CN112950242A - Information pushing method and device and electronic equipment - Google Patents
Information pushing method and device and electronic equipment Download PDFInfo
- Publication number
- CN112950242A CN112950242A CN201911257893.7A CN201911257893A CN112950242A CN 112950242 A CN112950242 A CN 112950242A CN 201911257893 A CN201911257893 A CN 201911257893A CN 112950242 A CN112950242 A CN 112950242A
- Authority
- CN
- China
- Prior art keywords
- person
- target person
- detected
- matrix
- item
- 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 56
- 239000011159 matrix material Substances 0.000 claims abstract description 143
- 230000014759 maintenance of location Effects 0.000 claims abstract description 22
- 238000012545 processing Methods 0.000 claims description 18
- 238000004590 computer program Methods 0.000 claims description 14
- 230000005484 gravity Effects 0.000 claims description 12
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000012163 sequencing technique Methods 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 3
- 238000004422 calculation algorithm Methods 0.000 description 17
- 238000010586 diagram Methods 0.000 description 13
- 238000007726 management method Methods 0.000 description 8
- 230000009286 beneficial effect Effects 0.000 description 5
- 238000001914 filtration Methods 0.000 description 5
- 239000000284 extract Substances 0.000 description 4
- 239000002699 waste material Substances 0.000 description 4
- 230000006399 behavior Effects 0.000 description 3
- 238000010276 construction Methods 0.000 description 3
- 238000007418 data mining Methods 0.000 description 3
- 230000002349 favourable effect Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 210000004709 eyebrow Anatomy 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 241000501105 Aeshnidae Species 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000001174 ascending effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000002716 delivery method Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 210000000887 face Anatomy 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 210000003128 head Anatomy 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000003825 pressing Methods 0.000 description 1
- 230000002035 prolonged effect Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 239000000725 suspension Substances 0.000 description 1
Images
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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Business, Economics & Management (AREA)
- Development Economics (AREA)
- Accounting & Taxation (AREA)
- Strategic Management (AREA)
- Multimedia (AREA)
- Finance (AREA)
- Game Theory and Decision Science (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Human Computer Interaction (AREA)
- Economics (AREA)
- Health & Medical Sciences (AREA)
- General Business, Economics & Management (AREA)
- General Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of Internet, and provides an information pushing method, an information pushing device and electronic equipment, wherein the method comprises the following steps: acquiring continuous frame images of a person to be detected, wherein the continuous frame images comprise face images of the person to be detected; selecting a person to be detected with the retention time of the face image reaching a preset time threshold as a target person, and extracting attribute information of the target person to construct a personal file; calculating an item preference matrix of the target person based on the attribute information in the personal profile; and pushing according to the recommendation list corresponding to the item preference matrix matching of the target person. The embodiment of the invention can improve the accuracy of advertisement pushing and realize accurate advertisement delivery.
Description
Technical Field
The invention relates to the technical field of internet, in particular to an information pushing method and device and electronic equipment.
Background
At present, the face recognition and human body structuring technology is widely applied to the fields of security and new retail, and the social security and the commercial analysis efficiency are greatly improved. At present, in the field of media propagation, screen advertisements and projection advertisements are widely applied to elevators, and a good flow conversion effect is achieved for merchants. However, for advertisement media, in the prior art, a common advertisement delivery method is random and wide delivery, so that the advertisement delivery has no accurate crowd direction, it is difficult to accurately find a user group, and a part of advertisement resources are wasted. Therefore, the problem of low advertisement pushing accuracy exists in the prior art.
Disclosure of Invention
The embodiment of the invention provides an information pushing method which can improve the accuracy of advertisement pushing, realize accurate advertisement delivery and reduce resource waste.
In a first aspect, an embodiment of the present invention provides an information pushing method, including the following steps:
acquiring continuous frame images of a person to be detected, wherein the continuous frame images comprise face images of the person to be detected;
selecting a person to be detected with the retention time of the face image reaching a preset time threshold as a target person, and extracting attribute information of the target person to construct a personal file;
calculating an item preference matrix of the target person based on the attribute information in the personal profile;
and pushing according to the recommendation list corresponding to the item preference matrix matching of the target person.
In a second aspect, an embodiment of the present invention further provides an information pushing apparatus, including:
the system comprises an acquisition module, a processing module and a display module, wherein the acquisition module is used for acquiring continuous frame images of a person to be detected, and the continuous frame images comprise face images of the person to be detected;
the extraction module is used for selecting the person to be detected with the retention time of the face image reaching a preset time threshold value as a target person and extracting attribute information of the target person to construct a personal file;
the calculation module is used for calculating an item preference matrix of the target person based on the attribute information in the personal file;
and the pushing module is used for pushing according to the recommendation list corresponding to the item preference matrix matching of the target person.
In a third aspect, an embodiment of the present invention further provides an electronic device, including: the information push method comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor executes the computer program to realize the steps in the information push method provided by the embodiment of the invention.
In a fourth aspect, the embodiment of the present invention further provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the computer program implements the steps in the information pushing method provided by the embodiment of the present invention.
In the embodiment of the invention, continuous frame images of a person to be detected are collected, wherein the continuous frame images comprise face images of the person to be detected; selecting a person to be detected with the retention time of the face image reaching a preset time threshold as a target person, and extracting attribute information of the target person to construct a personal file; calculating an item preference matrix of the target person based on the attribute information in the personal profile; and pushing according to the recommendation list corresponding to the item preference matrix matching of the target person. According to the embodiment of the invention, the stay time of each person is analyzed, the person to be detected with the stay time reaching the preset time threshold is taken as the target person, the person to be detected with the stay time not reaching the time threshold does not need to be matched with the recommendation list, and resource waste caused by random delivery is avoided; when the target person is selected, the attribute information of the target person is extracted to construct an article preference matrix of the target person, and a recommendation list corresponding to the article preference matrix is matched for recommendation, so that the accuracy of advertisement pushing is improved, and accurate advertisement delivery is realized.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a system architecture diagram of an information pushing method provided in an embodiment of the present invention;
fig. 2 is a flowchart of an information pushing method provided by an embodiment of the present invention;
fig. 3 is a flowchart of another information pushing method provided by an embodiment of the present invention;
fig. 4 is a flowchart of another information pushing method provided by an embodiment of the present invention;
fig. 5 is a flowchart of another information pushing method provided by an embodiment of the present invention;
FIG. 6a is a detailed diagram of an algorithm for an item preference matrix according to an embodiment of the present invention;
FIG. 6b is a detailed diagram of an algorithm for an alternative item preference matrix according to an embodiment of the present invention;
fig. 7 is a flowchart of another information pushing method provided by an embodiment of the present invention;
FIG. 8 is a schematic structural diagram of an information pushing apparatus according to an embodiment of the present invention;
fig. 9 is a schematic structural diagram of another information pushing apparatus provided in the embodiment of the present invention;
fig. 10 is a schematic structural diagram of another information pushing apparatus provided in the embodiment of the present invention;
fig. 11 is a schematic structural diagram of another information pushing apparatus provided in the embodiment of the present invention;
fig. 12 is a schematic structural diagram of another information pushing apparatus provided in the embodiment of the present invention;
fig. 13 is a schematic structural diagram of another information pushing apparatus provided in the embodiment of the present invention;
fig. 14 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Fig. 1 is a system architecture diagram 1001 for pushing information according to an embodiment of the present invention. The system can provide a framework of an operation scene for the information pushing method provided by the embodiment of the invention. The system architecture 1001 includes a projection device 1002, a server 1003, and a backend management device 1004. The projection device 1002 includes a camera module 1002a and a projection module 1002 b.
