[go: up one dir, main page]

CN114418605B - Interactive elevator advertising intelligent recommendation method and system - Google Patents

Interactive elevator advertising intelligent recommendation method and system Download PDF

Info

Publication number
CN114418605B
CN114418605B CN202111422668.1A CN202111422668A CN114418605B CN 114418605 B CN114418605 B CN 114418605B CN 202111422668 A CN202111422668 A CN 202111422668A CN 114418605 B CN114418605 B CN 114418605B
Authority
CN
China
Prior art keywords
passenger
information
advertisement
passengers
elevator
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.)
Active
Application number
CN202111422668.1A
Other languages
Chinese (zh)
Other versions
CN114418605A (en
Inventor
吴彬
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chengdu Xinchao Media Group Co Ltd
Original Assignee
Chengdu Xinchao Media Group Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Chengdu Xinchao Media Group Co Ltd filed Critical Chengdu Xinchao Media Group Co Ltd
Priority to CN202111422668.1A priority Critical patent/CN114418605B/en
Publication of CN114418605A publication Critical patent/CN114418605A/en
Application granted granted Critical
Publication of CN114418605B publication Critical patent/CN114418605B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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
    • G06Q30/0242—Determining effectiveness of advertisements
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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
    • G06F16/903—Querying
    • G06F16/9035—Filtering based on additional data, e.g. user or group profiles
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Accounting & Taxation (AREA)
  • Development Economics (AREA)
  • Finance (AREA)
  • Strategic Management (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Marketing (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Computational Linguistics (AREA)
  • Databases & Information Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Business, Economics & Management (AREA)
  • Evolutionary Biology (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • Computing Systems (AREA)
  • Molecular Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Indicating And Signalling Devices For Elevators (AREA)

Abstract

The invention discloses an intelligent recommendation method and system for interactive elevator advertisements, wherein the recommendation method comprises the steps of collecting passenger information in an elevator through a monitoring camera, collecting passenger interaction scene information through a touch advertisement screen, analyzing the passenger information based on a passenger attribute identification algorithm, storing the interaction scene information into a database module, constructing an advertisement information set and an interaction scene database according to the interaction scene information, intelligently recommending advertisements to passengers based on a scene-aware FP-growth recommendation algorithm on the basis of the interaction scene database, obtaining relevant advertisements interested by target passengers in the whole elevator cluster, locally correcting the results, searching the target passengers, and recommending the advertisements to the target passengers. The invention designs the intelligent recommendation method and system for the interactive elevator advertisement with certain robustness and higher accuracy based on the touch type advertisement display screen and combined with the passenger attribute and the interactive information, thereby realizing the maximum advertisement putting benefit.

