CN117408760A - Picture display method and system based on artificial intelligence - Google Patents
Picture display method and system based on artificial intelligence Download PDFInfo
- Publication number
- CN117408760A CN117408760A CN202311714899.9A CN202311714899A CN117408760A CN 117408760 A CN117408760 A CN 117408760A CN 202311714899 A CN202311714899 A CN 202311714899A CN 117408760 A CN117408760 A CN 117408760A
- Authority
- CN
- China
- Prior art keywords
- crowd
- advertisement
- advertisement picture
- picture
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 28
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 27
- 230000009471 action Effects 0.000 claims description 68
- 230000011218 segmentation Effects 0.000 claims description 22
- 238000013527 convolutional neural network Methods 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 7
- 230000000694 effects Effects 0.000 description 10
- 238000012549 training Methods 0.000 description 8
- 238000000605 extraction Methods 0.000 description 6
- 210000003128 head Anatomy 0.000 description 6
- 238000013459 approach Methods 0.000 description 3
- 238000004590 computer program Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000002372 labelling Methods 0.000 description 3
- 230000035484 reaction time Effects 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000011478 gradient descent method Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000000875 corresponding effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000008034 disappearance Effects 0.000 description 1
- 230000003370 grooming effect Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0241—Advertisements
- G06Q30/0251—Targeted advertisements
- G06Q30/0269—Targeted advertisements based on user profile or attribute
- G06Q30/0271—Personalized advertisement
-
- G—PHYSICS
- G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
- G09F—DISPLAYING; ADVERTISING; SIGNS; LABELS OR NAME-PLATES; SEALS
- G09F9/00—Indicating arrangements for variable information in which the information is built-up on a support by selection or combination of individual elements
- G09F9/30—Indicating arrangements for variable information in which the information is built-up on a support by selection or combination of individual elements in which the desired character or characters are formed by combining individual elements
Abstract
The invention provides a picture display method and a system based on artificial intelligence, and relates to the technical field of picture display, wherein the method comprises the steps of determining average age, number, flowing speed, ambient light color and ambient light intensity of people based on video shot by the people; determining the playing time length of the advertisement picture by using the playing time length determining model based on the average age of the crowd, the crowd quantity, the crowd flow speed and the advertisement picture; determining the lowest display brightness of the advertisement picture by using a brightness determination model based on the crowd quantity, the ambient light color, the ambient light intensity and the advertisement picture; determining the display position of the advertisement picture on the screen based on the average age of the crowd, the crowd quantity, the crowd flow speed and the screen information; the method can reduce the consumption of electric power energy brought by the display of the advertisement pictures on the basis of the playing time of the advertisement pictures, the lowest display brightness of the advertisement pictures and the display position of the advertisement pictures on the screen.
Description
Technical Field
The invention relates to the technical field of picture display, in particular to a picture display method and system based on artificial intelligence.
Background
At present, with the development of society, the display of advertisement pictures through a screen becomes a main propaganda means, and more enterprises display advertisement pictures through the screen. Because the screen is larger, the enterprise screen is simpler and rough to manage. Whether people are beside the screen or not, the advertisement pictures can be played on the screen all the day, and the advertisement pictures are displayed with high brightness and occupy the whole screen, so that a large amount of electric power energy is consumed.
Therefore, how to reduce the consumption of electric power caused by displaying advertisement pictures is a current urgent problem to be solved.
Disclosure of Invention
The invention mainly solves the technical problem of how to reduce the consumption of electric power energy caused by the display of advertisement pictures.
According to a first aspect, the present invention provides an artificial intelligence based picture display method, including: acquiring crowd shooting videos and advertisement pictures; determining average age of the crowd, the crowd quantity, the crowd flow speed, the ambient light color and the ambient light intensity by using a crowd determination model based on the crowd shooting video; determining the playing time length of the advertisement picture based on the average age of the crowd, the crowd quantity, the crowd flow speed and the playing time length determining model of the advertisement picture; determining the lowest display brightness of the advertisement picture by using a brightness determination model based on the crowd quantity, the ambient light color, the ambient light intensity and the advertisement picture; determining the display position of the advertisement picture on a screen based on the average age of the crowd, the crowd quantity, the crowd flow speed and screen information; and displaying the advertisement picture based on the playing time of the advertisement picture, the lowest display brightness of the advertisement picture and the display position of the advertisement picture on a screen.
Furthermore, the crowd-determining model is a gating circulation unit model, the input of the crowd-determining model is that the crowd shoots a video, and the output of the crowd-determining model is that the average age of the crowd, the crowd quantity, the crowd flow speed, the ambient light color and the ambient light intensity.
Further, the playing time length determining model is a convolutional neural network model, the input of the playing time length determining model is the average age of the crowd, the crowd number, the crowd flow speed and the advertisement picture, and the output of the playing time length determining model is the playing time length of the advertisement picture.
Still further, the method further comprises: acquiring action videos of people in the playing time of the advertisement pictures when the advertisement pictures are played; processing and determining the attention degree of the crowd to the advertisement picture by sequentially using a segmentation model, a person average watching duration determination model and an attention degree determination model based on the action video of the crowd in the playing duration of the advertisement picture; if the attention degree of the crowd to the advertisement picture is smaller than the attention degree threshold value, a new advertisement picture is replaced for playing; if the attention degree of the crowd to the advertisement picture is greater than an attention degree threshold value, repeatedly playing the advertisement picture, wherein the attention degree threshold value is determined by the following formula:
wherein->、/>、/>Is->Are all preset weights, are->For the average age of the population->For the number of people, ->For the crowd flow rate,/->For the playing time of the advertisement picture, < + >>Is a threshold of attention.
