understood as, frame each in default advertisement video is preset
The corresponding area image to be replaced of advertisement frames replaces with target facial image, or covers default advertisement using target facial image and regard
Each frame in frequency presets the corresponding area image to be replaced of advertisement frames.
It should be noted that when area image to be replaced and target facial image it is not of uniform size when, need target
Facial image zooms in and out processing, so that the size of area image to be replaced and target facial image is consistent.
It should be noted that it is divided into two kinds of situations when the quantity of target to be replaced and target person is inconsistent, one is
When the quantity of target to be replaced is greater than the quantity of target person, only by the area to be replaced where a part of target to be replaced therein
Area image is updated to target facial image；It, can be right when another kind is that the quantity of target to be replaced is less than the quantity of target person
Priority level initializing is carried out to target person according to shooting area and selects a part of mesh according to the sequence of priority from high to low
It marks facial image and participates in renewal process.
Referring to Fig. 4, step S3 can also include following sub-step:
S33 carries out personalisation process to target facial image, obtains individual face face image, wherein target facial image
It is consistent with personalized human face image section.
In embodiments of the present invention, individual face face image, which can be, carries out personalisation process to target facial image, obtains
The image arrived.The individual face face image obtained by personalisation process is consistent with target facial image part, needs to embody
It is that the individual face face image obtained after processing has certain identification for spectators, it can be seen that be to belong to
In who spectators.
The step of is carried out by personalisation process, obtains individual face face image for target facial image, it can be understood as, to mesh
It marks facial image and carries out sketch processing, and/or filtering processing, and/or cartooning's processing, obtain individual face face image.
The main flow for carrying out sketch processing to target facial image is: first, La Pula is carried out to target facial image
This image enhancement obtains 1# image；Second, Sobel gradient detection is carried out to target facial image, obtains 2# image；Third,
1# image is added according to a certain percentage (for example, 0.2 and 0.8) with 2# image, obtains 3# image；4th, 3# image is carried out
Sketch treatment process can be completed in bilateral filtering.
Target face is filtered, it can be understood as, it is that mill skin and whitening processing are carried out to target facial image,
Its main flow is: first, mill skin processing is carried out to facial image based on Steerable filter algorithm, obtains the first image；Second, it is right
Each pixel on first image carries out linear process to change the contrast and brightness of image, and whitening processing can be completed.
In other embodiments of the present invention, can also by the way of convolutional neural networks, to target facial image into
Row whitening processing specifically inputs target facial image in one four layers of convolutional neural networks, which is 64x
The target facial image of 64x 3, the target facial image after exporting 3 whitening of 64x 64x, structure are as follows:
First layer: kernel size=9, Stride=1, Padding=same, number of kernel=32 swashs
Function living is line rectification function (Rectified Linear Unit, ReLU).
The second layer: kernel size=3, Stride=1, Padding=same, number of kernel=64 swashs
Function living is ReLU.
Third layer: kernel size=3, Stride=1, Padding=same, number of kernel=64 swashs
Function living is ReLU.
4th layer: kernel size=3, Stride=1, Padding=same, number of kernel=3, nothing
Activation primitive.
Wherein, kernel size refers to the size of convolution kernel；Stride refers to the mobile step-length of convolution kernel；
Padding refers to edge whether is filled in convolution process, and Padding=same is mainly used for keeping boundary information, also
It is to utilize boundary information as far as possible, while keeping input picture consistent with the output size of image；number of kernel
Indicate the quantity of convolution kernel, the corresponding characteristic pattern (feature map) of each convolution kernel；Addition activation primitive purpose be
Help depth network that study is gone to understand complicated nonlinear model, ReLU function is very common at present, its expression formula are as follows:
F (x)=max (0, x)
ReLU function overcomes the problem of gradient disappears, while accelerating network training speed.
First layer network interpretation: the 64x 64x 3 of input picture size, wherein 64x 64 indicates the size of image, and 3 indicate
Picture depth is 3 because being rgb triple channel image.This layer convolution kernel of 32 9x 9, step-length Stride=1,
ReLU activation primitive has been used after carrying out convolution to input picture, simultaneously because Padding=same, therefore finally export 64x
The characteristic pattern of 64x 32, wherein 32 indicate depth, 64x 64 indicates feature map size.
The second layer is explained: being inputted as 64x 64x 32, this layer convolution kernel of 64 3x 3, step-length Stride=1 is right
Input picture has used ReLU activation primitive after carrying out convolution, simultaneously because Padding=same, therefore exporting is 64x 64x
64 characteristic pattern.