The camera module 1002a, the projection module 1002b, the server 1003 and the background management device 1004 are connected to each other through the internet to transmit data.
In the system, a plurality of camera modules 1002a can be arranged for collecting multi-angle images of the person to be detected so as to obtain corresponding retention time; the method can also be used for collecting multi-angle images of the target personnel to acquire the attribute information of the target personnel.
The plurality of projection modules 1002b may also be arranged for performing multi-advertisement projection according to the recommendation list transmitted to the projection module 1002b by the background management device 1004, so as to realize multi-directional and comprehensive advertisement showing.
The server 1003 may be configured to store attribute information of a target person, identify the attribute information of the target person collected by the camera module 1002a through a video structuring engine and a human body recognition algorithm, construct a user preference matrix after performing structuring processing on the identified attribute information, and store the user preference matrix in the server 1003. The system performs data mining on the user preference matrix stored in the server 1003, performs data analysis on the user preference matrix, and matches a recommendation list corresponding to the user preference matrix in combination with a recommendation algorithm.
The background management device 1004 may be a mobile terminal, and is configured to sort the list of the legged nails after performing similarity calculation between the recommendation list obtained through data mining and the user preference matrix, and send the sorted recommendation list to the projection module 1002b through a network for delivery.
The mobile terminal 100 may be an electronic device having a display screen and capable of reading, receiving and the like data transmitted by communication, and the mobile terminal 100 includes, but is not limited to, a smart phone, a tablet computer, a laptop portable computer, a desktop computer and the like. The mobile terminal can also be called a terminal, a user side, an intelligent terminal and the like.
Specifically, after the camera module 1002a collects the continuous frame images in the elevator, the continuous frame images are transmitted to the server 1003 through network transmission, the continuous frame images are analyzed in the server 1003 through a video structured engine and a human body recognition algorithm, target personnel are extracted, attribute information of the target personnel is extracted, and a user preference matrix is constructed based on the attribute information; the user preference matrix is subjected to data analysis, recommendation lists corresponding to the user preference matrix are matched based on a recommendation algorithm, the recommendation lists are sent to the background management device 1004 to be sorted, an ordered recommendation list is obtained, and the background management device sends the ordered recommendation list to the projection module 1002b to be released.
It should be understood that the numbers of the projection devices, the servers, the background management devices, and the network described above are merely illustrative, and may be specifically adjusted according to implementation needs. The embodiment of the invention can be applied to elevators and can also be applied to scenes such as large supermarkets, shopping malls and the like for pushing advertisements.
As shown in fig. 2, fig. 2 is a flowchart of an information pushing method provided in an embodiment of the present invention, which specifically includes the following steps:
201. and acquiring continuous frame images of the person to be detected, wherein the continuous frame images comprise face images of the person to be detected.
In this embodiment, the electronic device on which the information push method operates may acquire the continuous frame images and the like through a wired connection manner or a wireless connection manner. It should be noted that the Wireless connection manner may include, but is not limited to, a 3G/4G connection, a WiFi (Wireless-Fidelity) connection, a bluetooth connection, a wimax (worldwide Interoperability for Microwave access) connection, a Zigbee (low power local area network protocol), a uwb (ultra wideband) connection, and other Wireless connection manners known now or developed in the future. The embodiment of the invention can be applied to an advertisement putting scene of an elevator, the continuous frame images can be collected through image collecting equipment, and in the embodiment, the camera is adopted for image collection.
The person to be detected may represent a plurality of persons entering the elevator, and the person to be detected may include people of various ages, for example: the detector comprises a detector, a signal processing module and a signal processing module.
The continuous frame images may be images acquired by a camera in a certain period of time in the elevator, and the continuous frame images may include face images, human body images, and the like of the person to be detected, for example: the method comprises the steps of obtaining human faces, human body images and the like of a person to be detected in the elevator running from 5 pm to 5 pm for 5 minutes. A plurality of persons to be detected can be distinguished by collecting face images of the persons to be detected, and information corresponding to each person to be detected is convenient to acquire.
202. And selecting the person to be detected with the retention time of the face image reaching a preset time threshold as a target person, and extracting attribute information of the target person to construct a personal file.
The stay time represents the total stay time of the person to be detected corresponding to the face image from the elevator to the elevator in the continuous frame images, and the stay time can include the waiting time of temporary suspension of the elevator in the midway, such as: when 2 persons enter the elevator at the same time, a person A goes to a floor 4, and a person B goes to a floor 6, the waiting time of the person A going out of the elevator and then the ascending time is continuously started in the floor 4 is temporarily stopped, and the waiting time of the person B is represented; if there are more than 2 persons going to different floors, the waiting time of some persons may include a plurality of waiting times, so that the waiting time is prolonged.
The preset time threshold may be a showing time of the advertisement on the corresponding floor. The comparison of the stay time with the time threshold may be to determine whether the stay time of the person to be detected is greater than or equal to the time threshold, which indicates whether the person to be detected can see the pushed advertisement within the stay time, and the advertisement showing time is confirmed when the advertisement is produced. Of course, the time threshold may also be a manually preset time, such as: the fixed time threshold is 30s, 40s, etc.
When the stay time of the person to be detected meets the time threshold, the person to be detected is selected as a target person, and the target person represents a person for which the attribute information can be extracted to establish the personal profile. The attribute information may include a specific angle of a photographed picture, a human body of a target person, wearing, accessories, and the like.
The extracting of the attribute information of each target person to construct the personal archive may be filing the attribute information of each target person into its own personal archive, and classifying and storing the attribute information in the personal archive based on a clustering basic service, where the clustering basic service may refer to a process of dividing a set of physical or abstract objects into a plurality of classes composed of similar objects.