Further, the segmentation model, the average person viewing time length determination model and the attention degree determination model are all gating circulation unit models, the input of the segmentation model is an action video of a crowd in the playing time length of the advertisement picture, the output of the segmentation model is an action video of each person in the crowd in the playing time length of the advertisement picture, the input of the average person viewing time length determination model is an action video of each person in the playing time length of the advertisement picture, the output of the average person viewing time length determination model is an average person viewing time length of each person in the crowd for the advertisement, the input of the attention degree determination model is an action video of each person in the playing time length of the advertisement picture, the average person viewing time length of each person in the crowd for the advertisement, and the output of the attention degree determination model is the attention degree of the crowd for the advertisement picture.
According to a second aspect, the present invention provides an artificial intelligence based picture display system comprising: the acquisition module is used for acquiring crowd shooting videos and advertisement pictures; the crowd determining module is used for determining average age of the crowd, crowd quantity, crowd flow speed, ambient light color and ambient light intensity by using a crowd determining model based on the crowd shooting video; the time length determining module is used for determining the playing time length of the advertisement picture by using a playing time length determining model based on the average age of the crowd, the crowd quantity, the crowd flow speed and the advertisement picture; the brightness determining module is used for determining the lowest display brightness of the advertisement picture based on the crowd quantity, the ambient light color, the ambient light intensity and the advertisement picture by using a brightness determining model; the position determining module is used for determining the display position of the advertisement picture on a screen based on the average age of the crowd, the crowd quantity, the crowd flow speed and screen information; and the display module is used for displaying the advertisement pictures based on the playing time of the advertisement pictures, the lowest display brightness of the advertisement pictures and the display positions of the advertisement pictures on a screen.
Furthermore, the crowd-determining model is a gating circulation unit model, the input of the crowd-determining model is that the crowd shoots a video, and the output of the crowd-determining model is that the average age of the crowd, the crowd quantity, the crowd flow speed, the ambient light color and the ambient light intensity.
Further, the playing time length determining model is a convolutional neural network model, the input of the playing time length determining model is the average age of the crowd, the crowd number, the crowd flow speed and the advertisement picture, and the output of the playing time length determining model is the playing time length of the advertisement picture.
Still further, the system is further configured to: acquiring action videos of people in the playing time of the advertisement pictures when the advertisement pictures are played; processing and determining the attention degree of the crowd to the advertisement picture by sequentially using a segmentation model, a person average watching duration determination model and an attention degree determination model based on the action video of the crowd in the playing duration of the advertisement picture; if the attention degree of the crowd to the advertisement picture is smaller than the attention degree threshold value, a new advertisement picture is replaced for playing; if the attention degree of the crowd to the advertisement picture is greater than an attention degree threshold value, repeatedly playing the advertisement picture, wherein the attention degree threshold value is determined by the following formula:
wherein->、/>、/>Is->Are all preset weights, are->For the average age of the population->For the number of people, ->For the flow rate of the population in question,for the playing time of the advertisement picture, < + >>Is a threshold of attention.
Further, the segmentation model, the average person viewing time length determination model and the attention degree determination model are all gating circulation unit models, the input of the segmentation model is an action video of a crowd in the playing time length of the advertisement picture, the output of the segmentation model is an action video of each person in the crowd in the playing time length of the advertisement picture, the input of the average person viewing time length determination model is an action video of each person in the playing time length of the advertisement picture, the output of the average person viewing time length determination model is an average person viewing time length of each person in the crowd for the advertisement, the input of the attention degree determination model is an action video of each person in the playing time length of the advertisement picture, the average person viewing time length of each person in the crowd for the advertisement, and the output of the attention degree determination model is the attention degree of the crowd for the advertisement picture.
The invention provides a picture display method and a system based on artificial intelligence, wherein the method comprises the steps of obtaining crowd shooting videos and advertisement pictures; determining average age of the crowd, the crowd quantity, the crowd flow speed, the ambient light color and the ambient light intensity by using a crowd determination model based on the crowd shooting video; determining the playing time length of the advertisement picture based on the average age of the crowd, the crowd quantity, the crowd flow speed and the playing time length determining model of the advertisement picture; determining the lowest display brightness of the advertisement picture by using a brightness determination model based on the crowd quantity, the ambient light color, the ambient light intensity and the advertisement picture; determining the display position of the advertisement picture on a screen based on the average age of the crowd, the crowd quantity, the crowd flow speed and screen information; based on the playing time of the advertisement picture, the lowest display brightness of the advertisement picture and the display position of the advertisement picture on a screen, the method can reduce the consumption of electric power energy caused by the display of the advertisement picture.
Drawings
Fig. 1 is a schematic flow chart of an artificial intelligence-based picture display method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of an artificial intelligence based picture display system according to an embodiment of the present invention;
fig. 3 is a schematic diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
In an embodiment of the present invention, an artificial intelligence based picture display method is provided as shown in fig. 1, where the artificial intelligence based picture display method includes steps S1 to S6:
step S1, obtaining crowd shooting videos and advertisement pictures.
The crowd shooting video is obtained by shooting the crowd beside the advertisement screen through a camera. The crowd shoots a plurality of passers-by including passers-by that pass the advertising screen in the video. As an example, the crowd may take video for 5 seconds, 10 seconds, 30 seconds, 1 minute, etc.
An advertisement picture refers to a static image used for advertising. The advertising pictures may contain text, graphics, logos, product or brand information, etc. for attracting the attention of the target audience and conveying relevant information. The advertising pictures may be in JPG format.
Crowd shooting videos provides crowd activity information in an actual scene, including crowd quantity, crowd age, crowd flow speed, ambient light color, ambient light intensity, etc., while advertising pictures provide content that needs to be displayed in the scene. In this way, the system can perform subsequent analysis and processing based on these input data, by combining crowd behavior and advertising pictures, to determine an optimal advertising picture presentation strategy.
And S2, determining the average age of the crowd, the crowd number, the crowd flow speed, the ambient light color and the ambient light intensity by using a crowd determination model based on the crowd shooting video.
