WO2020056915A1 - 顾客到访分析方法、装置及存储介质 - Google Patents

顾客到访分析方法、装置及存储介质 Download PDF

Info

Publication number
WO2020056915A1
WO2020056915A1 PCT/CN2018/117286 CN2018117286W WO2020056915A1 WO 2020056915 A1 WO2020056915 A1 WO 2020056915A1 CN 2018117286 W CN2018117286 W CN 2018117286W WO 2020056915 A1 WO2020056915 A1 WO 2020056915A1
Authority
WO
WIPO (PCT)
Prior art keywords
customer
recognition result
time
result data
face recognition
Prior art date
Application number
PCT/CN2018/117286
Other languages
English (en)
French (fr)
Inventor
刘泽许
Original Assignee
图普科技（广州）有限公司
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 图普科技（广州）有限公司 filed Critical 图普科技（广州）有限公司
Priority to EP18934444.3A priority Critical patent/EP3855343A4/en
Publication of WO2020056915A1 publication Critical patent/WO2020056915A1/zh

Links

Images

Classifications

    • 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
    • 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
    • G06Q30/0201—Market modelling; Market analysis; Collecting market data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/40—Scenes; Scene-specific elements in video content
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects

Definitions

  • the present disclosure relates to the technical field of data analysis and processing, and in particular, to a customer visit analysis method, device, and storage medium.
  • the present disclosure provides a customer visit analysis method, device, and storage medium.
  • a first aspect of the present disclosure provides a customer visit analysis method, including:
  • the customer ID corresponds to a plurality of the photographed time periods, determining whether the time difference between two adjacent time periods in the plurality of photographed time periods is not greater than the stay duration threshold;
  • a time period can be understood as a time stamp.
  • the method further includes: clustering the face recognition result data of all customers according to the store to obtain the face recognition of the customers in each store Results data.
  • the method before clustering the face recognition result data of all customers according to the store, the method further includes:
  • packaging and sending the face recognition result data transmitted from the upper-layer service to a message queue includes:
  • preprocessing the face recognition result data includes completing the data verification, organizing the data structure, and compressing the data;
  • the reading and processing the face recognition result data from the message queue includes: decompressing the face recognition result data and arranging the data structure;
  • the processed face recognition result data is stored in a time-series database according to a preset saving rule.
  • the determining whether the time difference between two adjacent time periods in the plurality of captured time periods is not greater than the stay duration threshold includes:
  • the day is divided into a plurality of consecutive preset time periods, and each of the photographed time periods corresponding to the ID of each customer is mapped to the plurality of consecutive preset time periods.
  • Continuous time points in each photographed time period corresponding to the same customer ID are projected into consecutive preset time periods, where two consecutive preset time periods in a plurality of consecutive preset time periods
  • the time interval length of a segment is the same as the dwell duration threshold.
  • the method further includes: determining a stay time of each visit behavior of the ID of each customer.
  • the face recognition result data includes a date, and after determining the arrival time and the departure time of each store behavior, the method further includes:
  • the face recognition result data further includes a shooting area where the customer is photographed, and the method further includes:
  • the obtaining facial recognition result data of customers in the store includes:
  • the face data corresponding to the face recognition result data is the same as the face data of the employee stored in advance, it is determined that the face recognition result data corresponds to the face recognition result data of the employee;
  • employees include, but are not limited to, courier, cleaner, security, etc.
  • the method further includes:
  • the time period is regarded as a one-time visit behavior of the corresponding customer.
  • a second aspect of the present disclosure provides a customer visit analysis device, including:
  • a data acquisition module configured to acquire facial recognition result data of customers in the store, where the facial recognition result data includes the ID of the customer and the time period during which the customer was photographed;
  • a customer analysis module configured to cluster the face recognition result data of the customers in the store according to the customer's ID to obtain at least one of the photographed time periods corresponding to the ID of each customer;
  • a data processing module configured to determine whether the time difference between two adjacent time periods in the plurality of captured time periods is not greater than the stay when the customer ID corresponds to a plurality of the captured time periods A duration threshold; and when the time difference is not greater than the stay duration threshold, determining that the two adjacent time periods correspond to the same visit to the store;
  • the visit determination module is used to determine the arrival time and the departure time of each visit to the store.
  • the device further includes a data classification module for clustering the face recognition result data of all customers according to the stores to obtain the face recognition result data of the customers in each store.
  • a data classification module for clustering the face recognition result data of all customers according to the stores to obtain the face recognition result data of the customers in each store.
  • the data classification module is further configured to, before clustering the face recognition result data of all customers according to the store, package the face recognition result data transmitted from the upper-layer service and send it to the message queue; Read face recognition result data from the message queue for processing.
  • the data classification module is configured to preprocess the face recognition result data after obtaining the face recognition result data from the upper-layer service, and the preprocessing includes completing data verification, Organize the data structure and compress the data; pack the pre-processed face recognition result data into a message queue; save the processed face recognition result data into a time series database according to a preset save rule.
  • the data processing module includes:
  • a time mapping module configured to divide each day into multiple consecutive preset time periods, and map each of the photographed time periods corresponding to each customer's ID to the multiple consecutive preset time periods ;
  • An aggregation statistics module configured to determine a minimum time point and a maximum time point corresponding to the ID of the customer in each preset time period;
  • a threshold value judging module configured to determine a minimum time point of a preset preset time period and a preset preset time period of two adjacent preset time periods in a preset time period mapped with the captured time period. Whether the difference between the maximum time points of the time period is not greater than the stay duration threshold.