Third layer is explained: input size is 64x 64x 64, this layer convolution kernel of 64 3x 3, step-length Stride=
1, ReLU activation primitive has been used after carrying out convolution to input picture, simultaneously because Padding=same, therefore exporting is 64x
The characteristic pattern of 64x 64.
4th layer is explained: input size is 64x 64x 64, this layer convolution kernel of 3 3x 3, step-length Stride=
1, convolution is carried out to input picture, simultaneously because Padding=same, therefore export the coloured picture for being 64x 64x 3.
It can be found that network requirement input 64x 64x 3 image, it is therefore desirable to first by the size of target facial image into
Row adjustment, in addition the size of network final output image and the size of input picture are consistent.
Under normal circumstances as the intensification of network, the quantity of convolution kernel are stepped up, feature map is gradually reduced, this
Sample can sufficiently extract the representative feature of preceding layer.Since above-mentioned network structure is simple, whole network only has four layers, because
Feature map size does not change during this.Test discovery network speed is very fast, works well.
The main flow for carrying out cartooning's processing to target facial image is: first, detect people in target facial image
88 characteristic points of face, decompose the organ on face；Second, it is sought in true face library according to each organ of decomposition
Look for most like organ；Third finds the corresponding cartoon figure in pre-existing advertisement generating system according to most like organ,
It is assembled, finally obtains complete cartoon figure.
S34, the region to be replaced for presetting advertisement frames to each frame in default advertisement video are split, and obtain each frame
The default corresponding area image to be replaced of advertisement frames.
In embodiments of the present invention, region to be replaced can be the region where target to be replaced, and target to be replaced can be with
It is the object in default advertisement video or personage.For example, basketball, small fish, personage etc..Target to be replaced can be extensively
Accuse quotient storage advertisement be it is selected, can be what ad system detected automatically, the embodiment of the present invention is not limited thereto.
It should be noted that the quantity of target to be replaced can be one, it is also possible to multiple, the embodiment of the present invention is herein
It is not construed as limiting.For example, when the content of default advertisement frames is that 3 football players play soccer on football pitch, then this is default wide
The target to be replaced accused in frame can be football, can also be with one or more of which football player, can be with existing foot
Ball, and have football player.
The region to be replaced for presetting advertisement frames to each frame in default advertisement video is split, and it is default to obtain each frame
The step of advertisement frames corresponding area image to be replaced, it can be understood as, firstly, determining each frame based on above-mentioned mode
Target to be replaced in default advertisement frames, then, it is determined that target to be replaced is in the position where presetting in advertisement frames out, most
Afterwards, according to the region where target to be replaced, i.e., it is pre- to be partitioned into this in the position for presetting place in advertisement frames for region to be replaced
If the area image to be replaced in advertisement frames.
Each frame in default advertisement video is preset the corresponding area image to be replaced of advertisement frames and is updated to individual character by S35
Change facial image, obtains targeted advertisements.
In embodiments of the present invention, each frame in default advertisement video is preset into the corresponding administrative division map to be replaced of advertisement frames
As the step of being updated to individual face face image, obtain targeted advertisements, it can be understood as, frame each in default advertisement video is pre-
If the corresponding area image to be replaced of advertisement frames replaces with individual face face image, or is covered and preset using individual face face image
Each frame in advertisement video presets the corresponding area image to be replaced of advertisement frames.
In other embodiments of the present invention, in order to enable the interest of advertisement is more sufficient, multiple can be acquired with reference to figure
The corresponding target facial image of each reference picture is fused to the position to be replaced that a frame presets advertisement frames, Ke Yili by picture
The target facial image that first reference picture obtains is fused to first frame and presets advertisement frames, by second with reference to figure by Xie Wei
Advertisement frames are preset as obtained target facial image is fused to the second frame, and so on, then obtained targeted advertisements
It is multiple expressions, it is abundant in content.
A simply example is set forth below out, is illustrated come the whole process effect to the present embodiment:
Default advertisement video is stored in advertisement generating system by certain basketball training organization (advertiser), when having detected row
When people passes through shooting area, it will be filmed by the pedestrian in region, handled, obtain targeted advertisements, and on a display screen
The targeted advertisements are played, it is possible that the interesting picture that the user plays basketball in basketball court on display screen, and at
The expression of reason can seem very interesting, while can also be by the expression real-time display of the user on a display screen.