The average age of a population refers to the average of the sum of the ages of all the individuals of the population appearing in the video taken by the population divided by the number of individuals in the population.
The crowd number refers to the individual number of the crowd appearing in the crowd-shot video. In some embodiments, it may be determined whether the population is greater than a population threshold, and if the population is greater than the population threshold, the management terminal may be notified to conduct population grooming.
Crowd flow speed represents the speed of average flow of crowd walking in crowd-shot video, for example, crowd flow speed may be 1.5m/s.
Ambient light color represents the light color in the external environment presented in the crowd-sourced video. The light color in the external environment can be green, red, white, etc.
The ambient light intensity is the received light intensity. For example, the ambient light intensity is 1 kaleidos (Lux).
The crowd determining model is a gating circulating unit model, the input of the crowd determining model is that the crowd shoots videos, and the output of the crowd determining model is that the average age of the crowd, the crowd quantity, the crowd flowing speed, the ambient light color and the ambient light intensity.
The gated loop unit model includes a gated loop unit (Gated Recurrent Unit, GRU), which is one way of artificial intelligence.
The gating cycle unit comprises a memory unit, an update gate and a reset gate.
The memory unit stores the information of the previous time step and transmits the information to the input and hidden state of the current time step so as to keep the history information.
The refresh gate determines whether to refresh the information of the memory cell. It controls the update degree of the memory unit according to the input of the current time step and the hidden state of the previous time step. When the update gate approaches 1, more past information will be retained; more new information will be passed on when the update gate approaches 0.
The reset gate determines how to use the memory cell of the previous time step. It decides which information to discard based on the input of the current time step and the hidden state of the previous time step. The reset gate functions to better accommodate the current input.
The gating circulation unit model effectively solves the problem of gradient disappearance through the adjustment of the update gate and the reset gate, so that the model can better process the dependency relationship of long sequence data.
The crowd shooting videos in continuous time periods are processed through the gating circulating unit model, the relation in the crowd shooting video time sequence can be captured better, and the characteristics of the association relation among the crowd shooting videos comprehensively considering all time points can be output, so that the output characteristics are more accurate and comprehensive.
In some embodiments, the gated loop cell model may be trained by a gradient descent method, and training is stopped when a preset stopping condition is reached. The preset stop condition may be for reaching a maximum training round or for the loss function to converge.
And step S3, determining the playing time of the advertisement picture by using a playing time determining model based on the average age of the crowd, the crowd quantity, the crowd flow speed and the advertisement picture.
The average age of the population, the number of the population, the flow rate of the population, the advertisement picture may all be used as inputs to determine a length of time for which the advertisement picture is played.
As an example, viewers of different ages vary in information acceptance and attention duration. Young people may prefer fast-paced, short duration advertising content, while elderly people may require more time to receive information. Thus, the average age of the population can affect the length of time that the advertising picture is played. The larger the average age of the crowd, the longer the required playing time of the advertisement picture.
The number of people may indirectly reflect the competing relationship between the viewers. If the audience is large, the visual information overload may cause distraction to the audience. In this case, a shorter advertisement playing time period can more attract and maintain the attention of the viewer and effectively deliver advertisement information. Conversely, when the number of people is small, a longer playing time period can provide more time for the audience to get in depth to the advertisement content.
The faster the crowd flows, the faster the passers-by passes the screen, and the longer the advertising pictures should be played accordingly, so as to convey advertising information as soon as possible within a limited period of time. Conversely, if the crowd flow rate is slower, the advertising picture may be selected for a longer length of play to more fully attract the viewer's attention and provide more information.
The content, creative form, information density and other factors of the advertising pictures also influence the playing time. If the content of the advertisement pictures is concise and clear, the interest of the audience can be quickly aroused and core information is transmitted, the playing time length can be shorter, and vice versa. Accordingly, an advertising picture may be used as an input to determine a length of time for which the advertising picture is played.
The playing time of the advertisement picture is the playing time of the advertisement on the screen.
The play duration determining model is a convolutional neural network model. The Convolutional Neural Network (CNN) may be a multi-layer neural network (e.g., comprising at least two layers). The at least two layers may include at least one of a convolutional layer (CONV), a modified linear unit (ReLU) layer, a pooling layer (POOL), or a fully-connected layer (FC). Convolutional neural networks, which are one way of artificial intelligence, are able to extract useful features from an image and gradually understand and learn the contextual information of the image. The input of the playing time length determining model is the average age of the crowd, the crowd quantity, the crowd flow speed and the advertisement picture, and the output of the playing time length determining model is the playing time length of the advertisement picture. The play duration determination model can comprehensively consider the average age of the crowd, the crowd quantity, the crowd flow speed and the advertisement pictures and output the play duration of the advertisement pictures to obtain proper advertisement pictures.
And S4, determining the lowest display brightness of the advertisement picture by using a brightness determination model based on the crowd quantity, the ambient light color, the ambient light intensity and the advertisement picture.
The brightness determination model is a convolutional neural network model. The input of the brightness determination model is the ambient light color, the ambient light intensity and the advertisement picture, and the output of the brightness determination model is the lowest display brightness of the advertisement picture.
The lowest display brightness of the advertisement picture is the lowest brightness which can achieve the advertisement display effect when the advertisement picture is displayed on the screen. The brightness determination model can be obtained by training a plurality of groups of training samples through a gradient descent method. The input of the training samples is the number of sample people, the color of sample ambient light, the illumination intensity of sample ambient light and the sample advertisement picture, and the output label of the training samples is the lowest display brightness of the sample advertisement picture. The multiple groups of training samples can be obtained through manual labeling of staff, in the manual labeling process, the staff can label the sample crowd quantity, the sample environment light color, the sample environment light intensity and the sample advertisement picture, and when the advertisement display effect is considered during labeling, the minimum display brightness of the advertisement picture can be reduced as much as possible so as to achieve the effect of saving energy. The lowest display brightness of the advertisement picture is reduced as much as possible during marking, so that the brightness determination model after training is completed can also consider the advertisement display effect during outputting and reduce the lowest display brightness of the advertisement picture as much as possible.