203. An item preference matrix for the target person is calculated based on the attribute information in the personal profile.
The attribute information can be multidimensional attribute information, the multidimensional attribute information is used for constructing an article preference matrix of a target person, the multidimensional attribute information is identified through a human body identification algorithm, the identified attribute information is subjected to structuring processing on the attribute information based on video structuring, and the article preference matrix is calculated according to the structured attribute information. The horizontal axis of the item preference matrix may be a plurality of target persons, the vertical axis may represent a plurality of items (attribute information), and each corresponding vector in the matrix may represent the interest level of the target person in the corresponding item. The interest level can be used for knowing the favorite degree of the target person on the object, wherein the higher the interest level is, the more preferred the target person is, and the lower the score is, the less preferred the target person is.
204. And matching the corresponding recommendation list according to the item preference matrix of the target person for pushing.
After the item preference matrix of the target person is obtained, the interest degree of the target person in each item can be obtained, and the interest degree can be embodied through grading. According to the scores, the system can perform data mining, and preferentially match the recommendation list of the items with high interest degree based on a recommendation algorithm, for example: target person A scores 80 for professional garments and 10 for princess skirts, the recommendation algorithm will preferentially match for professional garments.
The recommendation list can comprise a plurality of recommendation items which are arranged in order, when the background management equipment pushes the recommendation list to the projection module, the projection module can perform ordered projection according to the recommendation items which are arranged in order in the recommendation list, and when a plurality of projection modules exist, the recommendation items which are arranged in order can be respectively pushed to the projection modules, so that the projection modules can be projected at the same time, and the improvement of the reach rate of the advertisement is facilitated.
In the embodiment of the invention, as the continuous frame images of the person to be detected are collected, the continuous frame images comprise the face images of the person to be detected; selecting a person to be detected with the retention time of the face image reaching a preset time threshold as a target person, and extracting attribute information of the target person to construct a personal file; calculating an item preference matrix of the target person based on the attribute information in the personal profile; and matching the corresponding recommendation list according to the item preference matrix of the target person for pushing. Therefore, the embodiment of the invention can select the target person according to the retention time and the preset time threshold, extract the attribute information of the target person for structuralization processing, construct the item preference matrix of the target person, and match the recommendation list corresponding to the item preference matrix for pushing, thereby being beneficial to improving the accuracy of advertisement pushing and realizing accurate advertisement delivery.
As shown in fig. 3, fig. 3 is a flowchart of another information pushing method provided in the embodiment of the present invention, which specifically includes the following steps:
301. and acquiring continuous frame images of the person to be detected, wherein the continuous frame images comprise face images of the person to be detected.
302. And recognizing the face image in the continuous frame images.
In the continuous frame images, face images of a plurality of persons to be retrieved may be included. The person to be retrieved can be in a head-down state, a head-up state, a steering state and the like in the elevator, if the camera cannot collect the face image of the person to be retrieved within the first time, the person can be temporarily abandoned to collect the face image, and when the face image can be collected in the subsequent detection, the face image can be continuously identified, for example: when the elevator just enters, the A face image is in a low head state, the camera cannot collect the face image, and after 30 seconds, the camera recognizes the A face. I.e. it means that the camera can capture successive frame images of the person to be detected in successive time periods.
303. And acquiring floor information corresponding to the face image, and calculating the stay time of the person to be detected according to the floor information.
Each face image corresponds to a person to be retrieved, the camera can calculate the staying time of the person to be detected in the elevator by acquiring the floor corresponding to each face image, for example, the person to be detected C presses the 6 th floor of an elevator button, the elevator is entered from 5 pm, and the elevator is separated from the elevator after 2 pm after 5 pm, which indicates that the staying time of the person to be detected C is 2 minutes. Of course, there may be a plurality of people to be detected entering the elevator at the same time, or people to be detected also enter in the midway, and many people go to the same floor, at this time, the required residence time from the floor entering the elevator to the floor that has been pressed in the elevator needs to be calculated respectively, the calculation may be performed on the basis of the floor entering the elevator in the first batch, and the midway entering may not be considered, for example: and (3) entering a person A and a person B in the floor 1, wherein the person A goes to the floor 5, the person B goes to the floor 6, and the person B enters the floor 3 and goes to the floor 5, and the residence time of the person going to the floor 5 and the person going to the floor 6 is calculated by taking the time of the person going to the floor 1 as a starting point.
304. And judging whether the retention time reaches a preset time threshold value, wherein the preset time threshold value comprises advertisement showing time.
Wherein, each floor can have a plurality of advertisements, the advertisement showing time of each advertisement is different, when judging whether the stay time reaches the preset time threshold, it can be judged that the stay time of the corresponding floor is compared with the advertisement showing time of a plurality of advertisements of the floor, and each advertisement showing time equivalent to each floor can be used as the preset time threshold, for example: when a person in the elevator goes to the 5 th floor, a person goes to the 10 th floor, the staying time for going to the 5 th floor is 1 minute, and the staying time for going to the 10 th floor is 2 minutes, the advertisement showing time of a plurality of advertisements of the 5 th floor and the advertisement showing time of a plurality of advertisements of the 10 th floor are respectively obtained, whether the advertisements of the 5 th floor are less than or equal to 1 minute in the advertisement showing time is judged, and whether the advertisements of the 10 th floor are less than or equal to 2 minutes in the advertisement showing time is judged.
305. And if so, selecting the person to be detected as a target person, and extracting attribute information of the target person to construct a personal file.
And in the retention time, if the advertisement with the advertisement showing time of the corresponding floor being less than or equal to the retention time exists, selecting the person to be detected pressing the floor as a target person, extracting the attribute information of the target person, and filing the attribute information to obtain a corresponding personal file.
As a possible embodiment, if the stay time is less than the advertisement showing time of all the advertisements in the floor where the user goes, it indicates that the user has already left the elevator without considering whether the advertisement is the advertisement preferred by the user, but the advertisement has not been played, so that accurate pushing cannot be realized, and resource waste is caused. Therefore, when the stay time is less than the advertisement showing time of all the advertisements in the floor, the start of the pushing system for pushing the advertisements to the person to be detected can be suspended.
306. An item preference matrix for the target person is calculated based on the attribute information in the personal profile.
307. And matching the corresponding recommendation list according to the item preference matrix of the target person for pushing.