Description

Interactive elevator advertisement intelligent recommendation method and system
Technical Field
The invention relates to the technical field of elevator advertisements, in particular to an intelligent recommendation method and system for interactive elevator advertisements.
Background
Along with the increase of the domestic elevator, the value of data information generated in the elevator is also more and more paid attention to. The elevator advertisement can be directly contacted with community users as a novel advertisement medium, and serves users better by utilizing mass information generated in the elevator, so that an intelligent elevator advertisement recommendation system meeting application scenes is necessary to be designed.
More intelligent recommending researches of the conventional elevator advertisement are unidirectional analysis of a recommending system, intelligent advertisement delivery is realized based on passenger attribute detection and the like, but the correlation analysis of interactive behavior data of passengers and advertisements is lacked, and the advertisement delivery benefit is not maximized.
Disclosure of Invention
The invention provides an intelligent recommendation method and system for interactive elevator advertisements, which are used for solving the problem that the prior art lacks of carrying out association analysis on interactive behavior data of passengers and advertisements and does not maximize advertisement delivery benefits.
On the premise of having a touch type advertisement display screen, the intelligent recommendation method and system for the elevator advertisement are combined with passenger attribute (system initiative analysis) and interaction information (interaction behaviors such as sliding the display screen by passengers, clicking the advertisement and the like, and passenger initiative feedback information), are further explored on the basis of the existing research, and the intelligent recommendation method and system for the elevator advertisement are designed, have certain robustness and high accuracy, are beneficial to users to obtain better advertisement service, and realize maximized advertisement putting benefit.
The technical scheme adopted by the invention is that the intelligent recommending method for the interactive elevator advertisement comprises the following steps:
passenger information is collected in the elevator through the monitoring camera, and passenger interaction scene information is collected through the touch advertisement screen;
analyzing the passenger information based on a passenger attribute identification algorithm;
storing the interaction scenario information into a database module, and constructing an advertisement information set and an interaction scenario database according to the interaction scenario information;
Based on an interaction scene database, intelligently recommending advertisements to passengers based on a scene-aware FP-growth recommendation method;
obtaining relevant advertisements interested by target passengers in the whole elevator cluster, and carrying out local correction on the results;
Searching for a target passenger and recommending an advertisement to the target passenger.
As a preferred mode of the intelligent recommendation method for the interactive elevator advertisement, the method for constructing the advertisement information set and the interactive scene database according to the interactive scene information comprises the following steps:
Constructing an advertisement information set, wherein S= { S 1,s2,...,sm }, S i＝{k1,k2,ki,....Km},1≤i≤m,si is a sequence of advertisements k i focused in a t 1 time period of a target passenger, the beginning time of the t 1 time period is when the passenger clicks an advertisement interface, and the ending time is when the passenger leaves an elevator;
An interaction scenario database is constructed, namely D= { D 1,d2,...,dn }, wherein D i＝{[x,y]:s1,s2,...},1≤i≤n,di represents specific interaction scenario data, namely a thing database formed by target passengers and advertisement information sets, wherein [ x, y ] represents target passenger attributes, x is gender attribute, and y is age attribute.
As a preferable mode of the intelligent recommendation method for the advertisement of the interactive elevator, the time threshold t 2、t3;t2 is increased in the process of constructing the advertisement information set, the camera continuously loses passenger information to indicate that passengers leave the elevator, the advertisement screen automatically exits from the interactive interface until the next click is received, the construction of the advertisement information set is restarted, if the time period from the time when the passengers click the advertisement k i to the time when the advertisement is switched is greater than t 3, k i is added to s i, and otherwise, no addition is carried out.
As a preferable mode of the intelligent recommendation method for the interactive elevator advertisement, the method for intelligently recommending the advertisement to the passenger based on the FP-growth recommendation algorithm of the context awareness based on the interactive context database comprises the following steps:
The method comprises the steps of establishing a head list, namely, the number of each k i in an advertisement information set S, defining the minimum support degree as T MS,TMS E (0, 1), scanning the advertisement information set S to obtain the count of all frequent 1 item sets (k 1,k2,ki) C k2,Cki..a.). Then deleting items with support below the threshold T MS, placing frequent 1 item sets into the item header table, and arranging in descending order of support;
Constructing an FP-tree, scanning an advertisement information set, eliminating the read original data without frequent 1 item set, arranging the read original data in descending order according to the support degree, reading the ordered data set, inserting the ordered data set into the FP-tree according to the ordered sequence, wherein the node with the front ordering is a root node and the node with the back ordering is a child node, adding 1 to the corresponding count if the common root node exists, and connecting the node corresponding to the head list with the new node through a node linked list if the new node exists until all the data are inserted into the FP-tree;