By determining the lowest display brightness of the advertisement picture, the consumption of electric power energy sources during the playing of the advertisement picture can be reduced.
The brightness determination model can determine the lowest display brightness of the advertisement picture after comprehensively considering the crowd quantity, the ambient light color, the ambient light intensity and the advertisement picture.
And S5, determining the display position of the advertisement picture on a screen based on the average age of the crowd, the crowd quantity, the crowd flow speed and screen information.
The screen information includes the size of the screen, screen resolution, screen refresh rate, screen reaction time. Meanwhile, screen information can be used for determining the display position of the advertisement picture on the screen.
The screen refresh rate refers to the number of refreshes per second of the screen, expressed in hertz (Hz). Typical refresh rates are 60Hz, 75Hz, 144Hz, etc. The higher the refresh rate, the higher the smoothness and response speed of the screen display.
The screen reaction time refers to the time required to display a pixel from black to white (or from one gray level to another). The shorter the reaction time, the less the blurring, smear, etc. will occur when the screen displays a moving picture.
The display position of the advertisement picture on the screen is the display position of the advertisement picture in the screen. For example, advertising pictures may be displayed at the top, bottom, left side, right side, middle, etc. of the screen. As an example, the advertisement picture may be completely displayed for the entire screen at a display position of the screen. Also as an example, an advertisement picture may be displayed on the top of the screen, while no advertisement is displayed elsewhere than on the top of the screen to save energy.
In some embodiments, an abscissa system may be established based on the screen, and the display position of the advertisement picture on the screen may be represented as an abscissa range and an ordinate range in the abscissa system. For example, the horizontal position range when the advertisement picture is displayed on the screen is indicated by the abscissa range, and the vertical position range when the advertisement picture is displayed on the screen is indicated by the ordinate range.
Because the screen is larger, if all the screen is used for displaying advertisements, a large amount of power is consumed, so that the advertisement pictures can be displayed only at the appointed position of the screen under the condition that the display effect of the advertisement pictures is ensured by determining the display position of the advertisement pictures, and the power energy is saved.
The average age of the crowd can be used for determining the display position of the advertisement picture on the screen. As an example, if the average age of the crowd is small, the advertisement picture may be displayed obliquely on the display position of the screen to attract the attention of young people.
The crowd count may be used to determine the display location of the advertising picture on the screen, as an example, if the crowd count is large, the advertising picture needs to be placed in a more obvious and easily noticeable location on the screen, and the location range of the advertising picture needs to be larger to attract the eyes of more people. Also as an example, if the number of people is small, the location range of the advertisement picture display may be small to save energy.
The crowd flow rate may be used to determine a display location of the advertising picture on a screen. As an example, if the crowd flows faster, the advertisement picture needs to be placed in the center of a screen that is more noticeable and noticeable, and the range of positions of the advertisement picture needs to be larger to attract the immediate attention of the passers-by. For another example, if the crowd flowing speed is slower, the advertisement picture can be selectively placed at a position almost higher than the crowd, so that the crowd does not need to raise the head all the time to see the advertisement picture, and the display position range of the advertisement picture can be smaller at the same time, so that energy sources are saved.
In some embodiments, the display position of the advertisement picture on the screen may be determined by a display position determination model. The display position determining model is a convolutional neural network model, the average age of the crowd, the crowd quantity, the crowd flow speed and screen information are input into the display position determining model, and the output of the display position determining model is the display position of the advertisement picture on a screen.
The display position determining model can comprehensively consider the average age of people, the number of people, the flow speed of people and screen information, reduces the consumption of electric power energy under the condition of ensuring the display effect of the advertisement pictures, and finally outputs and obtains the display position of the advertisement pictures on the screen.
And S6, displaying the advertisement picture based on the playing time of the advertisement picture, the lowest display brightness of the advertisement picture and the display position of the advertisement picture on a screen.
In some embodiments, the advertisement delivery can be performed by setting the playing time of the advertisement picture, the lowest display brightness of the advertisement picture and the display position of the advertisement picture on the screen.
In some embodiments, when the advertisement picture is played, the action video of the crowd in the playing time of the advertisement picture can be obtained, and the attention of the crowd to the advertisement picture is determined by processing the action video of the crowd in the playing time of the advertisement picture sequentially by using the segmentation model, the average person viewing time determining model and the attention determining model. The method comprises the steps that a segmentation model, a person-average viewing time length determination model and a concern degree determination model are all gating circulation unit models, the input of the segmentation model is an action video of a crowd in the playing time length of an advertisement picture, the output of the segmentation model is an action video of each person in the crowd in the playing time length of the advertisement picture, the input of the person-average viewing time length determination model is an action video of each person in the playing time length of the advertisement picture, the output of the person-average viewing time length determination model is a person-average viewing time length of each person in the crowd in the playing time length of the advertisement picture, the input of the concern degree determination model is an concern degree of each person in the crowd to the advertisement picture. The attention degree of the crowd to the advertisement picture can be a numerical value between 0 and 1, and the larger the numerical value is, the higher the attention degree of the crowd to the advertisement picture is. In some embodiments, if the attention degree of the crowd to the advertisement picture is smaller than the attention degree threshold, the new advertisement picture is updated to be played, and if the attention degree of the crowd to the advertisement picture is greater than the attention degree threshold, the advertisement picture is repeatedly played.
In some embodiments, the attention threshold may be manually set by the user in advance.
In some embodiments, the attention threshold may be determined by the following formula:
wherein->、/>、/>Is->Are all the weights of the preset weight, and the weight of the whole body is equal to the preset weight,for the average age of the population->For the number of people, ->For the crowd flow rate,/->For the playing time of the advertisement picture, < + >>Is a threshold of attention.