In the embodiment of the invention, the floor information corresponding to the face image is obtained by identifying the face image in the continuous frame image, the stay time of the person to be detected is calculated according to the floor information, whether the stay time reaches the advertisement showing time or not is judged, and the person to be detected corresponding to the stay time reaching the advertisement showing time is taken as the target person; after the target person is determined, the attribute information of the target person can be extracted to calculate the item preference matrix of the target person, and the recommendation list corresponding to the item preference matrix is pushed. Like this, through screening the target personnel in the elevator in advance, extract the ability to target personnel's attribute information again, can realize corresponding propelling movement, be favorable to improving the rate of accuracy of advertisement propelling movement, need not all to wait to detect in the elevator attribute information and all acquire, avoid causing the wasting of resources, consuming time overlength, the calculated amount is too big. As shown in fig. 4, fig. 4 is a flowchart of another information pushing method provided in the embodiment of the present invention, where the attribute information includes a tag attribute, and the method specifically includes the following steps:
401. and acquiring continuous frame images of the person to be detected, wherein the continuous frame images comprise face images of the person to be detected.
402. Selecting a person to be detected with the retention time of the face image reaching a preset time threshold value as a target person, carrying out image tracking on the target person, detecting the label attribute of the target person, and carrying out structured processing on the label attribute.
Wherein, can take a candid photograph at the in-process of tracking, take a candid photograph can include that take a candid photograph in a poor light, side take a candid photograph, have shelter from thing to take a candid photograph, shadow etc.. The tag attributes may include, but are not limited to, the target person's body type, height, age, and even brand of apparel and LOGO, type of shoes, color and brand of backpack, hat, glasses, earrings, and the like.
After the target person is determined, human body tracking can be performed on the target person through a Kalman filtering algorithm, and the tag attribute of the target person is detected in real time. In the process of tracking the image, the video image is often influenced by the design of a system instrument, measurement operation and the like, and when the video image is tracked, interference waves often appear, so that filtering can be performed through a Kalman algorithm. The Kalman filtering is an algorithm for carrying out optimal estimation on the system state by using a linear system state equation and inputting and outputting continuous frame image data through a system. The optimal estimation can also be seen as a filtering process, since the influence of noise and interference in the system is included in the successive frame image data. The Kalman filtering can estimate the state of a dynamic system from a series of data with measurement noise under the condition that the measurement variance is known, and can update and process the acquired continuous frame images in real time.
The structuring processing may be processing by a video structuring engine, detecting a plurality of tag attributes, identifying associations between the plurality of tag attributes, and implementing data structuring, and the tag attributes may be used to construct a matrix to calculate an item preference matrix of a target person.
403. And selecting the face image of the target person as an archive ID, and constructing a personal archive of the target person based on the archive ID and the label attribute.
The face can be used as a unique identifier, the recognition degree is high, the face image of each person is used as the file ID to establish the personal file of the target person, and the method has pertinence and safety. The personal profile corresponding to the ID can be directly searched through the profile ID. The personal profile can store tag attributes obtained by structuring tag attributes acquired by target personnel for many times, and the tag attributes acquired each time can be different.
The face image of the selected target person can be obtained by identifying continuous frame images collected by a camera through a face identification technology, the identification can be obtained by collecting a plurality of key points where the face is located, the plurality of face images are distinguished by analyzing the plurality of key points, and different face images are respectively constructed into different archive IDs. The key points may refer to a nose, eyes, eyebrows, a mouth, a cheek, and the like of the human face, and the analyzing of the key points may be analyzing coordinates, distances, and the like between the nose, the eyes, the eyebrows, the mouth, the cheek, and the like.
As an optional embodiment, the same person to be detected may appear in the same scene for multiple times, and when the camera acquires the face image of the person to be detected for multiple times, face recognition may be performed. If it is recognized that the face image is already used as a profile ID and a personal profile is established, the person to be retrieved may be determined as a target person and current attribute information of the target person may be acquired, where the current attribute information refers to information currently being acquired by a camera, for example: the target person's clothing, pants, shoes, hats, bags, jewelry, and the color, style, brand, etc. of the clothing, pants, shoes, hat, bags, jewelry.
And judging whether the specific gravity value of the current attribute information in the personal file reaches a preset specific gravity threshold value.
After the current attribute information is acquired, all attribute information in the target person may be extracted, and the two attribute information are compared, for example: all the attribute information of the target person is 4 and is a sports series, the current attribute information is 2, and the specific gravity value of the skirt and the high-heeled shoes can be represented to be 0. The specific gravity threshold described above may be set in advance, for example, to 1/2.
If so, it may be indicated that the attribute information of the target person is stable, for example: the specific gravity threshold is 3/5, and the specific gravity value is 4/5. And filing the current attribute information into the personal file of the target person, and calculating the item preference matrix of the target person by combining the current attribute information and the attribute information in the established personal file. If not, a personal file of the target person can be created based on the new attribute information, the personal file before the target person is deleted, updating of the personal file is achieved, the personal file is built in real time according to the current attribute information, and an item preference matrix is calculated to obtain a more accurate push list.
404. An item preference matrix for the target person is calculated based on the attribute information in the personal profile.
405. And matching the corresponding recommendation list according to the item preference matrix of the target person for pushing.
In the embodiment of the invention, the target person is selected by collecting continuous frame images of the person to be detected, the label attribute of the target person is detected by carrying out image tracking on the target person, the label attribute is subjected to video structuralization processing, a face image is used as a file ID, a personal file of the target person is constructed according to the structuralized file ID and the label attribute, an article preference matrix of the target person is calculated based on the matrix constructed by the label attribute in the file, and a corresponding recommendation list is matched according to the article preference matrix for pushing. Therefore, the label attribute is detected through image tracking, and the label attribute is subjected to structured processing, so that the method is favorable for accurately positioning advertisement push and covering crowds.
As shown in fig. 5, fig. 5 is a flowchart of another information pushing method provided in the embodiment of the present invention, which specifically includes the following steps:
501. and acquiring continuous frame images of the person to be detected, wherein the continuous frame images comprise face images of the person to be detected.
502. And selecting the person to be detected with the retention time of the face image reaching a preset time threshold as a target person, and extracting attribute information of the target person to construct a personal file.
503. And constructing a user-feature matrix according to the target personnel and the label attributes, wherein the user-feature matrix comprises the interest degree of the target personnel on the label attributes.
See FIG. 6a, in which the user-feature matrix P (user-class matrix) has matrix values Pi，jShows that useriFor classjThe interestingness can be solved by optimizing the loss function. Multiple items (feature attributes) items of the user can be used as a data set, and all items constitute a complete set of items. For each user, the item with past behavior is called as a positive sample, the specified interest RUI is 1, and the user also needs to randomly sample from the item corpus, select samples with the number equivalent to the number of the positive samples as negative samples, and the specified interest RUI is 0. Thus, the value range of interest is [0,1 ]]. And continuously carrying out iterative calculation optimization on the parameters by combining a loss function with a random gradient descent algorithm until the parameters are converged, so that the interestingness can be calculated.