The method comprises the steps of excavating an FP-tree, sequentially and upwards finding a condition mode base corresponding to an item of the item head table from the bottom item of the item head table, recursively excavating from the condition mode base to obtain a frequent item set of the item head table, defining the item number N FP of the frequent item set, and returning only the frequent item set meeting the item number requirement;
and defining a global weighting parameter, and carrying out batch normalization on the occurrence times of the specific items of the maximum frequent item set to serve as the global weighting parameter corresponding to each specific item.
As a preferred mode of the intelligent recommendation method for the interactive elevator advertisement, the method for locally correcting the result comprises the following steps:
clustering the interactive data scene information of specific points based on a K-means algorithm, wherein the size of the clusters is equal to the attribute category of the passengers, and carrying out batch normalization on the advertisement occurrence times in each cluster;
And defining an advertisement attention threshold T P, reserving advertisement items with scores higher than T P, and generating a new advertisement catalogue.
As a preferred mode of the intelligent recommendation method for the interactive elevator advertisement, the method for searching the target passengers comprises the following steps:
After a passenger enters an elevator, detecting the position of the passenger based on a deep learning algorithm, detecting the position of the human eyes based on an opencv human eye detection model (haarcascade _eye.xml), performing template matching according to a human eye state library, and performing human eye attention position detection on the passenger, wherein the human eye state library comprises 5 types of forward, left, right, upper and lower attention, and when the passenger faces the advertisement screen, the human eye state is considered to be forward due to the fact that the camera is positioned above the advertisement screen, and establishing a target passenger model:
F(x)=αR1(x)+βR2(x)
Wherein x is a specific passenger, R 1 (x) is the area of a passenger detection frame, R 2 (x) is the confidence level of the passenger facing the right front, alpha and beta are weight factors of corresponding factors respectively, F (x) is the final score, and the passengers are ranked according to the weight factors, the priority of the passenger is confirmed, the priority of the passenger is higher and is the main target passenger, and the advertisement catalogue is recommended based on the attribute of the passenger.
As a preferred mode of the intelligent interactive elevator advertisement recommending method, the method for recommending advertisements to the target passengers comprises the following steps:
generating a target passenger advertisement catalog based on a scene-aware FP-growth recommendation algorithm;
The advertisement catalogue of the target passenger x is played, the camera still detects the target passenger x, if no other target passenger exists, the content in the advertisement catalogue is played in turn, and if other target passenger x+i exists, the advertisement catalogue corresponding to x+i is played;
The advertisement catalogue of the target passenger x is not completely played, the target passenger x gets out of the ladder, if no other target passenger exists, the advertisement screen resumes default playing, and if other target passenger x+i exists, the advertisement catalogue corresponding to x+i is sequentially played according to the priority.
The invention also provides an intelligent interactive elevator advertisement recommendation system, which comprises:
the data center control center is responsible for calculating, storing and transmitting the data information of the whole elevator cluster;
the information acquisition module is used for acquiring attribute information and interaction information of passengers in the elevator;
the passenger attribute analysis module is used for obtaining passenger attribute information based on a deep learning algorithm;
The database module is used for storing the scene information of the interaction between the passenger and the touch advertisement screen in a data form;
The global recommendation module analyzes database information based on a scene-aware FP-growth recommendation algorithm and intelligently recommends advertisements to passengers based on the data of the whole elevator cluster;
the local correction module is used for further analyzing specific point location information based on K-means to make comprehensive judgment on the basis of global recommendation.
The intelligent recommendation method and system for the elevator advertisement have the advantages that on the premise of the touch type advertisement display screen, the intelligent recommendation method and system for the elevator advertisement are combined with passenger attributes and interaction information, are further explored on the basis of existing researches, have certain robustness and high accuracy, are beneficial to users to obtain better advertisement service, and achieve maximum advertisement putting benefit.
Drawings
Fig. 1 is a schematic flow chart of an intelligent interactive elevator advertisement recommendation method disclosed by the invention.
Fig. 2 is a block diagram of an interactive elevator advertisement intelligent recommendation system disclosed by the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the accompanying drawings, but embodiments of the present invention are not limited thereto.
Example 1:
referring to fig. 1, the embodiment discloses an intelligent recommendation method for an interactive elevator advertisement, which comprises the following steps:
and S1, collecting passenger information in the elevator through a monitoring camera, and collecting passenger interaction scene information through a touch advertisement screen.