  • the time mapping module is configured to project consecutive time points in each photographed time period corresponding to the ID of the same customer into consecutive preset time periods, where a plurality of consecutive It is assumed that a time interval length between two adjacent preset time periods in the time period is the same as the stay duration threshold.
  • the visit determination module is further configured to determine a stay duration of each visit to the store by the ID of each customer.
  • the face recognition result data includes a date
  • the visit determination module is further configured to determine an ID of each customer after determining the arrival time and the departure time of each visit to the store. The number of visits corresponding to the date and the total length of stay.
  • a third aspect of the present disclosure provides an electronic device including a processor, a memory, and a bus.
  • the memory stores machine-readable instructions executable by the processor.
  • the processor and The memories communicate through a bus, and when the machine-readable instructions are executed by the processor, the method according to the first aspect is performed.
  • a fourth aspect of the present disclosure provides a storage medium on which a computer program is stored, and when the computer program is executed by a processor, the method according to the first aspect is performed.
  • a fifth aspect of the present disclosure provides a computer program product that, when run on a computer, causes the computer to execute the method described in the first aspect.
  • the customer visit analysis method, device and storage medium can analyze the customer's visit behavior based on the results of face recognition. Obtain the face recognition result data in the store, and then analyze the visit information of the customers in the store, including the customer's arrival time and departure time, to achieve an accurate analysis of the customer's visit information.
  • FIG. 1 shows a schematic diagram of an application scenario provided by the present disclosure.
  • FIG. 2 shows a flowchart of a customer visit analysis method provided by the present disclosure
  • FIG. 3 shows another flowchart of a customer visit analysis method provided by the present disclosure
  • FIG. 4 shows a schematic diagram of a customer visit analysis device provided by the present disclosure.
  • Electronic equipment-10 memory-11; processor-12; network module-13; customer visit analysis device-14; data acquisition module-100; customer analysis module-101; data processing module-102; visit determination module- 103.
  • FIG. 1 it is a block diagram of an electronic device 10 provided by the present disclosure.
  • the electronic device 10 in the present disclosure may be a device having a data processing function, such as a server, a personal computer, a tablet computer, a smart phone, or the like.
  • the electronic device 10 includes a memory 11, a processor 12, a network module 13, and a customer visit analysis device 14.
  • the memory 11, the processor 12, and the network module 13 are directly or indirectly electrically connected to each other to implement data transmission or interaction. For example, these components can be electrically connected to each other through one or more communication buses or signal lines.
  • the memory 11 stores a customer visit analysis device 14.
  • the customer visit analysis device 14 includes at least one software function module that can be stored in the memory 11 in the form of software or firmware.
  • the processor 12 passes Software programs and modules stored in the memory 11 are run, such as the customer visit analysis device 14 in the present disclosure, to execute various functional applications and data processing, that is, to implement the data processing method performed by the electronic device 10 in the present disclosure.
  • the memory 11 may be, but is not limited to, Random Access Memory (RAM), Read-Only Memory (ROM), and Programmable Read-Only Memory (PROM). , Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Read-Only Memory (Electric Erasable Programmable Read-Only Memory, EEPROM), etc.
  • RAM Random Access Memory
  • ROM Read-Only Memory
  • PROM Programmable Read-Only Memory
  • EPROM Erasable Programmable Read-Only Memory
  • EEPROM Electrically Erasable Read-Only Memory
  • the processor 12 may be an integrated circuit chip and has data processing capabilities.
  • the aforementioned processor 12 may be a general-purpose processor, including a central processing unit (CPU), a network processor (Network Processor), and the like. Various methods, steps, and logic block diagrams provided in the present disclosure may be implemented or performed.
  • a general-purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
  • the network module 13 is configured to establish a communication connection between the electronic device 10 and an external communication terminal through a network, and implement a network signal and data sending and receiving operation.
  • the network signal may include a wireless signal or a wired signal.
  • the structure shown in FIG. 1 is merely schematic, and the electronic device 10 may further include more or fewer components than those shown in FIG. 1, or have a different configuration from that shown in FIG. 1.
  • Each component shown in FIG. 1 may be implemented by hardware, software, or a combination thereof.
  • the electronic device 10 may further include a face recognizer.
  • the electronic device 10 performs face recognition through the face recognizer included in the electronic device 10 to obtain face recognition result data.
  • the processor 12 in the electronic device 10 may be communicatively connected with a face recognizer independent of the electronic device 10, so as to obtain the face recognition result data recognized and obtained by the face recognizer.
  • This embodiment provides a customer visit analysis method.
  • the method includes the following steps, which can be executed by the processor 12 in the electronic device 10 shown in FIG. 1.
  • Step 110 Obtain facial recognition result data of customers in the store.
  • the face recognition result data transmitted by the camera in real time includes the employee's face data and the customer's face data. Therefore, if you want to obtain the customer's face recognition result data, you need to associate the employee's face recognition result data with the customer. To distinguish the face recognition result data.
  • One implementation method is to establish a database of employee information and facial data corresponding to employees in advance. When real-time face recognition result data is transmitted, it is determined whether the face data corresponding to the face recognition result data is the same as the face data of the employee stored in advance.