The per-person viewing duration of each person for an advertisement may reflect the attention of one person to the advertisement content. As an example, if a person views for a longer period of time during the playing of an advertisement picture, this generally means that interest in the advertisement content and deep observation are generated, the higher the attention is, and vice versa.
The action video of each person in the playing time of the advertisement picture can reflect the attention degree of one person to the advertisement content. The action video may provide information on a person's behavior and engagement. As an example, if a person is keeping looking at an advertisement picture and the hand is in a scratch, the user may be considered to show greater engagement and interest in the advertisement content, with a higher degree of attention. Conversely, if a person does not watch the advertisement within the playing time of the advertisement picture, the user is not particularly concerned with the advertisement.
The segmentation model is responsible for extracting the space information of the action video, the average person viewing time length determination model is responsible for determining the viewing time length of each person, and the attention degree determination model comprehensively considers a plurality of factors to determine the attention degree of the crowd to the advertisement. This multi-model cooperative approach can increase computational efficiency and processing speed, and reduce overall system complexity, providing more accurate results.
The method has the advantages that through the segmentation model, the average person viewing time length determination model and the attention degree determination model, the information such as the action video of the crowd in the advertisement picture playing time length and the average person viewing time length of each person to the advertisement is combined, the attention degree of each person and the whole crowd to the advertisement can be more accurately determined, and the system can intelligently make advertisement playing decisions. When the attention of the crowd to the advertisement is low, the crowd can switch to a new advertisement picture in time, so that the resource waste is avoided and the advertisement effect is improved. And when the attention of the crowd to the advertisement is higher, the advertisement picture can be repeatedly played, so that the advertisement display effect is further improved.
In some embodiments, the average viewing time length determination model includes an action extraction layer, an action consistency determination layer, and a viewing time length determination layer. The input of the action extraction layer is an action video of each person in the crowd within the playing time of the advertisement picture, the output of the action extraction layer is a hand action video of each person in the crowd, an eye action video of each person in the crowd and a head action video of each person, the input of the action consistency judgment layer is a hand action video of each person in the crowd, an eye action video of each person in the crowd and a head action video of each person, the output of the action consistency judgment layer is an action consistency degree of each person in the crowd, the input of the watching time length determination layer is an action consistency degree of each person, an eye action video of each person in the crowd, a head action video of each person and an action consistency degree of each person, and the output of the watching time length determination layer is a watching time length of each person in the crowd for the advertisement. The action extraction layer, the action consistency judging layer and the watching duration determining layer all adopt a gating circulating unit (Gated Recurrent Unit, GRU) structure. The action extraction layer outputs hand action video, eye action video, head action video and the like of each person in the crowd, and the action extraction layer is used for extracting action information of each person in the crowd. The action consistency judging layer outputs the action consistency degree of each person in the crowd, and the action of the layer judges whether the action of each person in the crowd is consistent or not, so that a reference is provided for predicting the watching duration. The output of the viewing time length determining layer is the person-average viewing time length of each person in the crowd for the advertisement, and the effect of the layer is to calculate the person-average viewing time length of each person for the advertisement according to the action information and the action consistency degree of each person in the crowd.
When a person views an advertisement, the person usually has corresponding action performances such as eye gaze, hand movements, head turning and the like. When a person is interested or invested in an advertisement, their actions tend to be relatively consistent and consistent. Thus, the degree of action consistency of each person in the crowd may reflect the degree of concentration and engagement of the advertisement, and thus be related to the viewing duration. When a person is interested in advertisements, their attention will be focused more and the performance will be more consistent. Conversely, if a person is not interested or unrelated in an advertisement, their attention and performance may be more abrupt. Therefore, the action consistency degree of each person in the crowd can indirectly reflect the experience consistency of the person when watching advertisements, and further is related to the watching duration.
Based on the same inventive concept, fig. 2 is a schematic diagram of an artificial intelligence based picture display system according to an embodiment of the present invention, where the artificial intelligence based picture display system includes:
an acquisition module 21, configured to acquire crowd-captured videos and advertisement pictures;
a crowd determining module 22, configured to determine an average age of the crowd, a number of the crowd, a flow rate of the crowd, an ambient light color, and an ambient light intensity using a crowd determining model based on the crowd-captured video;
a duration determining module 23, configured to determine a playing duration of the advertisement picture based on the average age of the crowd, the crowd number, the crowd flow speed, and the advertisement picture using a playing duration determining model;
a brightness determination module 24, configured to determine a lowest display brightness of the advertisement picture based on the crowd amount, the ambient light color, the ambient light intensity, and the advertisement picture using a brightness determination model;
a position determining module 25, configured to determine a display position of the advertisement picture on a screen based on the average age of the crowd, the crowd number, the crowd flow speed, and screen information;
and the display module 26 is configured to display the advertisement picture based on a playing duration of the advertisement picture, a minimum display brightness of the advertisement picture, and a display position of the advertisement picture on a screen.
Based on the same inventive concept, an embodiment of the present invention provides an electronic device, as shown in fig. 3, including:
comprising the following steps: a processor 31; a memory 32; a computer program; wherein the computer program is stored in the memory 32 and configured to be executed by the processor 31 to implement an artificial intelligence based picture display method as provided above, the method comprising: acquiring crowd shooting videos and advertisement pictures; determining average age of the crowd, the crowd quantity, the crowd flow speed, the ambient light color and the ambient light intensity by using a crowd determination model based on the crowd shooting video; determining the playing time length of the advertisement picture based on the average age of the crowd, the crowd quantity, the crowd flow speed and the playing time length determining model of the advertisement picture; determining the lowest display brightness of the advertisement picture by using a brightness determination model based on the crowd quantity, the ambient light color, the ambient light intensity and the advertisement picture; determining the display position of the advertisement picture on a screen based on the average age of the crowd, the crowd quantity, the crowd flow speed and screen information; and displaying the advertisement picture based on the playing time of the advertisement picture, the lowest display brightness of the advertisement picture and the display position of the advertisement picture on a screen.