When the same person is detected for multiple times to obtain different tag attributes, the tag attributes with less occurrence times in the personal file of the target person can be deleted, and the tag attributes with more use times are reserved. Therefore, the method is beneficial to calculating the prediction score of the target person on the object more accurately, so that the corresponding recommendation list is matched more accurately.
504. A feature-item matrix is constructed from the tag attributes and the items, the feature-item matrix including weights for the items in the tag attributes.
Wherein, in the feature-item matrix Q (class-item matrix), Qi，jRepresenting itemjIn classiThe higher the weight in (1) is, the more representative the class is. The weights may also be calculated by the optimization penalty function described above.
505. And multiplying the user-feature matrix and the feature-item matrix to obtain an item preference matrix of the target person, wherein the item preference matrix comprises the interest degree of the target person in the item.
Specifically, the behavior data set of the target person (the data set includes all users, all items, and an item list of behaviors of each user) may be modeled by using an LFM (Latent factor model), and assuming that there are 3 users in the data set and 4 items in the data set, the number of classes (class) modeled by the LFM is 4, so as to obtain the model shown in fig. 6 a.
Referring to FIG. 6a, wherein the R matrix is a user-item matrix, the matrix value Ri，jShows that useriTo itemjInterest degree of (1), and useriTo itemjThe interest level of (b) is a value to be solved. For each user, after the interest degrees of the user to all items are calculated, the user can be ranked and recommended. And calculating the interest degrees and the weights to obtain a user-class matrix and a class-item matrix of the target person. The LFM algorithm extracts a plurality of subjects from the data set, the subjects are used as a bridge for connecting between a user and the item, and the R matrix is represented as multiplication of a P matrix and a Q matrix, so that the user-item matrix of the target person can be obtained. Higher interestingness indicates that user prefers the item. Multiplying the user-class matrix with the class-item matrix to obtain a user-item matrix, and calculating the useriTo itemjThe specific algorithm of interest level of (c) can be referred to as shown in FIG. 6b, wherein Pu，kShows that useruFor classkInterest degree of (1), Qk，iRepresenting itemkIn classiThe weight in (1).
506. And matching the corresponding recommendation list according to the item preference matrix of the target person for pushing.
In the embodiment of the invention, the user preference matrix is obtained by obtaining the label attribute of the target person, constructing the user-feature matrix according to the label attribute, constructing the feature-article matrix according to the label attribute and the article, multiplying the label attribute and the article to obtain the user preference matrix, and finally, the corresponding recommendation list is matched according to the article preference matrix of the target person for pushing. Therefore, the user preference matrix is calculated according to the interest degree of the target person on the label attribute and the weight of the object in the label attribute, and the targeted advertisement pushing method is beneficial to more accurately pushing the advertisement to the target person in a targeted mode.
As shown in fig. 7, fig. 7 is a flowchart of another information pushing method provided in the embodiment of the present invention, which specifically includes the following steps:
601. and acquiring continuous frame images of the person to be detected, wherein the continuous frame images comprise face images of the person to be detected.
602. And selecting the person to be detected with the retention time of the face image reaching a preset time threshold as a target person, and extracting attribute information of the target person to construct a personal file.
603. An item preference matrix for the target person is calculated based on the attribute information in the personal profile.
604. And matching the pre-push list according to the item preference matrix of the target person.
The pre-push list is obtained by the system according to a recommendation algorithm and aiming at item matching in an item preference matrix of the target person. Including in the pre-push list a plurality of items similar to the items in the item preference matrix, for example: the item preference matrix comprises rugby shoes and ball coats of adidas, and the recommendation algorithm can judge that the target person is a sports darner and search advertisements of sports series items in the floor where the target person goes to form a pre-push list.
As a possible embodiment, when there are multiple target persons, there is an item preference matrix for each target person. Therefore, the item preference matrix of the target person may be sorted according to the preset target person priority, for example: and 3 persons go to the same floor, the article preference matrix of the No. 1 target person is used for judging that the article is mainly an infant pregnant article, the article preference matrix of the No. 2 target person is used for judging that the article is mainly an old garment, and the article preference matrix of the No. 3 target person is used for judging that the article is mainly a young sports garment, so that the advertisement related to the infant pregnant article can be pushed firstly, then the advertisement related to the young sports garment is pushed, and then the advertisement related to the old garment is pushed. Of course, the above is only an example, and the pushing sequence may be in other manners, which is not limited in the embodiment of the present invention.
605. And calculating the item similarity of the items in the pre-push list and the items in the item preference matrix of the target person.
Wherein the item similarity may represent a similarity between features of two items, for example: the attributes, roles, types, included elements, etc. of the two items are compared. For comparing the similarity of two items, items to be compared may be specified and a corresponding similarity value may be matched for each item, for example: in the pre-recommendation list, adidas's jersey is present, adidas's sneakers are present in the item preference matrix, and when matching, adidas is a common logo, and both the jersey and the sneakers are sports series, the similarity for the same logo match is 50%, the similarity for both sports series matches is 20%, and the item similarity for both items is 70%.
606. And constructing an article similarity matrix, and calculating the prediction preference value of the target personnel.
After the article similarity among a plurality of articles is calculated, a similarity matrix can be constructed by the article similarities, a similarity threshold value can be preset for the article similarities, the article similarity which does not reach the similarity threshold value is deleted, and the waste of resources caused by invalid pushing is avoided. Matrix values in the similarity matrix represent the similarity between the articles, and the sizes of a plurality of prediction preference values of the target person are calculated according to the similarity of the articles.
607. And sequencing the prediction preference values, and pushing a recommendation list corresponding to the sequenced prediction preference values in a matching manner, wherein the recommendation list comprises topN articles.
After calculating the plurality of prediction preference values of the target person, the system can sequence the prediction preference values in sequence according to the magnitude of the prediction preference values, and sequence the prediction preference values with large numerical values in the front. After sorting, a recommendation list with topN articles can be obtained, and the recommendation list is pushed to a target person in an elevator.
In the embodiment of the invention, the pre-push list is matched according to the item preference matrix of the target person, the item similarity between the pre-push list and the items in the item preference matrix is calculated, the item similarity matrix is constructed, the prediction preference value of the target person is calculated, the prediction preference values are sorted, and the recommendation list corresponding to the sorted prediction preference values is matched for pushing. By calculating the similarity of the articles, constructing a similarity matrix to calculate the prediction preference values, and then sequencing the prediction preference values, the method is favorable for preferentially recommending the articles with high prediction preference values to target personnel, and realizes accurate pushing.