  • the piece of face recognition result data corresponds to the employee's face recognition result data. If the face data corresponding to the face recognition result data transmitted is different from the face data of the employee stored in advance, it is determined that the piece of face recognition result data corresponds to the face recognition result data of the customer in the store.
  • the face recognition result data includes a customer's ID (identification identification number) and a time period during which the customer was photographed.
  • each face data is converted into an ID through the face recognition technology, and the face data of different people is distinguished by the ID, which is convenient for each Customer visit behavior is analyzed.
  • the data transmitted back by the camera can be a video stream, which recognizes the face in the video stream screen, and can obtain the time period that each customer was photographed in the camera.
  • the video stream can also be understood as a stream of pictures, and the camera transmits the captured pictures in real time to recognize the faces in the pictures. For example, if a customer is captured by the first camera from 9:05 to 9:10, the time period in which a customer's face recognition result data is captured is correspondingly from 9:05 to 9:10.
  • Step 120 Cluster the face recognition result data of the customers in the store according to the customer ID.
  • a piece of face recognition result data includes the customer's ID and the time period when the customer was photographed. Of course, there can also be other information, for example, it can also include the area where the customer was photographed. If there are multiple stores, it can include the Information on the stores that were photographed.
  • clustering is performed according to the customer IDs in the face recognition result data to obtain one or more photographed time periods corresponding to each customer ID, thereby achieving the face recognition results. Clustering of data.
  • each customer's ID has multiple time periods that are photographed. When there is only one camera in the store, and when the customer has been photographed by the camera, there may be only one time period that is photographed.
  • Step 130 Determine whether two adjacent time periods correspond to the same visit to the store.
  • this time period is deemed to correspond to the customer's one-time visit behavior. If the customer's ID corresponds to multiple photographed time periods, it is necessary to determine multiple photographed time periods. Whether the time difference between two adjacent time periods in the time period is not greater than a preset stay duration threshold.
  • the customer when the area in the store is large, the customer may be located in some dead-end areas of the store within a certain period of time after being photographed, and the camera does not capture the person of the customer Face, therefore, in the face recognition result data, there may be a certain time blank in a plurality of time periods in which the customer is photographed. For example, if the customer is photographed from 9:05 to 9:10 and the customer is photographed from 9:20 to 9:40, the time difference between the two adjacent time periods is 10 minutes.
  • the threshold of the length of stay corresponding to the customer's identity is set in advance, for example, 30 minutes, that is, when the time difference between the two adjacent time periods of the customer is 10 minutes, because the length of the stay is not greater than the preset threshold of 30 minutes, The time periods captured in these two data correspond to the same visit behavior.
  • customers when purchasing goods in the store, they may be captured by multiple cameras at the same time. For example, customers are captured by the first camera from 9:05 to 9:20, and captured by the second camera from 9:10 to 9:30. In this case, the time difference is obviously not due to the overlapping time periods of customers. Greater than the preset length of stay threshold, it is also considered a one-time visit.
  • the stay duration threshold can be flexibly set, for example, different stay duration thresholds can be set for different stores. For example, if the stay duration threshold is set according to the area size of each store, the stay duration threshold set for a store with a larger area may be larger than the stay duration threshold set for a store with a smaller area. As another example, different dwell duration thresholds can be set for different customer IDs. For example, if historical data collection and analysis is performed for each customer, for customers who have stayed in the store for a longer period of time, the threshold for the length of stay may be greater than for customers who have been analyzed for a shorter stay in the store The duration of the threshold. For another example, for different stores and customers, the stay duration threshold may be set to the same fixed duration.
  • Step 140 Determine the arrival time and the departure time of each visit.
  • the arrival time and departure time of each arrival behavior can be determined according to the time period captured in the face recognition result data, so as to obtain each time in the store. Customer visit information.
  • the visit behavior of customers can be analyzed for a certain store. Collect the images in the store through the camera. After obtaining the data of the face recognition results in the store, you can analyze the visit information of each customer in each store, including the customer's arrival time and departure time.
  • step 110 it may also include:
  • the face recognition result data of all customers are clustered according to the stores, and the face recognition result data of customers in each store is obtained.
  • the face recognition result data of customers in each store is determined according to the store, and the face recognition result data in each store is analyzed separately, which can effectively improve the analysis Efficiency, reducing the time of data analysis.
  • the face recognition result data transmitted from the upper-level service needs to be packaged and sent to the message.
  • Queue when processing the face recognition result data, read data from the message queue.
  • a face recognition device may be integrated into the camera. After the camera captures the pictures in the store, the face recognition device performs face recognition processing on the pictures collected by the camera to obtain face recognition result data.
  • the processor 12 of the electronic device 10 includes a message queue for buffering face recognition result data.
  • the face recognizer After the face recognizer obtains the face recognition result data, it sends the obtained face recognition result data to the processor 12, and the processor 12 packs the face recognition result data transmitted from the face recognizer and sends it to the message queue.
  • the processor 12 processes the face recognition result data, the processor 12 reads the face recognition result data from the message queue and processes them in the order of the cache. This ensures data analysis reliability.
  • this embodiment introduces a message queue as a buffer for data processing.
  • each data can be stored in a database.
  • an asynchronous mechanism is used to process big data, which can effectively reduce the load of data processing by the database.
  • step 130 the step of determining whether the time difference between two adjacent time periods in the multiple captured time periods is not greater than the stay duration threshold can be implemented with reference to FIG. 3.