Based on the same inventive concept, the present embodiment provides a computer readable storage medium having stored thereon a computer program which when executed by a processor 31 implements the artificial intelligence based picture display method provided above, the method comprising obtaining a crowd taken video and an advertisement picture; determining average age of the crowd, the crowd quantity, the crowd flow speed, the ambient light color and the ambient light intensity by using a crowd determination model based on the crowd shooting video; determining the playing time length of the advertisement picture based on the average age of the crowd, the crowd quantity, the crowd flow speed and the playing time length determining model of the advertisement picture; determining the lowest display brightness of the advertisement picture by using a brightness determination model based on the crowd quantity, the ambient light color, the ambient light intensity and the advertisement picture; determining the display position of the advertisement picture on a screen based on the average age of the crowd, the crowd quantity, the crowd flow speed and screen information; and displaying the advertisement picture based on the playing time of the advertisement picture, the lowest display brightness of the advertisement picture and the display position of the advertisement picture on a screen.
Finally, it should be understood that the embodiments described in this specification are merely illustrative of the principles of the embodiments of this specification. Other variations are possible within the scope of this description. Thus, by way of example, and not limitation, alternative configurations of embodiments of the present specification may be considered as consistent with the teachings of the present specification. Accordingly, the embodiments of the present specification are not limited to only the embodiments explicitly described and depicted in the present specification.
Claims (10)
1. An artificial intelligence based picture display method, comprising:
acquiring crowd shooting videos and advertisement pictures;
determining average age of the crowd, the crowd quantity, the crowd flow speed, the ambient light color and the ambient light intensity by using a crowd determination model based on the crowd shooting video;
determining the playing time length of the advertisement picture based on the average age of the crowd, the crowd quantity, the crowd flow speed and the playing time length determining model of the advertisement picture;
determining the lowest display brightness of the advertisement picture by using a brightness determination model based on the crowd quantity, the ambient light color, the ambient light intensity and the advertisement picture;
determining the display position of the advertisement picture on a screen based on the average age of the crowd, the crowd quantity, the crowd flow speed and screen information;
and displaying the advertisement picture based on the playing time of the advertisement picture, the lowest display brightness of the advertisement picture and the display position of the advertisement picture on a screen.
2. The artificial intelligence based picture display method of claim 1, wherein the crowd-sourced model is a gated loop unit model, the crowd-sourced model is input to capture video for the crowd, and the crowd-sourced model is output to the crowd average age, the crowd number, the crowd flow rate, the ambient light color, the ambient light intensity.
3. The artificial intelligence based picture display method according to claim 1, wherein the play duration determination model is a convolutional neural network model, the input of the play duration determination model is the average age of the population, the number of the population, the flow speed of the population, the advertisement picture, and the output of the play duration determination model is the play duration of the advertisement picture.
4. The artificial intelligence based picture display method according to claim 1, wherein the method further comprises:
acquiring action videos of people in the playing time of the advertisement pictures when the advertisement pictures are played;
processing and determining the attention degree of the crowd to the advertisement picture by sequentially using a segmentation model, a person average watching duration determination model and an attention degree determination model based on the action video of the crowd in the playing duration of the advertisement picture;
if the attention degree of the crowd to the advertisement picture is smaller than the attention degree threshold value, a new advertisement picture is replaced for playing;
if the attention degree of the crowd to the advertisement picture is greater than an attention degree threshold value, repeatedly playing the advertisement picture, wherein the attention degree threshold value is determined by the following formula:
wherein->、/>、/>Is->Are all preset weights, are->For the average age of the population->For the number of people, ->For the crowd flow rate,/->For the playing time of the advertisement picture, < + >>Is a threshold of attention.
5. The method for displaying pictures based on artificial intelligence according to claim 4, wherein the segmentation model, the average person viewing time length determination model and the attention degree determination model are all gating circulation unit models, the input of the segmentation model is an action video of a crowd in the playing time length of the advertisement picture, the output of the segmentation model is an action video of each person in the crowd in the playing time length of the advertisement picture, the input of the average person viewing time length determination model is an action video of each person in the playing time length of the advertisement picture, the output of the average person viewing time length determination model is an average person viewing time length of each person in the crowd in the playing time length of the advertisement picture, the input of the attention degree determination model is an action video of each person in the playing time length of the advertisement picture, and the output of the attention degree determination model is the attention degree of the crowd to the advertisement picture.
6. An artificial intelligence based picture display system, comprising:
the acquisition module is used for acquiring crowd shooting videos and advertisement pictures;
the crowd determining module is used for determining average age of the crowd, crowd quantity, crowd flow speed, ambient light color and ambient light intensity by using a crowd determining model based on the crowd shooting video;
the time length determining module is used for determining the playing time length of the advertisement picture by using a playing time length determining model based on the average age of the crowd, the crowd quantity, the crowd flow speed and the advertisement picture;
the brightness determining module is used for determining the lowest display brightness of the advertisement picture based on the crowd quantity, the ambient light color, the ambient light intensity and the advertisement picture by using a brightness determining model;
the position determining module is used for determining the display position of the advertisement picture on a screen based on the average age of the crowd, the crowd quantity, the crowd flow speed and screen information;
and the display module is used for displaying the advertisement pictures based on the playing time of the advertisement pictures, the lowest display brightness of the advertisement pictures and the display positions of the advertisement pictures on a screen.
7. The artificial intelligence based picture display system of claim 6, wherein the crowd-sourced model is a gated loop unit model, wherein the crowd-sourced model is input to take video of the crowd, and wherein the crowd-sourced model is output to the crowd average age, the crowd quantity, the crowd flow rate, the ambient light color, the ambient light intensity.