Compared with prior art, the embodiment of the present invention has the advantage that
Firstly, being more clear reference picture by image preprocessing, and then improve the precision of human face region segmentation.
Secondly, by by the facial image of spectators and each frame for being fused in default advertisement video preset advertisement frames to
Position is replaced, the targeted advertisements comprising spectators' facial image is generated, enhances the interest of advertisement, increases advertisement and spectators
Interaction allows the viewer to generate biggish interest to the targeted advertisements.
Finally, target facial image, which is carried out personalisation process, can more improve advertisement to enhance the interest of target face
The interest of itself.
For the method flow of above-mentioned Fig. 2, Fig. 3 and Fig. 4, a kind of possible reality of advertisement generating means 200 is given below
Existing mode, the advertisement generating means 200 can use the device architecture of the advertisement generating system 100 in above-described embodiment to realize,
Or the processor 101 in the advertisement generating system 100 is realized, is mentioned referring to Fig. 5, Fig. 5 shows the embodiment of the present invention
The block diagram of the advertisement generating means of confession.Advertisement generating means 200 include obtaining module 201 and processing module 202.
Module 201 is obtained, for obtaining the reference picture comprising target person of camera acquisition.
Processing module 202 obtains the corresponding target of target person for being split to the human face region in reference picture
Facial image；Target facial image is fused to the position to be replaced that each frame in default advertisement video presets advertisement frames, is obtained
Targeted advertisements.
In embodiments of the present invention, default advertisement video includes that multiframe presets advertisement frames, and every frame presets advertisement frames and includes
Area image to be replaced, position to be replaced are the position where area image to be replaced, and processing module 202 is executed target face
The step of image co-registration each frame into default advertisement video presets the position to be replaced of advertisement frames, obtains targeted advertisements, specifically
For: the region to be replaced for presetting advertisement frames to each frame in default advertisement video is split, and it is default wide to obtain each frame
Accuse the corresponding area image to be replaced of frame；Each frame in default advertisement video is preset into the corresponding administrative division map to be replaced of advertisement frames
As being updated to target facial image, targeted advertisements are obtained.
In embodiments of the present invention, processing module 202, which is executed, presets advertisement frames pair for each frame in default advertisement video
The step of area image to be replaced answered is updated to target facial image, obtains targeted advertisements, is specifically used for: default advertisement is regarded
Each frame presets the corresponding area image to be replaced of advertisement frames and replaces with target facial image in frequency；Or utilize target facial image
The each frame covered in default advertisement video presets the corresponding area image to be replaced of advertisement frames.
In embodiments of the present invention, default advertisement video includes that multiframe presets advertisement frames, and every frame presets advertisement frames and includes
Area image to be replaced, position to be replaced are the position where area image to be replaced, and processing module 202 is executed target face
The step of image co-registration each frame into default advertisement video presets the position to be replaced of advertisement frames, obtains targeted advertisements, specifically
For: personalisation process is carried out to target facial image, obtains individual face face image, wherein target facial image and individual character
It is consistent to change facial image part；The region to be replaced for presetting advertisement frames to each frame in default advertisement video is split, and is obtained
The corresponding area image to be replaced of advertisement frames is preset to each frame；It is corresponding that each frame in default advertisement video is preset into advertisement frames
Area image to be replaced be updated to individual face face image, obtain targeted advertisements.
In embodiments of the present invention, processing module 202, which is executed, carries out personalisation process to target facial image, obtains individual character
The step of changing facial image, is specifically used for: carrying out sketch processing, and/or filtering processing, and/or cartoon to target facial image
Change processing, obtains individual face face image.
In embodiments of the present invention, processing module 202 is also used to: image preprocessing is carried out to reference picture, so that with reference to
Image is more clear.
It is apparent to those skilled in the art that for convenience and simplicity of description, the advertisement of foregoing description
The specific work process of generating means 200, can refer to corresponding processes in the foregoing method embodiment, and details are not described herein.
In conclusion the embodiment of the present invention provides a kind of advertisement generation method, apparatus and system, which comprises obtain
The reference picture comprising target person for taking camera to acquire；Human face region in reference picture is split, target is obtained
The corresponding target facial image of personage；By target facial image be fused to each frame in default advertisement video preset advertisement frames to
Position is replaced, targeted advertisements are obtained.By the way that the facial image of spectators is preset with each frame being fused in default advertisement video
The position to be replaced of advertisement frames, generates the targeted advertisements comprising spectators' facial image, enhances the interest of advertisement, increases wide
Announcement is interacted with spectators', allows the viewer to generate biggish interest to the targeted advertisements.