  • a detailed implementation of this step is:
  • Step 210 Divide each day into multiple consecutive preset time periods, and map each photographed time period corresponding to each customer's ID to multiple consecutive preset time periods.
  • Continuous time points in each photographed time period corresponding to the ID of the same customer may be projected into continuous preset time periods.
  • the interval length between two consecutive preset time periods in the multiple consecutive preset time periods may be the same as the stay duration threshold.
  • Step 220 Determine a minimum time point and a maximum time point corresponding to the customer ID in each preset time period.
  • the minimum time point and the maximum time point among all consecutive time points in each preset time period are obtained through the aggregation function.
  • Step 230 In the preset time period mapped with the captured time period, determine the minimum time point of the next preset time period and the maximum time of the previous preset time period in two adjacent preset time periods. Whether the difference between the points is not greater than the stay duration threshold.
  • the minimum time point and the maximum time point in each preset time period of each photographed time period corresponding to each customer are obtained through the above algorithm analysis, so as to determine the customer's once-to-store behavior through analysis calculation.
  • the processor 12 in the electronic device 10 pre-processes the face recognition result data after obtaining the face recognition result data from the upper-layer service.
  • the pre-processing may include completing the verification of the data, organizing the data structure, and compressing the data.
  • the pre-processed face recognition result data is packaged and sent to a message queue.
  • Post-processing can include decompressing the face recognition result data and arranging the data structure.
  • the face recognition result data after post-processing is saved to the time series database according to certain preset saving rules, for example, according to the store or according to the shooting area, and customer visit analysis is performed on these data.
  • the time-series database can be understood as that the face recognition result data in the database is stored in chronological order.
  • the time sequence here mainly refers to that each piece of data has a time field. When obtaining data, you can obtain it according to certain conditions according to this time field, for example, to obtain data in order according to the entry time, or to obtain data in reverse order according to the departure time.
  • the time-series database and the aforementioned database in the present disclosure may refer to the same database. For example, it may be a Structured Query Language (SQL) type database, or a non-relational database NoSQL.
  • SQL Structured Query Language
  • the obtained face recognition result data is the face recognition result data of all the stores in the country and the pictures taken by each camera in each shooting area in the store. These data are used to distinguish the data of different stores, and at the same time, useless data is filtered.
  • the useless data can include the face data of the identified employees and some blurred face data.
  • the customer visit analysis is performed on the face recognition result data in each store to construct data analysis conditions.
  • the data analysis conditions may include fields such as the ID of the customer to be analyzed, the data of which day, and the customer's staying time threshold. Construct a query aggregation statement of time series data according to the data analysis conditions, and initiate a query aggregation operation request to the time series database.
  • the time series database For example, if you want to analyze the visit behavior of a certain customer ’s ID on a certain day, then query the time series database for multiple photographed time periods corresponding to that customer ’s ID on that day, and compare the time points in these time periods. It is arranged and corresponding to each preset unit time, the length of the unit time interval is the threshold of the customer's stay time in the data analysis conditions. If the customer's stay time threshold is 30 minutes, the 24 hours a day is divided into 30 minutes, and the minimum time point and maximum time for each preset unit time that the customer is photographed are obtained through the aggregation function. point.
  • each preset unit time interval is also 30 minutes. If the customer is captured by the first camera at 9:05 to 9:20, it will be at 9:10 to 9 : 25 was captured by the second camera, and was captured by the third camera from 9:45 to 9:50, then the time points corresponding to these time periods are arranged according to time.
  • a preset unit time from 9:00 to 9:30, and another preset unit time from 9:30 to 10:00, and the minimum time point from 9:00 to 9:30 is obtained through the aggregation function. 9:05 and the maximum time point 9:25, get the minimum time point 9:45 and the maximum time point 9:50 from 9:30 to 10:00.
  • the minimum time point and the maximum time point in each preset unit time corresponding to each customer's ID After determining the minimum time point and the maximum time point in each preset unit time corresponding to each customer's ID through aggregation statistics, determine the minimum time point in the later preset unit time and the previous preset unit time Whether the difference between the maximum time points within is not greater than a preset dwell duration threshold. If it is not greater than, for example, the difference between 9:45 and 9:25 is 20 minutes, and the length of stay is not greater than the threshold of 30 minutes, it is determined as a visit behavior of the customer, and the arrival time and The check-out time is to confirm that the customer in the above example arrives at 9:05 and leaves at 9:50. If the difference is greater than the preset dwell time threshold, it is determined as another visiting behavior.
  • the visit behavior of each customer in different stores can be analyzed, and the shooting area can be added to the above data analysis conditions, so that the visit behavior of customers in each area can be obtained by region analysis, such as freshness Area, clothing area, living area, etc.
  • the length of stay, the number of visits and the total length of stay of each day can be obtained. Further, according to the daily analysis data, the total number of visits per week, month, and year for each customer, the length of stay for each visit, and the total length of stay can also be obtained.
  • Integrate the data obtained during the entire analysis process such as customer ID, store, shooting area, date, arrival time of each visit behavior, departure time, threshold length of stay, length of stay, number of visits, and total Data such as length of stay are integrated, and these data are saved to the database according to preset storage conditions, such as by store or region.