The elevator environment is that an elevator touch advertisement screen is arranged in an elevator and on the left side or the right side of the same side as an elevator door, one side of a key is avoided, the elevator passengers face the elevator, a monitoring camera is positioned at the top of a car and right above the advertisement screen, the whole elevator area can be monitored, and passengers in front of the advertisement screen can be accurately analyzed.
The touch advertisement screen is provided with a main playing window and an advertisement catalog window, advertisements can be daily rolled and played in the main playing window, a rolling bar can be slid in the advertisement catalog window, any interested advertisement can be clicked, and if the currently played advertisement is the last advertisement in the advertisement catalog, the advertisement screen can be used for broadcasting in turn.
The information acquisition is divided into two parts, wherein one part is used for acquiring passenger information through the monitoring camera, and the other part is used for acquiring passenger interaction scene information through the touch advertisement screen. The passenger information is fed back to S2, and the interaction scenario information is fed back to S3.
And S2, analyzing the passenger information based on a passenger attribute identification algorithm.
The passenger information is analyzed based on the passenger attribute recognition algorithm, and the embodiment only feeds back and analyzes the sex information and the age information of the passenger. In addition, it is also possible to consider adding attribute information such as the height of the passenger, clothing, whether to wear the clothing, and the like.
Passenger attribute recognition is paid attention to in the field of video monitoring, and in order to improve accuracy of passenger attribute recognition, a deep learning method is adopted. The present embodiment identifies the gender, age of the passenger based on the attribute-localization model (Attribute Localization Module, ALM). The model does not need additional region marking (realized by STN (Spatial Transformer Networks)), can perform end-to-end training under multi-scale depth supervision, and is obviously superior to most of the existing methods in three pedestrian attribute data sets of PETA, RAP and PA-100K, wherein the Females (sexes: males, females) accuracy in the RAP data set reaches 96%, and the AgeLess accuracy reaches 88%, which are superior to the baseline model of the RAP data set. The RAP data set has 41585 pedestrian samples, contains 72 attributes, is collected from 26 cameras monitored indoors, has resolution ranging from 36×92 to 344×554, and has a certain similarity with an elevator scene.
And S3, storing the interaction scenario information into a database module, and constructing an advertisement information set and an interaction scenario database according to the interaction scenario information.
The passenger interacts with the advertising screen, selects the advertisement of interest through the catalog window, clicks the jump to obtain advertisement details. The recommendation system stores the passenger attributes and the interaction scenario information in the database module. The interaction scenario information refers to a sequence of advertisements that the targeted passenger is interested in over a period of time. The specific process is as follows.
Constructing an advertisement information set, wherein s= { S 1,s2,...,sm }, S i＝{k1,k2,ki,....Km},1≤i≤m,si is a sequence of advertisements k i focused by a target passenger in a t 1 time period, the beginning time of the t 1 time period is the time period when the passenger clicks an advertisement interface, the ending time of the t 1 time period is the time period when the passenger leaves an elevator, the continuous loss of the passenger information by a camera in an increasing time threshold t 2、t3;t2 time period indicates that the passenger leaves the elevator, the advertisement screen automatically exits from the interactive interface until the next click is received, the advertisement information set is restarted, and if the time period from the time when the passenger clicks the advertisement k i to the time when the advertisement k 3 is greater than t 3, k i is added into S i, otherwise, the advertisement screen is not added.
An interaction scenario database is constructed, namely D= { D 1,d2,...,dn }, wherein D i＝{[x,y]:s1,s2,...},1≤i≤n,di represents specific interaction scenario data, namely a thing database formed by target passengers and advertisement information sets, wherein [ x, y ] represents target passenger attributes, x is gender attribute, and y is age attribute. In addition, other attributes such as height, clothes, whether to wear the ornaments and the like can also be added.
S4, based on the interaction scene database, intelligently recommending advertisements to passengers based on a scene-aware FP-growth recommendation algorithm.
The FP-growth is an algorithm for mining data association rules, effectively compresses a database into a data structure with a small storage space, overcomes the defect of multiple-scanning the database in the classical algorithm Aprior, only needs to scan the database twice, and converts the problem of finding a long frequent pattern into a strategy of recursion pattern growth, thereby avoiding generating a large number of candidate sets and greatly reducing the time complexity of the algorithm.
The specific process is as follows:
The method comprises the steps of establishing a head list, namely, the number of each k i in an advertisement information set S, defining the minimum support degree as T MS,TMS E (0, 1), scanning the advertisement information set S to obtain the count of all frequent 1 item sets (k 1,k2,ki) C k2,Cki..a.). Then deleting items with support below the threshold T MS, placing frequent 1 item sets into the item header table, and arranging in descending order of support;