Referring to fig. 8, fig. 8 is a schematic structural diagram of an information pushing apparatus according to an embodiment of the present invention, where the apparatus 800 specifically includes:
the acquisition module 801 is used for acquiring continuous frame images of a person to be detected, wherein the continuous frame images comprise face images of the person to be detected;
the extraction module 802 is configured to select a person to be detected, as a target person, for whom the retention time of the face image reaches a preset time threshold, and extract attribute information of the target person to construct a personal file;
a calculating module 803, configured to calculate an item preference matrix of the target person based on the attribute information in the personal profile;
and the pushing module 804 is configured to push the item preference matrix of the target person according to the recommendation list corresponding to the item preference matrix matching.
Optionally, as shown in fig. 9, the extracting module 802 includes:
an identifying unit 8021, configured to identify a face image in the continuous frame images;
the first calculating unit 8022 is configured to obtain floor information corresponding to the face image, and calculate the staying time of the person to be detected according to the floor information;
a judging unit 8023, configured to judge whether the retention time reaches a preset time threshold, where the preset time threshold includes an advertisement showing time;
the selecting unit 8024 is used for selecting the person to be detected as the target person if the target person is reached.
Optionally, the attribute information includes a tag attribute, as shown in fig. 10, the extracting module 802 further includes:
the processing unit 8025 is configured to perform image tracking on the target person, detect a tag attribute of the target person, and perform structuring processing on the tag attribute;
the first construction unit 8026 is configured to select a face image of the target person as an archive ID, and construct a personal archive of the target person based on the archive ID and the tag attribute.
Optionally, as shown in fig. 11, the calculation module 803 includes:
a second constructing unit 8031, configured to construct a user-feature matrix according to the target person and the tag attribute, where the user-feature matrix includes interest of the target person in the tag attribute;
the second construction unit 8031 is further configured to construct a feature-item matrix according to the tag attributes and the items, where the feature-item matrix includes weights of the items in the tag attributes;
the second calculating unit 8032 is configured to multiply the user-feature matrix and the feature-item matrix to obtain an item preference matrix of the target person, where the item preference matrix includes interest level of the target person in the item.
Optionally, as shown in fig. 12, the pushing module 804 includes:
a matching unit 8041, configured to match the pre-push list according to the item preference matrix of the target person;
a third calculating unit 8042, configured to calculate item similarity between the items in the pre-push list and the items in the item preference matrix of the target person;
the second construction unit 8031 is further configured to construct an item similarity matrix, and calculate a predicted preference value of the target person;
the sorting unit 8043 is configured to sort the prediction preference values, and push a recommendation list matched with the sorted prediction preference values, where the recommendation list includes topN items.
Optionally, as shown in fig. 13, the apparatus 800 further includes:
an obtaining module 805, configured to obtain a face image of a person to be retrieved, and if it is recognized that the face image is already used as a file ID and a personal file is established, determine the person to be retrieved as a target person and obtain current attribute information;
a judging module 806, configured to judge whether a specific gravity value of the current attribute information in the personal profile reaches a preset specific gravity threshold;
a filing module 807 for filing the current attribute information into the personal file of the target person if the current attribute information is reached;
and a new creating module 808, configured to create a personal profile of the target person based on the new attribute information if the personal profile does not reach the new attribute information.
The information pushing device provided by the embodiment of the invention can realize each process realized by the information pushing method in the embodiment of the method and can achieve the same beneficial effect, and in order to avoid repetition, the details are not repeated.
Referring to fig. 14, fig. 14 is a schematic structural diagram of an electronic device according to an embodiment of the present invention, as shown in fig. 14, including: a processor 1401, a memory 1402, a network interface 1403, and computer programs stored on the memory 1402 and executable on the processor 1401, wherein:
the processor 1401 is used for calling the computer program stored in the memory 1402, and executing the following steps:
acquiring continuous frame images of a person to be detected, wherein the continuous frame images comprise face images of the person to be detected;
selecting a person to be detected with the retention time of the face image reaching a preset time threshold as a target person, and extracting attribute information of the target person to construct a personal file;
calculating an item preference matrix of the target person based on the attribute information in the personal profile;
and matching the corresponding recommendation list according to the item preference matrix of the target person for pushing.
Optionally, the step performed by the processor 1401 to select a person to be detected whose staying time of the face image reaches a preset time threshold as a target person includes:
identifying face images in the continuous frame images;
acquiring floor information corresponding to the face image, and calculating the stay time of a person to be detected according to the floor information;
judging whether the retention time reaches a preset time threshold value, wherein the preset time threshold value comprises advertisement showing time;
and if so, selecting the person to be detected as the target person.
Optionally, the attribute information executed by the processor 1401 includes a tag attribute, and the step of extracting the attribute information of the target person to construct the personal profile includes:
carrying out image tracking on the target personnel, detecting the label attribute of the target personnel, and carrying out structural processing on the label attribute;
and selecting the face image of the target person as an archive ID, and constructing a personal archive of the target person based on the archive ID and the label attribute.
Optionally, the step of calculating the item preference matrix of the target person based on the attribute information in the personal profile performed by processor 1401 comprises:
constructing a user-feature matrix according to the target personnel and the label attributes, wherein the user-feature matrix comprises interest degrees of the target personnel in the label attributes;
constructing a feature-item matrix according to the label attributes and the items, wherein the feature-item matrix comprises weights of the items in the label attributes;
and multiplying the user-feature matrix and the feature-item matrix to obtain an item preference matrix of the target person, wherein the item preference matrix comprises the interest degree of the target person in the item.
Optionally, the step of pushing, performed by processor 1401, the recommendation list corresponding to the item preference matrix matching of the target person includes:
matching the pre-push list according to the item preference matrix of the target person;
calculating the article similarity of the articles in the pre-push list and the articles in the article preference matrix of the target person;
constructing an article similarity matrix, and calculating a prediction preference value of a target person;
and sequencing the prediction preference values, and pushing a recommendation list corresponding to the sequenced prediction preference values in a matching manner, wherein the recommendation list comprises topN articles.
Optionally, the processor 1401 is further configured to execute acquiring a face image of the person to be retrieved, and if it is recognized that the face image is already used as a profile ID and a personal profile is established, determine the person to be retrieved as a target person and acquire current attribute information;
judging whether the specific gravity value of the current attribute information in the personal file reaches a preset specific gravity threshold value or not;
if so, filing the current attribute information into the personal file of the target person;
and if not, establishing a personal file of the target person based on the new attribute information.