  • This embodiment provides a customer visit analysis device. Referring to FIG. 4, it includes:
  • the data obtaining module 100 is configured to obtain face recognition result data of a customer in the store, where the face recognition result data includes an ID of the customer and a time period during which the customer was photographed.
  • step S110 in FIG. 2 For the specific implementation of the data acquisition module 100, reference may be made to the related description of step S110 in FIG. 2, and therefore will not be repeated here.
  • the customer analysis module 101 is configured to perform clustering on the face recognition result data of the customers in the store according to the IDs of the customers to obtain at least one of the photographed time periods corresponding to the IDs of each customer.
  • step S120 in FIG. 2 For the specific implementation manner of the customer analysis module 101, reference may be made to the related description of step S120 in FIG. 2, and details are not described herein.
  • the data processing module 102 is configured to determine whether the time difference between two adjacent time periods in the plurality of time periods is not greater than a plurality of the time periods when the customer ID corresponds to the plurality of time periods A dwell time threshold; and when the time difference is not greater than the dwell time threshold, it is determined that the two adjacent time periods correspond to the same visit behavior.
  • step S130 in FIG. 2 For the specific implementation of the data processing module 102, reference may be made to the related description of step S130 in FIG. 2, and therefore will not be repeated here.
  • the visit determination module 103 is configured to determine the arrival time and the departure time of each store behavior.
  • step S140 for the specific implementation of the visit determination module 103, reference may be made to the related description of step S140 in FIG. 2, and therefore will not be repeated here.
  • the device further includes a data classification module for clustering the face recognition result data of all customers according to the stores to obtain the face recognition result data of the customers in each store.
  • a data classification module for clustering the face recognition result data of all customers according to the stores to obtain the face recognition result data of the customers in each store.
  • the data classification module is further configured to, before clustering the face recognition result data of all customers according to the store, package the face recognition result data transmitted from the upper-layer service and send it to the message queue; Read face recognition result data from the message queue for processing.
  • the data classification module is configured to preprocess the face recognition result data after obtaining the face recognition result data from the upper-layer service, and the preprocessing includes completing data verification, Organize the data structure and compress the data; pack the pre-processed face recognition result data into a message queue; save the processed face recognition result data into a time series database according to a preset save rule.
  • the data processing module includes: a time mapping module, configured to divide each day into a plurality of consecutive preset time periods, and map each of the captured time periods corresponding to the ID of each customer To the plurality of consecutive preset time periods; an aggregate statistics module for determining a minimum time point and a maximum time point corresponding to the ID of the customer in each preset time period; a threshold judgment module for judging whether the mapping has In the preset time period of the photographed time period, between a minimum time point of a later preset time period in two adjacent preset time periods and a maximum time point of a previous preset time period Whether the difference is not greater than the stay duration threshold.
  • a time mapping module configured to divide each day into a plurality of consecutive preset time periods, and map each of the captured time periods corresponding to the ID of each customer To the plurality of consecutive preset time periods
  • an aggregate statistics module for determining a minimum time point and a maximum time point corresponding to the ID of the customer in each preset time period
  • a threshold judgment module for
  • the time mapping module is configured to project consecutive time points in each photographed time period corresponding to the ID of the same customer into consecutive preset time periods, where a plurality of consecutive It is assumed that a time interval length between two adjacent preset time periods in the time period is the same as the stay duration threshold.
  • the visit determination module 103 is further configured to determine the length of stay of each visit to the store by the ID of each customer.
  • the face recognition result data includes a date
  • the visit determination module 103 is further configured to determine the time of each customer after determining the arrival time and departure time of each store behavior. The number of visits corresponding to the ID on the date and the total length of stay.
  • the present disclosure also provides an electronic device including a processor, a memory, and a bus.
  • the memory stores machine-readable instructions executable by the processor.
  • the processor and the The memories communicate through a bus, and when the machine-readable instructions are executed by the processor, the method according to the first embodiment is performed.
  • the present disclosure also provides a storage medium on which a computer program is stored.
  • the computer program is executed by a processor, the method described in the first embodiment is executed.
  • the present disclosure also provides a computer program product that, when run on a computer, causes the computer to execute the method described in the first embodiment.
  • each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, which contains one or more components for implementing a specified logical function Executable instructions.
  • the functions marked in the blocks may also occur in a different order than those marked in the drawings.
  • each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts can be implemented in a dedicated hardware-based system that performs the specified function or action. , Or it can be implemented with a combination of dedicated hardware and computer instructions.
  • the functional modules in the various embodiments of the present disclosure may be integrated together to form an independent part, or each of the modules may exist alone, or two or more modules may be integrated to form an independent part.
  • the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium.
  • the technical solution of the present disclosure is essentially a part that contributes to the existing technology or a part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium, including Several instructions are used to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in various embodiments of the present disclosure.
  • the foregoing storage media include: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disks or optical disks and other media that can store program codes .
  • relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that these entities or operations There is any such actual relationship or order among them.
  • the terms "including”, “comprising”, or any other variation thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or device that includes a series of elements includes not only those elements but also those that are not explicitly listed Or other elements inherent to such a process, method, article, or device. Without more restrictions, the elements defined by the sentence “including a " do not exclude the existence of other identical elements in the process, method, article, or equipment including the elements.