Constructing an FP-tree, scanning an advertisement information set, eliminating the read original data without frequent 1 item set, arranging the read original data in descending order according to the support degree, reading the ordered data set, inserting the ordered data set into the FP-tree according to the ordered sequence, wherein the node with the front ordering is a root node and the node with the back ordering is a child node, adding 1 to the corresponding count if the common root node exists, and connecting the node corresponding to the head list with the new node through a node linked list if the new node exists until all the data are inserted into the FP-tree;
The method comprises the steps of excavating an FP-tree, sequentially and upwards finding a condition mode base corresponding to an item of the item head table from the bottom item of the item head table, recursively excavating from the condition mode base to obtain a frequent item set of the item head table, defining the item number N FP of the frequent item set, and returning only the frequent item set meeting the item number requirement;
and defining a global weighting parameter, and carrying out batch normalization on the occurrence times of the specific items of the maximum frequent item set to serve as the global weighting parameter corresponding to each specific item.
S5, obtaining relevant advertisements interested by target passengers in the whole elevator cluster, and carrying out local correction on the results.
Relevant advertisements of interest to the target passengers in the whole elevator cluster are obtained from the global recommendation of S4, but each specific point should also take into account its own relative independence when recommending. Based on a K-means algorithm, clustering the interactive data scene information of specific points, wherein the size of the clusters is equal to the attribute category of passengers, and carrying out batch normalization on the advertisement occurrence times in each cluster. And (3) weighting and reordering similar commodities in the clustering result by combining the global recommendation of the S4. An advertisement attention threshold T P is defined, advertisement items with scores higher than T P are reserved, and a new advertisement catalog is generated.
And S6, searching for the target passengers and recommending advertisements to the target passengers.
After a passenger enters an elevator, detecting the position of the passenger based on a deep learning algorithm, detecting the position of a human eye based on an opencv eye detection model (haarcascade _eye.xml), and performing template matching according to a human eye state library to detect the attention position of the passenger, wherein the human eye state library comprises 5 types of attention to the front, left, right, upper and lower directions, and when the passenger faces the advertisement screen, the human eye state is considered to be the front, and a target passenger model is established because the camera is positioned above the advertisement screen:
F(x)=αR1(x)+βR2(x)
Wherein x is a specific passenger, R 1 (x) is the area of a passenger detection frame, R 2 (x) is the confidence level of the passenger facing the right front, alpha and beta are weight factors of corresponding factors respectively, F (x) is the final score, and the passengers are ranked according to the weight factors, the priority of the passenger is confirmed, the priority of the passenger is higher and is the main target passenger, and the advertisement catalogue is recommended based on the attribute of the passenger.
The method for recommending advertisements to the target passengers comprises the following steps:
And generating a target passenger advertisement catalog based on a scene-aware FP-growth recommendation algorithm. And after the advertisement catalogue of the target passenger x is played, the camera still detects the target passenger x, if no other target passenger exists, the contents in the advertisement catalogue are played in turn, and if other target passenger x+i exists, the advertisement catalogue corresponding to x+i is played. The advertisement catalogue of the target passenger x is not completely played, the target passenger x gets out of the ladder, if no other target passenger exists, the advertisement screen resumes default playing, and if other target passenger x+i exists, the advertisement catalogue corresponding to x+i is sequentially played according to the priority.
Example 2
Referring to fig. 2, the embodiment discloses an interactive elevator advertisement intelligent recommendation system, which includes:
the data center control center is responsible for calculating, storing and transmitting the data information of the whole elevator cluster;
the information acquisition module is used for acquiring attribute information and interaction information of passengers in the elevator;
the passenger attribute analysis module is used for obtaining passenger attribute information based on a deep learning algorithm;
The database module is used for storing the scene information of the interaction between the passenger and the touch advertisement screen in a data form;
The global recommendation module analyzes database information based on a scene-aware FP-growth recommendation algorithm and intelligently recommends advertisements to passengers based on the data of the whole elevator cluster;
the local correction module is used for further analyzing specific point location information based on K-means to make comprehensive judgment on the basis of global recommendation, so that advertisement recommendation is more accurate and efficient.
The foregoing embodiments are merely for illustrating the technical solution of the present invention, but not for limiting the same, and although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those skilled in the art that modifications may be made to the technical solution described in the foregoing embodiments or equivalents may be substituted for parts of the technical features thereof, and such modifications or substitutions do not depart from the spirit and scope of the technical solution of the embodiments of the present invention in essence.