In embodiment provided herein, it should be understood that disclosed device and method, it can also be by other
Mode realize.The apparatus embodiments described above are merely exemplary, for example, the flow chart and block diagram in attached drawing are shown
Device, the architectural framework in the cards of method and computer program product, function of multiple embodiments according to the present invention
And operation.In this regard, each box in flowchart or block diagram can represent one of a module, section or code
Point, a part of the module, section or code includes one or more for implementing the specified logical function executable
Instruction.It should also be noted that function marked in the box can also be attached to be different from some implementations as replacement
The sequence marked in figure occurs.For example, two continuous boxes can actually be basically executed in parallel, they sometimes may be used
To execute in the opposite order, this depends on the function involved.It is also noted that each of block diagram and or flow chart
The combination of box in box and block diagram and or flow chart can be based on the defined function of execution or the dedicated of movement
The system of hardware is realized, or can be realized using a combination of dedicated hardware and computer instructions.
In addition, each functional module in each embodiment of the present invention can integrate one independent portion of formation together
Point, it is also possible to modules individualism, an independent part can also be integrated to form with two or more modules.
It, can be with if the function is realized and when sold or used as an independent product in the form of software function module
It is stored in a computer readable storage medium.Based on this understanding, technical solution of the present invention is substantially in other words
The part of the part that contributes to existing technology or the technical solution can be embodied in the form of software products, the meter
Calculation machine software product is stored in a storage medium, including some instructions are used so that a computer equipment (can be a
People's computer, server or network equipment etc.) it performs all or part of the steps of the method described in the various embodiments of the present invention.
And storage medium above-mentioned include: USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), RAM, magnetic disk or
The various media that can store program code such as person's CD.
It should be noted that, in this document, relational terms such as first and second and the like are used merely to a reality
Body or operation are distinguished with another entity or operation, are deposited without necessarily requiring or implying between these entities or operation
In any actual relationship or order or sequence.Moreover, the terms "include", "comprise" or its any other variant are intended to
Non-exclusive inclusion, so that the process, method, article or equipment including a series of elements is not only wanted including those
Element, but also including other elements that are not explicitly listed, or further include for this process, method, article or equipment
Intrinsic element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that
There is also other identical elements in process, method, article or equipment including the element.
The foregoing is only a preferred embodiment of the present invention, is not intended to restrict the invention, for the skill of this field
For art personnel, the invention may be variously modified and varied.All within the spirits and principles of the present invention, made any to repair
Change, equivalent replacement, improvement etc., should all be included in the protection scope of the present invention.It should also be noted that similar label and letter exist
Similar terms are indicated in following attached drawing, therefore, once being defined in a certain Xiang Yi attached drawing, are then not required in subsequent attached drawing
It is further defined and explained.
Claims (10)
1. a kind of advertisement generation method, which is characterized in that applied to the processor of advertisement generating system, the advertisement generating system
It further include the camera with the processor communication, which comprises
Obtain the reference picture comprising target person of the camera acquisition；
Human face region in the reference picture is split, the corresponding target facial image of the target person is obtained；
The target facial image is fused to the position to be replaced that each frame in default advertisement video presets advertisement frames, obtains mesh
Mark advertisement.
2. the method according to claim 1, wherein the default advertisement video include multiframe preset advertisement frames,
It includes area image to be replaced that advertisement frames are preset described in every frame, and the position to be replaced is the area image place to be replaced
Position, it is described that the target facial image is fused to the position to be replaced that each frame in default advertisement video presets advertisement frames
The step of setting, obtaining targeted advertisements, comprising:
The region to be replaced for presetting advertisement frames to each frame in default advertisement video is split, and is obtained each frame and is preset advertisement
The corresponding area image to be replaced of frame；
Each frame in the default advertisement video is preset into the corresponding area image to be replaced of advertisement frames and is updated to the target
Facial image obtains targeted advertisements.
3. according to the method described in claim 2, it is characterized in that, each frame by the default advertisement video is preset
The step of corresponding area image to be replaced of advertisement frames is updated to the target facial image, obtains targeted advertisements, comprising:
Each frame in the default advertisement video is preset into the corresponding area image to be replaced of advertisement frames and replaces with the target person
Face image；
Or
It is corresponding to be replaced that the default advertisement frames of each frame in the default advertisement video are covered using the target facial image
Area image.