  • the customer visit analysis method, device and storage medium provided by the present disclosure can analyze the customer's visit behavior based on the face recognition results, and then analyze the visit information of the customer in the store, including the customer's arrival time and departure Time to achieve accurate analysis of customer visit information.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • Accounting & Taxation (AREA)
  • Development Economics (AREA)
  • Theoretical Computer Science (AREA)
  • Finance (AREA)
  • Strategic Management (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Health & Medical Sciences (AREA)
  • General Business, Economics & Management (AREA)
  • Game Theory and Decision Science (AREA)
  • Economics (AREA)
  • Data Mining & Analysis (AREA)
  • Marketing (AREA)
  • General Health & Medical Sciences (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Human Computer Interaction (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Collating Specific Patterns (AREA)
  • Image Analysis (AREA)

Abstract

本公开提供了顾客到访分析方法、装置及存储介质，顾客到访分析方法包括：获取门店内顾客的人脸识别结果数据，人脸识别结果数据包括顾客的ID以及顾客被拍摄到的时间段；根据顾客的ID对门店内顾客的人脸识别结果数据进行聚类，获得每个顾客的ID对应的至少一个被拍摄到的时间段；若顾客的ID对应多个被拍摄到的时间段，则判断多个被拍摄到的时间段中相邻两个时间段之间的时间差，是否不大于停留时长阈值；若不大于，则确定相邻两个时间段对应同一次到店行为；确定每次到店行为的到店时间以及离店时间。本公开提供的顾客到访分析方法能够通过人脸识别结果分析出每个顾客的到访信息，包括该顾客的到店时间以及离店时间。

Description

顾客到访分析方法、装置及存储介质
相关申请的交叉引用
本公开要求于2018年09月18日提交中国专利局的申请号为CN201811093386X，名称为“顾客到访分析方法、装置及存储介质”的中国专利申请的优先权，其全部内容通过引用结合在本公开中。
技术领域
本公开涉及数据分析处理技术领域，具体而言，涉及一种顾客到访分析方法、装置及存储介质。
背景技术
随着人工智能的日益发展，机器识别人脸的准确度已经与人类的识别准确度相差无几，基于人脸识别的应用也越来越多。在商业智能的场景下，基于人脸识别的技术，摄像头拍下的每位到访的访客基本上可以被正确辨识，但如何根据这些被辨识到的访客的人脸准确分析出每位访客的到访记录，现阶段还未提出有效的方案。
发明内容
本公开提供一种顾客到访分析方法、装置及存储介质。
本公开第一方面提供一种顾客到访分析方法，包括：
获取门店内顾客的人脸识别结果数据，所述人脸识别结果数据包括所述顾客的ID以及顾客被拍摄到的时间段；
根据顾客的ID对所述门店内顾客的人脸识别结果数据进行聚类，获得每个顾客的ID对应的至少一个所述被拍摄到的时间段；
若顾客的ID对应多个所述被拍摄到的时间段，则判断多个所述被拍摄到的时间段中相邻两个时间段之间的时间差，是否不大于停留时长阈值；
若不大于，则确定所述相邻两个时间段对应同一次到店行为；
确定每次到店行为的到店时间以及离店时间。本公开中，时间段可以理解为时间戳。
可选地，在所述获取门店内顾客的人脸识别结果数据之前，所述方法还包括：将所有顾客的人脸识别结果数据按照门店进行聚类，获得每个门店内顾客的人脸识别结果数据。
可选地，在将所有顾客的人脸识别结果数据按照门店进行聚类之前，所述方法还包括：
对上层服务传来的人脸识别结果数据进行打包并发送至消息队列；
从所述消息队列中读取人脸识别结果数据进行处理。
可选地，所述对上层服务传来的人脸识别结果数据进行打包并发送至消息队列，包括：
在获取到上层服务传来的人脸识别结果数据后，对该人脸识别结果数据进行预处理，所述预处理包括完成数据的校验、整理数据结构以及压缩数据；
将预处理后的人脸识别结果数据打包发送至消息队列中；
所述从所述消息队列中读取人脸识别结果数据进行处理，包括：对人脸识别结果数据进行解压以及整理数据结构；
将进行处理后的人脸识别结果数据按照预设保存规则保存至时序数据库中。
可选地，所述判断多个所述被拍摄到的时间段中相邻两个时间段之间的时间差，是否不大于停留时长阈值，包括：