Claims (4)

1.一种交互式电梯广告智能推荐方法，其特征在于，包括如下步骤：1. An interactive elevator advertisement intelligent recommendation method, characterized in that it comprises the following steps: 在电梯内通过监控摄像头对乘客信息进行采集，以及通过触摸广告屏对乘客交互情景信息进行采集；Collect passenger information through surveillance cameras in the elevator, and collect passenger interaction scenario information through touch advertising screens; 基于乘客属性识别算法对乘客信息进行分析；Analyze passenger information based on passenger attribute recognition algorithm; 将交互情景信息存储至数据库模块中，并根据交互情景信息构造广告信息集以及交互情景数据库；The interactive scenario information is stored in a database module, and an advertisement information set and an interactive scenario database are constructed according to the interactive scenario information; 所述根据交互情景信息构造广告信息集以及交互情景数据库的方法包括：The method for constructing an advertisement information set and an interactive scenario database according to interactive scenario information comprises: 构造广告信息集：，其中，，为目标乘客时间段内关注的广告的序列，时间段的开始时间为乘客点击广告界面，结束时间为乘客离开电梯；Construct an advertisement information set: ,in , , Target passengers Ads viewed during the time period sequence, The start time of the time period is when the passenger clicks the advertising interface, and the end time is when the passenger leaves the elevator; 构造交互情景数据库：，其中，，dl表示具体的交互情景信息，即目标乘客与广告信息集构成的事物数据库，其中表示目标乘客属性，为性别属性，为年龄属性；Constructing an interactive scenario database: ,in , , d l represents the specific interactive scenario information, that is, the object database consisting of the target passengers and the advertising information set, where represents the target passenger attributes, is the gender attribute, is the age attribute; 在交互情景数据库的基础之上，基于情景感知的FP-growth推荐算法向乘客智能推荐广告；Based on the interactive scenario database, the FP-growth recommendation algorithm based on scenario perception intelligently recommends advertisements to passengers; 获得整个电梯集群中目标乘客感兴趣的关联广告，并对结果进行局部校正；Obtain relevant advertisements that target passengers in the entire elevator cluster are interested in, and perform local correction on the results; 寻找目标乘客，并对目标乘客推荐广告；Find target passengers and recommend advertisements to them; 在构造广告信息集的过程中增加时间阈值、；时间段内摄像头连续丢失乘客信息表示乘客离开电梯，广告屏自动退出交互界面，直到接收到下一次点击则重新开始构造广告信息集；乘客点击广告至切换界面的时长若大于，则将加入，否则不加入；Add time threshold in the process of constructing advertisement information set , ; If the camera loses passenger information continuously within a time period, it means that the passenger has left the elevator, and the advertising screen automatically exits the interactive interface until the next click is received, and then the advertising information set is reconstructed; the passenger clicks the ad If the time to switch interface is greater than , then join in , otherwise do not join; 所述在交互情景数据库的基础之上，基于情景感知的FP-growth推荐算法向乘客智能推荐广告的方法包括：The method for intelligently recommending advertisements to passengers based on the interactive scenario database and the scenario-aware FP-growth recommendation algorithm includes: 建立项头表，即广告信息集中每个出现个数，定义最小支持度为，；扫描广告信息集，得到所有频繁1项集的计数；然后删除支持度低于阈值的项，将频繁1项集放入项头表，并按照支持度降序排列；Create a header table, i.e., an advertisement information set Each The number of occurrences, the minimum support is defined as , ; Scan the advertising information set , get all frequent 1-item sets Count ; Then delete the support value below the threshold , put the frequent 1-item set into the item header table and arrange them in descending order of support; 构造FP-tree，扫描广告信息集，将读到的原始数据剔除非频繁1项集，并按照支持度降序排列；读入排序后的数据集，按照排序后的顺序，插入FP-tree中，排序靠前的节点是根节点，而靠后的是子节点；如果有共用的根节点，则对应计数加1；如果有新节点出现，则项头表对应的节点会通过节点链表链接上新节点；直到所有的数据都插入到FP-tree；Construct FP-tree, scan the advertisement information set, remove the infrequent 1-item set from the original data, and sort them in descending order of support; read the sorted data set, and insert it into the FP-tree in the sorted order. The nodes at the front are the root nodes, and the nodes at the back are the child nodes; if there is a common root node, the corresponding count is increased by 1; if a new node appears, the node corresponding to the item header table will be linked to the new node through the node linked list; until all the data are inserted into the FP-tree; 挖掘FP-tree，从项头表的底部项依次向上找到项头表项对应的条件模式基；从条件模式基递归挖掘得到项头表项的频繁项集；定义频繁项集的项数，只返回满足项数要求的频繁项集；Mining FP-tree, from the bottom item of the header table upwards to find the conditional pattern base corresponding to the item in the header table; recursively mining from the conditional pattern base to obtain the frequent item set of the item in the header table; defining the number of items in the frequent item set , only frequent itemsets that meet the number of item requirements are returned; 定义全局加权参数，对最大频繁项集的具体条项出现次数进行批归一化，作为每个具体条项对应的全局加权参数；Define the global weight parameter, and perform batch normalization on the number of occurrences of specific items in the maximum frequent itemset as the global weight parameter corresponding to each specific item; 所述对结果进行局部校正的方法包括：The method for locally correcting the result comprises: 基于K-means算法对具体点位的交互情景信息进行聚类，簇的大小等于乘客属性类别，对每一簇中广告出现次数进行批归一化；The interactive scenario information of specific points is clustered based on the K-means algorithm. The size of the cluster is equal to the passenger attribute category, and the number of advertisement appearances in each cluster is batch normalized. 