8. The artificial intelligence based picture display system of claim 6, wherein the playback time length determination model is a convolutional neural network model, wherein the inputs of the playback time length determination model are the average age of the population, the number of the population, the flow rate of the population, the advertisement picture, and wherein the output of the playback time length determination model is the playback time length of the advertisement picture.
9. The artificial intelligence based picture display system according to claim 6, wherein the system is further configured to:
acquiring action videos of people in the playing time of the advertisement pictures when the advertisement pictures are played;
processing and determining the attention degree of the crowd to the advertisement picture by sequentially using a segmentation model, a person average watching duration determination model and an attention degree determination model based on the action video of the crowd in the playing duration of the advertisement picture;
if the attention degree of the crowd to the advertisement picture is smaller than the attention degree threshold value, a new advertisement picture is replaced for playing;
if the attention degree of the crowd to the advertisement picture is greater than an attention degree threshold value, repeatedly playing the advertisement picture, wherein the attention degree threshold value is determined by the following formula:
wherein->、/>、/>Is->Are all preset weights, are->For the average age of the population->For the number of people, ->For the crowd flow rate,/->For the playing time of the advertisement picture, < + >>Is a threshold of attention.
10. The artificial intelligence based picture display system of claim 9, wherein the segmentation model, the people-average viewing time length determination model and the attention degree determination model are all gating circulation unit models, the input of the segmentation model is an action video of a crowd in the playing time length of the advertisement picture, the output of the segmentation model is an action video of each person in the crowd in the playing time length of the advertisement picture, the input of the people-average viewing time length determination model is an action video of each person in the playing time length of the advertisement picture, the output of the people-average viewing time length determination model is an people-average viewing time length of each person in the crowd for the advertisement, the input of the attention degree determination model is an action video of each person in the playing time length of the advertisement picture, and the output of the attention degree determination model is the attention degree of the crowd for the advertisement picture.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311714899.9A CN117408760B (en) | 2023-12-14 | 2023-12-14 | Picture display method and system based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311714899.9A CN117408760B (en) | 2023-12-14 | 2023-12-14 | Picture display method and system based on artificial intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117408760A true CN117408760A (en) | 2024-01-16 |
| CN117408760B CN117408760B (en) | 2024-02-27 |
Family
ID=89496576
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311714899.9A Active CN117408760B (en) | 2023-12-14 | 2023-12-14 | Picture display method and system based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117408760B (en) |
Citations (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040194128A1 (en) * | 2003-03-28 | 2004-09-30 | Eastman Kodak Company | Method for providing digital cinema content based upon audience metrics |
| US20040194127A1 (en) * | 2003-03-28 | 2004-09-30 | Eastman Kodak Company | Method and system for modifying digital cinema frame content |
| US20110251896A1 (en) * | 2010-04-09 | 2011-10-13 | Affine Systems, Inc. | Systems and methods for matching an advertisement to a video |
| CN106241584A (en) * | 2016-08-23 | 2016-12-21 | 西尼电梯(杭州)有限公司 | A kind of intelligent video monitoring system based on staircase safety and method |
| CN107146096A (en) * | 2017-03-07 | 2017-09-08 | 浙江工业大学 | A kind of intelligent video advertisement methods of exhibiting and device |
| CN107371042A (en) * | 2017-08-31 | 2017-11-21 | 深圳创维－Rgb电子有限公司 | Advertisement placement method, device, equipment and storage medium |
| CN107590446A (en) * | 2017-08-28 | 2018-01-16 | 北京工业大学 | The system and implementation method of Intelligent Measurement crowd's attention rate |
| CN109635721A (en) * | 2018-12-10 | 2019-04-16 | 山东大学 | Video human fall detection method and system based on track weighting depth convolution sequence poolization description |
| CN109934625A (en) * | 2019-03-01 | 2019-06-25 | 统云信息科技有限公司 | A kind of artificial intelligence display screen advertisement dynamic throwing system and method |
| CN111242683A (en) * | 2020-01-10 | 2020-06-05 | 深圳市胜博恩科技有限公司 | Method and device for self-service delivery of advertisement content, storage medium and computer equipment |
| CN111736510A (en) * | 2020-07-17 | 2020-10-02 | 詹晨 | Outdoor multimedia advertisement intelligent regulation and control management system based on big data |
| WO2022023831A1 (en) * | 2020-07-29 | 2022-02-03 | Akhavan Mahdi | Smart display application with potential to exhibit collected outdoor information content using iot and ai platforms |
| CN114387038A (en) * | 2022-01-20 | 2022-04-22 | 郭一帆 | Billboard system combining artificial intelligence and computer vision |
| CN114973954A (en) * | 2022-05-28 | 2022-08-30 | 深圳市卓越鸿业信息技术有限公司 | Intelligent advertisement putting platform and method thereof |
| CN115204920A (en) * | 2021-04-09 | 2022-10-18 | 上海眼控科技股份有限公司 | Advertisement putting method, system and computer readable storage medium |
| CN115601062A (en) * | 2022-10-27 | 2023-01-13 | 临沂职业学院（Cn） | Intelligent advertisement display device, control method thereof and intelligent advertisement display device |