将每天划分为多个连续的预设时间段，并将每个顾客的ID对应的每个所述被拍摄到的时间段映射至所述多个连续的预设时间段；
确定每个预设时间段中所述顾客的ID对应的最小时间点以及最大时间点；
判断映射有所述被拍摄到的时间段的预设时间段中，相邻两个预设时间段中在后的预设时间段的最小时间点与在前的预设时间段的最大时间点之间的差值是否不大于所述停留时长阈值。
可选地，所述将每天划分为多个连续的预设时间段，并将每个顾客的ID对应的每个所述被拍摄到的时间段映射至所述多个连续的预设时间段，包括：
将同一顾客的ID对应的每个被拍摄到的时间段中的连续的时间点投射到连续的预设时间段中，其中，多个连续的预设时间段中相邻的两个预设时间段的时间间隔长度与所述停留时长阈值相同。
可选地，在所述确定每次到店行为的到店时间以及离店时间之后，所述方法还包括：确定每个顾客的ID的每次到店行为的停留时长。
可选地，所述人脸识别结果数据中包括日期，在所述确定每次到店行为的到店时间以及离店时间之后，所述方法还包括：
确定每个顾客的ID在所述日期对应的到访次数以及总停留时长。
可选地，所述人脸识别结果数据还包括顾客被拍摄到的拍摄区域，所述方法还包括：
将每个顾客的ID对应的所述到店时间、所述离店时间、所述停留时长、所述总停留时长以及所述到访次数中的至少一个进行整合并按照预设保存条件保存至数据库，所述预设保存条件包括门店和/或拍摄区域。
可选地，所述获取门店内顾客的人脸识别结果数据，包括：
在实时的人脸识别结果数据传来时，判断该人脸识别结果数据对应的脸部数据是否与预先存储的员工的脸部数据相同；
若该人脸识别结果数据对应的脸部数据与预先存储的员工的脸部数据相同，则判定该人脸识别结果数据对应的为员工的人脸识别结果数据；
若该人脸识别结果数据对应的脸部数据与预先存储的员工的脸部数据不相同，则判定该人脸识别结果数据对应的为门店内顾客的人脸识别结果数据，并获取该人脸识别结果数据。本公开中，员工包括但不限于快递员、清洁工、保安等。
可选地，所述方法还包括：
若顾客的ID对应一个所述被拍摄到的时间段，则该时间段认定为对应顾客的一次到店行为。
本公开第二方面提供一种顾客到访分析装置，包括：
数据获取模块，用于获取门店内顾客的人脸识别结果数据，所述人脸识别结果数据包括所述顾客的ID以及顾客被拍摄到的时间段；
顾客分析模块，用于根据顾客的ID对所述门店内顾客的人脸识别结果数据进行聚类，获得每个顾客的ID对应的至少一个所述被拍摄到的时间段；
数据处理模块，用于在顾客的ID对应多个所述被拍摄到的时间段时，判断多个所述被拍摄到的时间段中相邻两个时间段之间的时间差，是否不大于停留时长阈值；以及在所述时间差不大于所述停留时长阈值时，确定所述相邻两个时间段对应同一次到店行为；
到访确定模块，用于确定每次到店行为的到店时间以及离店时间。
可选地，所述装置还包括：数据归类模块，用于将所有顾客的人脸识别结果数据按照门店进行聚类，获得每个门店内顾客的人脸识别结果数据。
可选地，所述数据归类模块还配置为，在将所有顾客的人脸识别结果数据按照门店进行聚类之前，对上层服务传来的人脸识别结果数据进行打包并发送至消息队列；从所述消息队列中读取人脸识别结果数据进行处理。
可选地，所述数据归类模块配置为，在获取到上层服务传来的人脸识别结果数据后，对该人脸识别结果数据进行预处理，所述预处理包括完成数据的校验、整理数据结构以及压缩数据；将预处理后的人脸识别结果数据打包发送至消息队列中；将进行处理后的人脸识别结果数据按照预设保存规则保存至时序数据库中。
可选地，所述数据处理模块包括：
时间映射模块，用于将每天划分为多个连续的预设时间段，并将每个顾客的ID对应的每个所述被拍摄到的时间段映射至所述多个连续的预设时间段；
聚合统计模块，用于确定每个预设时间段中所述顾客的ID对应的最小时间点以及最大时间点；
阈值判断模块，用于判断映射有所述被拍摄到的时间段的预设时间段中，相邻两个预设时间段中在后的预设时间段的最小时间点与在前的预设时间段的最大时间点之间的差值是否不大于所述停留时长阈值。
可选地，所述时间映射模块配置为，将同一顾客的ID对应的每个被拍摄到的时间段中的连续的时间点投射到连续的预设时间段中，其中，多个连续的预设时间段中相邻的两个预设时间段的时间间隔长度与所述停留时长阈值相同。
可选地，所述到访确定模块还配置为，确定每个顾客的ID的每次到店行为的停留时长。
可选地，所述人脸识别结果数据中包括日期，所述到访确定模块还配置为，在所述确定每次到店行为的到店时间以及离店时间之后，确定每个顾客的ID在所述日期对应的到访次数以及总停留时长。
本公开第三方面提供一种电子设备，包括：处理器、存储器和总线，所述存储器存储有所述处理器可执行的机器可读指令，当所述电子设备运行时，所述处理器与所述存储器之间通过总线通信，所述机器可读指令被所述处理器执行时执行第一方面所述的方法。
本公开第四方面提供一种存储介质，所述存储介质上存储有计算机程序，所述计算机程序被处理器运行时执行第一方面所述的方法。
本公开第五方面提供一种计算机程序产品，所述计算机程序产品在计算机上运行时，使得计算机执行第一方面所述的方法。
相对现有技术，本公开提供的顾客到访分析方法、装置及存储介质，能够基于人脸识别结果对顾客的到访行为进行分析，在通过摄像头采集门店内的画面并进行人脸识别后，获得门店内的人脸识别结果数据，进而分析出门店内顾客的到访信息，包括该顾客的到店时间以及离店时间，实现对顾客到访信息的准确分析。
为使本公开的上述目的、特征和优点能更明显易懂，下文特举可选实施例，并配合所附附图，作详细说明如下。
附图说明
为了更清楚地说明本公开实施方式的技术方案，下面将对实施方式中所需要使用的附图作简单地介绍，应当理解，以下附图仅示出了本公开的某些实施例，因此不应被看作是对范围的限定，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他相关的附图。