对聚类结果中的同类广告进行加权，重新排序；定义广告关注度阈值，保留分值高于的广告项，生成新的广告目录。Weight and re-rank similar ads in clustering results; define ad attention thresholds , the retention score is higher than Ad items to generate a new ad catalog. 2.根据权利要求1所述的交互式电梯广告智能推荐方法，其特征在于，所述寻找目标乘客的方法包括：2. The interactive elevator advertisement intelligent recommendation method according to claim 1, characterized in that the method of finding target passengers comprises: 乘客进入电梯后，基于深度学习算法检测出乘客位置，基于opencv人眼检测模型haarcascade_eye.xml检测人眼位置，根据人眼状态库进行模板匹配，对乘客进行人眼关注位置检测，其中人眼状态库包括向正前方、左、右、上、下关注5种类型，由于摄像头位于广告屏上方，当乘客面向广告屏时，认为人眼状态为正前方，建立目标乘客模型：After the passenger enters the elevator, the passenger position is detected based on the deep learning algorithm, the eye position is detected based on the opencv eye detection model haarcascade_eye.xml, and template matching is performed according to the eye state library to detect the passenger's eye focus position. The eye state library includes five types: forward, left, right, up, and down. Since the camera is located above the advertising screen, when the passenger faces the advertising screen, the eye state is considered to be forward, and the target passenger model is established: 其中，为具体乘客，为乘客检测框面积，为乘客面向正前方置信度，分别为对应因素的权重因子，为最后得分，据此排序，确认乘客优先级，优先级较高的为主要目标乘客，基于该乘客属性推荐广告目录。in, For specific passengers, is the passenger detection box area, Confidence for passengers facing forward, are the weight factors of the corresponding factors, The final score is used for sorting and confirming the passenger priority. Passengers with higher priority are the main target passengers, and advertising catalogs are recommended based on the passenger attributes. 3.根据权利要求1所述的交互式电梯广告智能推荐方法，其特征在于，所述对目标乘客推荐广告的方法包括：3. The interactive elevator advertisement intelligent recommendation method according to claim 1, characterized in that the method of recommending advertisements to target passengers comprises: 基于情景感知的FP-growth推荐算法，生成目标乘客广告目录；Generate target passenger advertising catalog based on context-aware FP-growth recommendation algorithm; 目标乘客的广告目录播放结束，摄像头依然检测到目标乘客，若没有其他目标乘客，广告目录中的内容轮次播放，若存在其他目标乘客，则播放对应的广告目录；Target passengers The advertising catalogue ends, and the camera still detects the target passenger If there are no other target passengers, the contents in the advertising catalog will be played in turn. If there are other target passengers , then play Corresponding advertising catalogue; 目标乘客的广告目录没有播放完，目标乘客出梯，若没有其他目标乘客，广告屏恢复默认播放，若存在其他目标乘客，则根据优先级依次播放对应的广告目录。Target passengers The advertising catalogue has not been played to the end, and the target passengers When exiting the elevator, if there are no other target passengers, the advertising screen will resume the default playback. If there are other target passengers , then play in order according to priority The corresponding advertising catalog. 4.一种交互式电梯广告智能推荐系统，所述系统用于实现权利要求1-3中任一项所述的交互式电梯广告智能推荐方法，其特征在于，包括：4. An interactive elevator advertisement intelligent recommendation system, the system being used to implement the interactive elevator advertisement intelligent recommendation method according to any one of claims 1 to 3, characterized in that it comprises: 数据中控中心，负责整个电梯集群的数据信息的计算、存储以及传输；The data control center is responsible for the calculation, storage and transmission of data information of the entire elevator cluster; 信息采集模块，用于对电梯内的乘客属性信息以及交互信息进行采集；An information collection module is used to collect passenger attribute information and interaction information in the elevator; 乘客属性分析模块，基于深度学习算法获得乘客属性信息；Passenger attribute analysis module, which obtains passenger attribute information based on deep learning algorithm; 数据库模块，将乘客与触摸广告屏进行交互的情景信息以数据的形式存放；The database module stores the situational information of the passengers' interaction with the touch advertising screen in the form of data; 全局推荐模块，基于情景感知的FP-growth推荐算法分析数据库信息，以整个电梯集群的数据为基础，对乘客进行广告的智能推荐；The global recommendation module analyzes database information based on the context-aware FP-growth recommendation algorithm and makes intelligent advertising recommendations to passengers based on the data of the entire elevator cluster. 局部矫正模块，是在全局推荐的基础之上，基于K-means进一步分析具体点位信息做出综合判断。The local correction module is based on the global recommendation and further analyzes the specific point information based on K-means to make a comprehensive judgment.
CN202111422668.1A 2021-11-26 2021-11-26 Interactive elevator advertising intelligent recommendation method and system Active CN114418605B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111422668.1A CN114418605B (en) 2021-11-26 2021-11-26 Interactive elevator advertising intelligent recommendation method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111422668.1A CN114418605B (en) 2021-11-26 2021-11-26 Interactive elevator advertising intelligent recommendation method and system