| CN116664198A (en) * | 2023-04-28 | 2023-08-29 | 浙江新再灵科技股份有限公司 | Elevator intelligent screen remote control management method based on Internet of things |
| CN117219003A (en) * | 2023-11-09 | 2023-12-12 | 深圳市东陆科技有限公司 | Content display method and device of LED display module |
-
2023
- 2023-12-14 CN CN202311714899.9A patent/CN117408760B/en active Active
Patent Citations (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040194128A1 (en) * | 2003-03-28 | 2004-09-30 | Eastman Kodak Company | Method for providing digital cinema content based upon audience metrics |
| US20040194127A1 (en) * | 2003-03-28 | 2004-09-30 | Eastman Kodak Company | Method and system for modifying digital cinema frame content |
| US20110251896A1 (en) * | 2010-04-09 | 2011-10-13 | Affine Systems, Inc. | Systems and methods for matching an advertisement to a video |
| CN106241584A (en) * | 2016-08-23 | 2016-12-21 | 西尼电梯(杭州)有限公司 | A kind of intelligent video monitoring system based on staircase safety and method |
| CN107146096A (en) * | 2017-03-07 | 2017-09-08 | 浙江工业大学 | A kind of intelligent video advertisement methods of exhibiting and device |
| CN107590446A (en) * | 2017-08-28 | 2018-01-16 | 北京工业大学 | The system and implementation method of Intelligent Measurement crowd's attention rate |
| CN107371042A (en) * | 2017-08-31 | 2017-11-21 | 深圳创维－Rgb电子有限公司 | Advertisement placement method, device, equipment and storage medium |
| CN109635721A (en) * | 2018-12-10 | 2019-04-16 | 山东大学 | Video human fall detection method and system based on track weighting depth convolution sequence poolization description |
| CN109934625A (en) * | 2019-03-01 | 2019-06-25 | 统云信息科技有限公司 | A kind of artificial intelligence display screen advertisement dynamic throwing system and method |
| CN111242683A (en) * | 2020-01-10 | 2020-06-05 | 深圳市胜博恩科技有限公司 | Method and device for self-service delivery of advertisement content, storage medium and computer equipment |
| CN111736510A (en) * | 2020-07-17 | 2020-10-02 | 詹晨 | Outdoor multimedia advertisement intelligent regulation and control management system based on big data |
| WO2022023831A1 (en) * | 2020-07-29 | 2022-02-03 | Akhavan Mahdi | Smart display application with potential to exhibit collected outdoor information content using iot and ai platforms |
| CN115204920A (en) * | 2021-04-09 | 2022-10-18 | 上海眼控科技股份有限公司 | Advertisement putting method, system and computer readable storage medium |
| CN114387038A (en) * | 2022-01-20 | 2022-04-22 | 郭一帆 | Billboard system combining artificial intelligence and computer vision |
| CN114973954A (en) * | 2022-05-28 | 2022-08-30 | 深圳市卓越鸿业信息技术有限公司 | Intelligent advertisement putting platform and method thereof |
| CN115601062A (en) * | 2022-10-27 | 2023-01-13 | 临沂职业学院（Cn） | Intelligent advertisement display device, control method thereof and intelligent advertisement display device |
| CN116664198A (en) * | 2023-04-28 | 2023-08-29 | 浙江新再灵科技股份有限公司 | Elevator intelligent screen remote control management method based on Internet of things |
| CN117219003A (en) * | 2023-11-09 | 2023-12-12 | 深圳市东陆科技有限公司 | Content display method and device of LED display module |
Non-Patent Citations (2)
| Title |
|---|
| 楼小龙: "面向大屏显示的体感交互效率评估与优化方法研究", 《中国博士学位论文全文数据库工程科技Ⅱ辑》, no. 12, 15 December 2018 (2018-12-15), pages 028 - 4 * |
| 陈文花: "基于深度学习的图片广告点击率预估研究", 《中国优秀硕士学位论文全文数据库信息科技辑》, no. 01, 15 January 2020 (2020-01-15), pages 140 - 332 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117408760B (en) | 2024-02-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10911829B2 (en) | Vehicle video recommendation via affect | |
| CN111416950B (en) | Video processing method and device, storage medium and electronic equipment | |
| US10289898B2 (en) | Video recommendation via affect | |
| US9503786B2 (en) | Video recommendation using affect | |
| Gao et al. | Fast video multi-style transfer | |
| US20160191995A1 (en) | Image analysis for attendance query evaluation | |
| US20150206000A1 (en) | Background analysis of mental state expressions | |
| US20170171614A1 (en) | Analytics for livestreaming based on image analysis within a shared digital environment | |
| US9271035B2 (en) | Detecting key roles and their relationships from video | |
| CN110390048A (en) | Information-pushing method, device, equipment and storage medium based on big data analysis | |
| US20200026347A1 (en) | Multidevice multimodal emotion services monitoring | |
| US9195880B1 (en) | Interactive viewer for image stacks | |
| US20180192160A1 (en) | Context based augmented advertisement | |
| McDuff et al. | Crowdsourced data collection of facial responses | |
| CN112997477A (en) | Display device and display control method | |
| CN114821734A (en) | Method and device for driving expression of virtual character | |
| US11847818B2 (en) | Method for extracting video clip, device for extracting video clip, and storage medium | |
| WO2021042364A1 (en) | Method and device for taking picture | |
| WO2021103610A1 (en) | Display object control method and apparatus, electronic device and storage medium | |
| Yeh et al. | Visual-attention-based pixel dimming technique for OLED displays of mobile devices | |
| Feng et al. | Exploring CNN-based viewport prediction for live virtual reality streaming | |
| CN117408760B (en) | Picture display method and system based on artificial intelligence | |
| US10209523B1 (en) | Apparatus, system, and method for blur reduction for head-mounted displays | |
| CN116013184A (en) | Display screen processing method, device, equipment and medium | |
| CN116229311A (en) | Video processing method, device and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20240410 Address after: No.112 Yuanwang street, industrial development community, Ranghulu District, Daqing City, Heilongjiang Province Patentee after: Daqing suofelectronic Technology Development Co.,Ltd. Country or region after: China Address before: Room 102-940, 1st Floor, Unit 1, Building 11, No. 88 Niushahou Street, Jinjiang District, Chengdu City, Sichuan Province, 610000 Patentee before: Chengdu Yadu Kesheng Technology Co.,Ltd. Country or region before: China |
|
| TR01 | Transfer of patent right |