4. the method according to claim 1, wherein the default advertisement video include multiframe preset advertisement frames,
It includes area image to be replaced that advertisement frames are preset described in every frame, and the position to be replaced is the area image place to be replaced
Position, it is described that the target facial image is fused to the position to be replaced that each frame in default advertisement video presets advertisement frames
The step of setting, obtaining targeted advertisements, comprising:
Personalisation process is carried out to the target facial image, obtains individual face face image, wherein the target facial image
It is consistent with the personalized human face image section；
The region to be replaced for presetting advertisement frames to each frame in default advertisement video is split, and is obtained each frame and is preset advertisement
The corresponding area image to be replaced of frame；
Each frame in the default advertisement video is preset into the corresponding area image to be replaced of advertisement frames and is updated to the individual character
Change facial image, obtains targeted advertisements.
5. according to the method described in claim 4, it is characterized in that, described carry out at personalization the target facial image
The step of managing, obtaining individual face face image, comprising:
Sketch processing, and/or filtering processing, and/or cartooning's processing are carried out to the target facial image, obtain individual face
Face image.
6. the method according to claim 1, wherein being carried out in the human face region in the reference picture
Segmentation, before the step of obtaining the target person corresponding target facial image, the method also includes:
Image preprocessing is carried out to the reference picture, so that the reference picture is more clear.
7. a kind of advertisement generating means, which is characterized in that applied to the processor of advertisement generating system, the advertisement generating system
It further include the camera with the processor communication, the advertisement generating means include:
Module is obtained, for obtaining the reference picture comprising target person of the camera acquisition；
Processing module obtains the corresponding mesh of the target person for being split to the human face region in the reference picture
Mark facial image；The target facial image is fused to the position to be replaced that each frame in default advertisement video presets advertisement frames
It sets, obtains targeted advertisements.
8. a kind of advertisement generating system, which is characterized in that the advertisement generating system includes processor, and with the processor
The camera of communication；
The camera, for acquiring the reference picture comprising target person and being sent to the processor；
It is corresponding to obtain the target person for being split to the human face region in the reference picture for the processor
Target facial image；The target facial image is fused to the position to be replaced that each frame in default advertisement video presets advertisement frames
It sets, obtains targeted advertisements.
9. advertisement generating system according to claim 8, which is characterized in that the default advertisement video includes that multiframe is default
Advertisement frames, it includes area image to be replaced that advertisement frames are preset described in every frame, and the position to be replaced is the region to be replaced
Position where image, the processor are also used to:
The region to be replaced for presetting advertisement frames to each frame in default advertisement video is split, and is obtained each frame and is preset advertisement
The corresponding area image to be replaced of frame；
Each frame in the default advertisement video is preset into the corresponding area image to be replaced of advertisement frames and is updated to the target
Facial image obtains targeted advertisements.
10. advertisement generating system according to claim 9, which is characterized in that the processor is also used to:
Each frame in the default advertisement video is preset into the corresponding area image to be replaced of advertisement frames and replaces with the target person
Face image；
Or
It is corresponding to be replaced that the default advertisement frames of each frame in the default advertisement video are covered using the target facial image
Area image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910820682.3A CN110517187A (en) | 2019-08-30 | 2019-08-30 | Advertisement generation method, apparatus and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910820682.3A CN110517187A (en) | 2019-08-30 | 2019-08-30 | Advertisement generation method, apparatus and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110517187A true CN110517187A (en) | 2019-11-29 |
Family
ID=68630117