The electronic device 1400 provided by the embodiment of the present invention can implement each implementation manner in the information pushing method embodiment, and has corresponding beneficial effects, and for avoiding repetition, details are not described here.
It is noted that only 1401 and 1403 having components are shown, but it is understood that not all of the illustrated components are required and that more or fewer components can alternatively be implemented. As will be understood by those skilled in the art, the electronic device 1400 is a device capable of automatically performing numerical calculation and/or information processing according to a preset or stored instruction, and the hardware includes, but is not limited to, a microprocessor, an Application Specific Integrated Circuit (ASIC), a Programmable gate array (FPGA), a Digital Signal Processor (DSP), an embedded device, and the like.
The electronic device 1400 may be a desktop computer, a notebook, a palm top computer, a cloud server, or other computing devices. The electronic device 1400 may interact with the client through a keyboard, a mouse, a remote control, a touch pad, or a voice control device.
The memory 1402 includes at least one type of readable storage medium including a flash memory, a hard disk, a multimedia card, a card-type memory (e.g., SD or DX memory, etc.), a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a Read Only Memory (ROM), an Electrically Erasable Programmable Read Only Memory (EEPROM), a Programmable Read Only Memory (PROM), a magnetic memory, a magnetic disk, an optical disk, etc. In some embodiments, the storage 1401 may be an internal storage unit of the electronic device 1400, such as a hard disk or a memory of the electronic device 1400. In other embodiments, the memory 1401 may also be an external storage device of the electronic device 1400, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like, provided on the electronic device 1400. Of course, the memory 1401 may also include both internal and external storage devices for the electronic device 1400. In this embodiment, the memory 1401 is generally configured to store an operating system installed in the electronic device 1400 and various types of application software, such as program codes of an information pushing method. The memory 1401 may also be used to temporarily store various types of data that have been output or are to be output.
The network interface 1403 may include a wireless network interface or a wired network interface, and the network interface 1403 is generally used for establishing a communication connection between the electronic device 1400 and other electronic devices.
The embodiment of the present invention further provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by the processor 1401, the computer program implements each process of the information push method provided in the embodiment of the present invention, and can achieve the same technical effect, and in order to avoid repetition, the computer program is not described herein again.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solution of the present application may be substantially or partially embodied in the form of a software product, where the computer software product is stored in a storage medium (e.g., ROM/RAM, magnetic disk, optical disk), and includes several instructions for enabling a terminal device (e.g., a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the information pushing method of the embodiments of the present application.
The above disclosure is only for the purpose of illustrating the preferred embodiments of the present invention, and it is therefore to be understood that the invention is not limited by the scope of the appended claims.
Claims (10)
1. An information pushing method is characterized by comprising the following steps:
acquiring continuous frame images of a person to be detected, wherein the continuous frame images comprise face images of the person to be detected;
selecting a person to be detected with the retention time of the face image reaching a preset time threshold as a target person, and extracting attribute information of the target person to construct a personal file;
calculating an item preference matrix of the target person based on the attribute information in the personal profile;
and pushing according to the recommendation list corresponding to the item preference matrix matching of the target person.
2. The method of claim 1, wherein the step of selecting the person to be detected with the retention time of the face image reaching a preset time threshold as the target person comprises:
identifying a face image in the continuous frame image;
acquiring floor information corresponding to the face image, and calculating the stay time of the person to be detected according to the floor information;
judging whether the retention time reaches a preset time threshold value, wherein the preset time threshold value comprises advertisement showing time;
and if so, selecting the person to be detected as a target person.
3. The information pushing method according to claim 1, wherein the attribute information includes a tag attribute, and the step of extracting the attribute information of the target person to construct a personal profile includes:
carrying out image tracking on the target personnel, detecting the label attribute of the target personnel, and carrying out structuralization processing on the label attribute;
and selecting the face image of the target person as an archive ID, and constructing a personal archive of the target person based on the archive ID and the label attribute.
4. The information push method according to claim 3, wherein the step of calculating the item preference matrix of the target person based on the attribute information in the personal profile comprises:
constructing a user-feature matrix according to the target personnel and the label attributes, wherein the user-feature matrix comprises interest degrees of the target personnel in the label attributes;
constructing a feature-item matrix according to the tag attributes and the items, wherein the feature-item matrix comprises weights of the items in the tag attributes;
and multiplying the user-feature matrix and the feature-item matrix to obtain an item preference matrix of the target person, wherein the item preference matrix comprises the interest degree of the target person in the item.
5. The information push method according to claim 1, wherein the step of pushing by matching the corresponding recommendation list according to the item preference matrix of the target person comprises:
matching a pre-push list according to the item preference matrix of the target person;
calculating the item similarity of the items in the pre-push list and the items in the item preference matrix of the target person;
constructing an article similarity matrix, and calculating the prediction preference value of the target personnel;
and sequencing the prediction preference values, and matching a recommendation list corresponding to the sequenced prediction preference values for pushing, wherein the recommendation list comprises topN articles.
6. The information pushing method of claim 3, wherein the method further comprises:
acquiring a face image of a person to be retrieved, if the face image is identified as a file ID and a personal file is established, judging the person to be retrieved as a target person and acquiring current attribute information;
judging whether the specific gravity value of the current attribute information in the personal file reaches a preset specific gravity threshold value or not;
if so, filing the current attribute information in the personal file of the target person;
and if not, establishing the personal profile of the target person based on the new attribute information.
7. An information pushing apparatus, comprising:
the system comprises an acquisition module, a processing module and a display module, wherein the acquisition module is used for acquiring continuous frame images of a person to be detected, and the continuous frame images comprise face images of the person to be detected;
the extraction module is used for selecting the person to be detected with the retention time of the face image reaching a preset time threshold value as a target person and extracting attribute information of the target person to construct a personal file;
the calculation module is used for calculating an item preference matrix of the target person based on the attribute information in the personal file;
and the pushing module is used for pushing according to the recommendation list corresponding to the item preference matrix matching of the target person.
8. The information pushing apparatus according to claim 7, wherein the extracting module includes:
the recognition unit is used for recognizing the face image in the continuous frame image;
the calculating unit is used for acquiring floor information corresponding to the face image and calculating the stay time of the person to be detected according to the floor information;
the judging unit is used for judging whether the staying time reaches a preset time threshold value, and the preset time threshold value comprises advertisement showing time;
and the selecting unit is used for selecting the person to be detected as the target person if the person to be detected is reached.