Publications (2)

Publication Number Publication Date
CN114418605A CN114418605A (en) 2022-04-29
CN114418605B true CN114418605B (en) 2025-06-20

Family

ID=81266378

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111422668.1A Active CN114418605B (en) 2021-11-26 2021-11-26 Interactive elevator advertising intelligent recommendation method and system

Country Status (1)

Country Link
CN (1) CN114418605B (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN209625270U (en) * 2019-03-08 2019-11-12 福建省软众数字科技股份有限公司 A kind of device based on user preferences recommendation information
CN111382154A (en) * 2018-12-29 2020-07-07 赫狮网络科技(上海)有限公司 Advertisement matching system based on FP tree and most frequent item and working method thereof
CN113627984A (en) * 2020-12-10 2021-11-09 北京交通大学 An intelligent elevator advertising screen system and advertising delivery method based on scene perception

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6665669B2 (en) * 2000-01-03 2003-12-16 Db Miner Technology Inc. Methods and system for mining frequent patterns
CN108717654B (en) * 2018-05-17 2022-03-25 南京大学 A multi-e-commerce cross recommendation method based on clustering feature migration
CN110378752A (en) * 2019-07-26 2019-10-25 京东方科技集团股份有限公司 Advertisement recommended method, device, electronic equipment and storage medium
CN112036948A (en) * 2020-09-01 2020-12-04 杭州岁丰信息技术有限公司 Precise media information delivery system for elevator and control method thereof
CN113657933A (en) * 2021-08-16 2021-11-16 浙江新再灵科技股份有限公司 Preparation method of elevator advertisement recommendation data

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111382154A (en) * 2018-12-29 2020-07-07 赫狮网络科技(上海)有限公司 Advertisement matching system based on FP tree and most frequent item and working method thereof
CN209625270U (en) * 2019-03-08 2019-11-12 福建省软众数字科技股份有限公司 A kind of device based on user preferences recommendation information
CN113627984A (en) * 2020-12-10 2021-11-09 北京交通大学 An intelligent elevator advertising screen system and advertising delivery method based on scene perception

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
基于Fp树的加权频繁模式挖掘算法;陈文;计算机工程;20120331;第38卷(第6期);第63-65页 *

Also Published As

Publication number Publication date
CN114418605A (en) 2022-04-29

Similar Documents

Publication Publication Date Title
Aran et al. Broadcasting oneself: Visual discovery of vlogging styles
US20140068676A1 (en) Method and system for video collection management, invalid video replacement and continuous video replay
CN113158023A (en) Public digital life accurate classification service method based on mixed recommendation algorithm
CN111310019A (en) Information recommendation method, information processing method, system and equipment
CN110430476A (en) Direct broadcasting room searching method, system, computer equipment and storage medium
CN107071578A (en) IPTV program commending methods
CN105100894A (en) Face automatic labeling method and system
CN102165464A (en) Method and system for automated annotation of persons in video content
CN117440182B (en) Intelligent recommendation method and system based on video content analysis and user labels
CN112685596B (en) Video recommendation method and device, terminal, storage medium
CN113474769A (en) Image retrieval device and supervised data extraction method
US20110179003A1 (en) System for Sharing Emotion Data and Method of Sharing Emotion Data Using the Same
Akiyama et al. Proposal and Evaluation of Serendipitous Recommendation Method Using General Unexpectedness.
CN107885857B (en) A search result page user behavior pattern mining method, device and system
CN115936939A (en) Project recommendation method and system based on multiple preference degrees
CN109391829A (en) Video gets position analysis system, analysis method and storage media ready
CN114529352A (en) Intelligent marketing promotion method and system based on big data analysis
CN111538751A (en) Tagged user portrait generation system and method for Internet of things data
CN111581435A (en) Video cover image generation method and device, electronic equipment and storage medium
CN119671650A (en) A brand promotion method and system based on user positioning
CN114418605B (en) Interactive elevator advertising intelligent recommendation method and system
US20220075804A1 (en) Method and device for providing guide information for enhancement of artist's reputation
CN114154063A (en) Internet big data label analysis-based paired activity recommendation method
CN118042229B (en) Interactive network television service method and system
CN119322860B (en) Method and system for managing text creation display based on artificial intelligence

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
TA01 Transfer of patent application right

Effective date of registration: 20230216

Address after: 610000 High-tech Zone, Chengdu City, Sichuan Province, No. 99, No. 1, No. 2, No. 15, No. 1, No. 1505, No. 1, No. 1, No. 1, No. 1, No. 1, No. 1, No. 1, No. 1, No. 1, No

Applicant after: CHENGDU XINCHAO MEDIA GROUP Co.,Ltd.

Address before: No.1505, 15th floor, unit 2, building 1, No.99, Jinhui West 1st Street, Chengdu hi tech Zone, 610000, Sichuan Province

Applicant before: CHENGDU XINCHAO MEDIA GROUP Co.,Ltd.

Applicant before: Chengdu Baixin Zhilian Technology Co.,Ltd.

TA01 Transfer of patent application right
GR01 Patent grant
GR01 Patent grant