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910820682.3A Pending CN110517187A (en) | 2019-08-30 | 2019-08-30 | Advertisement generation method, apparatus and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110517187A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111145189A (en) * | 2019-12-26 | 2020-05-12 | 成都市喜爱科技有限公司 | Image processing method, image processing device, electronic equipment and computer readable storage medium |
| WO2022055424A3 (en) * | 2020-09-09 | 2022-05-05 | 脸萌有限公司 | Video photographing method and apparatus, and device and storage medium |
| US12041374B2 (en) | 2020-09-09 | 2024-07-16 | Lemon Inc. | Segmentation-based video capturing method, apparatus, device and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103455800A (en) * | 2013-09-09 | 2013-12-18 | 苏州大学 | Advertisement system based on intelligent identification and method for pushing corresponding advertisement |
| CN105118082A (en) * | 2015-07-30 | 2015-12-02 | 科大讯飞股份有限公司 | Personalized video generation method and system |
| CN106792147A (en) * | 2016-12-08 | 2017-05-31 | 天脉聚源(北京)传媒科技有限公司 | A kind of image replacement method and device |
| WO2018033143A1 (en) * | 2016-08-19 | 2018-02-22 | 北京市商汤科技开发有限公司 | Video image processing method, apparatus and electronic device |
| CN107767325A (en) * | 2017-09-12 | 2018-03-06 | 深圳市朗形网络科技有限公司 | Method for processing video frequency and device |
| CN109996107A (en) * | 2017-12-29 | 2019-07-09 | 百度在线网络技术（北京）有限公司 | Video generation method, device and system |
-
2019
- 2019-08-30 CN CN201910820682.3A patent/CN110517187A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103455800A (en) * | 2013-09-09 | 2013-12-18 | 苏州大学 | Advertisement system based on intelligent identification and method for pushing corresponding advertisement |
| CN105118082A (en) * | 2015-07-30 | 2015-12-02 | 科大讯飞股份有限公司 | Personalized video generation method and system |
| WO2018033143A1 (en) * | 2016-08-19 | 2018-02-22 | 北京市商汤科技开发有限公司 | Video image processing method, apparatus and electronic device |
| CN106792147A (en) * | 2016-12-08 | 2017-05-31 | 天脉聚源(北京)传媒科技有限公司 | A kind of image replacement method and device |
| CN107767325A (en) * | 2017-09-12 | 2018-03-06 | 深圳市朗形网络科技有限公司 | Method for processing video frequency and device |
| CN109996107A (en) * | 2017-12-29 | 2019-07-09 | 百度在线网络技术（北京）有限公司 | Video generation method, device and system |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111145189A (en) * | 2019-12-26 | 2020-05-12 | 成都市喜爱科技有限公司 | Image processing method, image processing device, electronic equipment and computer readable storage medium |
| CN111145189B (en) * | 2019-12-26 | 2023-08-08 | 成都市喜爱科技有限公司 | Image processing method, apparatus, electronic device, and computer-readable storage medium |
| WO2022055424A3 (en) * | 2020-09-09 | 2022-05-05 | 脸萌有限公司 | Video photographing method and apparatus, and device and storage medium |
| US12041374B2 (en) | 2020-09-09 | 2024-07-16 | Lemon Inc. | Segmentation-based video capturing method, apparatus, device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhao et al. | Dehazing evaluation: Real-world benchmark datasets, criteria, and baselines | |
| CN104881642B (en) | A kind of content delivery method, device and equipment | |
| US9013489B2 (en) | Generation of avatar reflecting player appearance | |
| CN105139004B (en) | Facial expression recognizing method based on video sequence | |
| KR100355382B1 (en) | Apparatus and method for generating object label images in video sequence | |
| US7941002B2 (en) | Apparatus and methods of producing photorealistic image thumbnails | |
| CN108549874A (en) | A kind of object detection method, equipment and computer readable storage medium | |
| CN108388882A (en) | Based on the gesture identification method that the overall situation-part is multi-modal RGB-D | |
| Amirshahi et al. | Jenaesthetics subjective dataset: analyzing paintings by subjective scores | |
| CN109325988A (en) | A kind of facial expression synthetic method, device and electronic equipment | |
| CN107644213A (en) | Video person extraction method and device | |
| CN108280426B (en) | Dark light source expression identification method and device based on transfer learning | |
| CN110378945A (en) | Depth map processing method, device and electronic equipment | |
| CN105849685A (en) | Editing options for image regions | |
| CN103984778B (en) | A kind of video retrieval method and system | |
| CN107820018A (en) | User's photographic method, device and equipment | |
| CN104794693A (en) | Human image optimization method capable of automatically detecting mask in human face key areas | |
| CN110517187A (en) | Advertisement generation method, apparatus and system | |
| CN107506738A (en) | Feature extracting method, image-recognizing method, device and electronic equipment | |
| CN109388725A (en) | The method and device scanned for by video content | |
| CN111143615A (en) | Short video emotion classification recognition device | |
| Qi et al. | Global and local information based deep network for skin lesion segmentation | |
| Zhao et al. | Emotion based image musicalization | |
| JP2004030007A (en) | Makeup simulation apparatus, makeup simulation method, makeup simulation program and recording medium with program recorded thereon | |
| CN112132750B (en) | Video processing method and device |
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 |