9. An electronic device, comprising: memory, processor and computer program stored on the memory and executable on the processor, the processor implementing the steps in the information pushing method according to any one of claims 1 to 6 when executing the computer program.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps in the information push method according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911257893.7A CN112950242A (en) | 2019-12-10 | 2019-12-10 | Information pushing method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911257893.7A CN112950242A (en) | 2019-12-10 | 2019-12-10 | Information pushing method and device and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112950242A true CN112950242A (en) | 2021-06-11 |
Family
ID=76225389
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911257893.7A Pending CN112950242A (en) | 2019-12-10 | 2019-12-10 | Information pushing method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112950242A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115034813A (en) * | 2022-06-01 | 2022-09-09 | 中通服慧展科技有限公司 | Advertisement display system integrating multiple communication functions |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104573619A (en) * | 2014-07-25 | 2015-04-29 | 北京智膜科技有限公司 | Method and system for analyzing big data of intelligent advertisements based on face identification |
| CN105355158A (en) * | 2015-11-12 | 2016-02-24 | 日立电梯（中国）有限公司 | Elevator advertisement posting method and system |
| US20160150260A1 (en) * | 2014-11-23 | 2016-05-26 | Christopher Brian Ovide | System And Method For Creating Individualized Mobile and Visual Advertisments Using Facial Recognition |
| CN107066476A (en) * | 2016-12-13 | 2017-08-18 | 江苏途致信息科技有限公司 | A kind of real-time recommendation method based on article similarity |
| CN107239993A (en) * | 2017-05-24 | 2017-10-10 | 海南大学 | A kind of matrix decomposition recommendation method and system based on expansion label |
| CN108021708A (en) * | 2017-12-27 | 2018-05-11 | 广州启生信息技术有限公司 | Content recommendation method, device and computer-readable recording medium |
| CN108197971A (en) * | 2017-12-08 | 2018-06-22 | 北京天正聚合科技有限公司 | Information collecting method, information processing method, apparatus and system |
| CN109146626A (en) * | 2018-08-14 | 2019-01-04 | 中山大学 | A kind of fashion clothing collocation recommended method based on user's dynamic interest analysis |
| CN109345305A (en) * | 2018-09-28 | 2019-02-15 | 广州凯风科技有限公司 | A kind of elevator electrical screen advertisement improvement analysis method based on face recognition technology |
| CN109558535A (en) * | 2018-11-05 | 2019-04-02 | 重庆中科云丛科技有限公司 | The method and system of personalized push article based on recognition of face |
| CN109903114A (en) * | 2017-12-11 | 2019-06-18 | 日立楼宇技术（广州）有限公司 | The method and device of information triggering response in ladder |
-
2019
- 2019-12-10 CN CN201911257893.7A patent/CN112950242A/en active Pending
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104573619A (en) * | 2014-07-25 | 2015-04-29 | 北京智膜科技有限公司 | Method and system for analyzing big data of intelligent advertisements based on face identification |
| US20160150260A1 (en) * | 2014-11-23 | 2016-05-26 | Christopher Brian Ovide | System And Method For Creating Individualized Mobile and Visual Advertisments Using Facial Recognition |
| CN105355158A (en) * | 2015-11-12 | 2016-02-24 | 日立电梯（中国）有限公司 | Elevator advertisement posting method and system |
| CN107066476A (en) * | 2016-12-13 | 2017-08-18 | 江苏途致信息科技有限公司 | A kind of real-time recommendation method based on article similarity |
| CN107239993A (en) * | 2017-05-24 | 2017-10-10 | 海南大学 | A kind of matrix decomposition recommendation method and system based on expansion label |
| CN108197971A (en) * | 2017-12-08 | 2018-06-22 | 北京天正聚合科技有限公司 | Information collecting method, information processing method, apparatus and system |
| CN109903114A (en) * | 2017-12-11 | 2019-06-18 | 日立楼宇技术（广州）有限公司 | The method and device of information triggering response in ladder |
| CN108021708A (en) * | 2017-12-27 | 2018-05-11 | 广州启生信息技术有限公司 | Content recommendation method, device and computer-readable recording medium |
| CN109146626A (en) * | 2018-08-14 | 2019-01-04 | 中山大学 | A kind of fashion clothing collocation recommended method based on user's dynamic interest analysis |
| CN109345305A (en) * | 2018-09-28 | 2019-02-15 | 广州凯风科技有限公司 | A kind of elevator electrical screen advertisement improvement analysis method based on face recognition technology |
| CN109558535A (en) * | 2018-11-05 | 2019-04-02 | 重庆中科云丛科技有限公司 | The method and system of personalized push article based on recognition of face |
Non-Patent Citations (1)
| Title |
|---|
| 何明;要凯升;杨;张久伶;: "基于标签信息特征相似性的协同过滤个性化推荐", 计算机科学, no. 1 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115034813A (en) * | 2022-06-01 | 2022-09-09 | 中通服慧展科技有限公司 | Advertisement display system integrating multiple communication functions |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20210406960A1 (en) | Joint-based item recognition | |
| Li et al. | A richly annotated dataset for pedestrian attribute recognition | |
| CN106776619B (en) | Method and device for determining attribute information of target object | |
| Nixon et al. | On soft biometrics | |
| US10747826B2 (en) | Interactive clothes searching in online stores | |
| US11086924B2 (en) | Image search device and image search method | |
| Hadi Kiapour et al. | Where to buy it: Matching street clothing photos in online shops | |
| US8724845B2 (en) | Content determination program and content determination device | |
| US9111147B2 (en) | Assisted video surveillance of persons-of-interest | |
| WO2017114237A1 (en) | Image query method and device | |
| JP6254836B2 (en) | Image search apparatus, control method and program for image search apparatus | |
| JP6500374B2 (en) | Image processing apparatus and image processing program | |
| US9959480B1 (en) | Pixel-structural reference image feature extraction | |
| JP2008203916A (en) | Image processing apparatus, program, and image processing method | |
| CN112150349A (en) | Image processing method and device, computer equipment and storage medium | |
| Li et al. | Personrank: Detecting important people in images | |
| AU2017231602A1 (en) | Method and system for visitor tracking at a POS area | |
| JP2017084078A (en) | Style search apparatus, method, and program | |
| JP2014229129A (en) | Combination presentation system and computer program | |
| Jiang et al. | A unified tree-based framework for joint action localization, recognition and segmentation | |
| CN112950242A (en) | Information pushing method and device and electronic equipment | |
| KR102323861B1 (en) | System for selling clothing online | |
| JP2019185205A (en) | Information processor and information processing method and program | |
| JP2017130061A (en) | Image processing system, image processing method and program | |
| CN111447260A (en) | Information pushing and information publishing 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 |