图1示出了本公开所提供的应用场景示意图。
图2示出了本公开所提供的顾客到访分析方法的流程图；
图3示出了本公开所提供的顾客到访分析方法的另一流程图；
图4示出了本公开所提供的顾客到访分析装置的示意图。
图标：
电子设备-10；存储器-11；处理器-12；网络模块-13；顾客到访分析装置-14；数据获取模块-100；顾客分析模块-101；数据处理模块-102；到访确定模块-103。
具体实施方式
下面将结合本公开中附图，对本公开中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本公开一部分实施例，而不是全部的实施例。通常在此处附图中描述和示出的本公开的组件可以以各种不同的配置来布置和设计。因此，以下对在附图中提供的本公开的实施例的详细描述并非旨在限制要求保护的本公开的范围，而是仅仅表示本公开的选定实施例。基于本公开的实施例，本领域技术人员在没有做出创造性劳动的前提下所获得的所有其他实施例，都属于本公开保护的范围。
应注意到：相似的标号和字母在下面的附图中表示类似项，因此，一旦某一项在一个附图中被定义，则在随后的附图中不需要对其进行进一步定义和解释。
下面结合附图，对本公开的一些实施方式作详细说明。在不冲突的情况下，下述的实施例及实施例中的特征可以相互组合。
如图1所示，是本公开提供的电子设备10的一种方框示意图。本公开中的电子设备10可以为具有数据处理功能的设备，如服务器、个人计算机、平板电脑、智能手机等。如图1所示，电子设备10包括：存储器11、处理器12、网络模块13及顾客到访分析装置14。
所述存储器11、处理器12以及网络模块13相互之间直接或间接地电性连接，以实现数据的传输或交互。例如，这些元件相互之间可通过一条或多条通讯总线或信号线实现电性连接。存储器11中存储有顾客到访分析装置14，所述顾客到访分析装置14包括至少一个可以软件或固件(firmware)的形式存储于所述存储器11中的软件功能模块，所述处理器12通过运行存储在存储器11内的软件程序以及模块，如本公开中的顾客到访分析装置14，从而执行各种功能应用以及数据处理，即实现本公开中由电子设备10执行的数据处理方法。
其中，所述存储器11可以是，但不限于，随机存取存储器(Random Access Memory，RAM)，只读存储器(Read Only Memory，ROM)，可编程只读存储器(Programmable Read-Only Memory，PROM)，可擦除只读存储器(Erasable Programmable Read-Only Memory，EPROM)，电可擦除只读存储器(Electric Erasable Programmable Read-Only Memory，EEPROM)等。其中，存储器11用于存储程序，所述处理器12在接收到执行指令后，执行所述程序。
所述处理器12可能是一种集成电路芯片，具有数据的处理能力。上述的处理器12可以是通用处理器，包括中央处理器(Central Processing Unit，CPU)、网络处理器(Network Processor，NP)等。可以实现或者执行本公开中提供的各方法、步骤及逻辑框图。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。
网络模块13用于通过网络建立电子设备10与外部通信终端之间的通信连接，实现网络信号及数据的收发操作。上述网络信号可包括无线信号或者有线信号。
可以理解，图1所示的结构仅为示意，电子设备10还可包括比图1中所示更多或者更少的组件，或者具有与图1所示不同的配置。图1中所示的各组件可以采用硬件、软件或其组合实现。例如，电子设备10还可以包括人脸识别器，电子设备10通过自身所包括的人脸识别器进行人脸识别，得到人脸识别结果数据。又例如，电子设备10中的处理器12可以与独立于电子设备10的人脸识别器通信连接，从而获取人脸识别器识别并得到的人脸识别结果数据。
第一实施例
本实施例提供了一种顾客到访分析方法，请参阅图2，包括以下步骤，这些步骤可以由图1所示电子设备10中的处理器12执行。
步骤110：获取门店内顾客的人脸识别结果数据。
以某一个门店为例，在该门店内设置有多个摄像头，每个摄像头会对门店内的画面进行采集。在画面中，多个人员在画面中行走，画面中包括有员工以及顾客。摄像头实时传输回的人脸识别结果数据中包括有员工的脸部数据以及顾客的脸部数据，因此若想要获取到顾客的人脸识别结果数据，需要将员工的人脸识别结果数据与顾客的人脸识别结果数据进行区分。其中一种实现方式为，事先建立员工信息以及员工对应的脸部数据的数据库。在实时的人脸识别结果数据传来时，判断该人脸识别结果数据对应的脸部数据是否与预先存储的员工的脸部数据相同，若该人脸识别结果数据对应的脸部数据与预先存储的员工的脸部数据相同，则判定该条人脸识别结果数据对应的为员工的人脸识别结果数据。若传来的人脸识别结果数据对应的脸部数据与预先存储的员工的脸部数据不相同，则判定该条人脸识别结果数据对应的为门店内顾客的人脸识别结果数据。
上述人脸识别结果数据包括有顾客的ID(身份标识号)以及顾客被拍摄到的时间段。在摄像头拍摄到的画面中存在多个员工以及多个顾客时，通过人脸识别技术将每个脸部数据转化为ID，通过ID将不同的人的脸部数据进行区分，便于后期对每个顾客的到访行为进行分析。在门店内设置有多个摄像头，摄像头传输回的数据可以为视频流，对视频流画面中的人脸进行识别，能够获取到每个顾客在该摄像头中被拍摄到的时间段。也可以将视 频流理解为图片流，摄像头实时将被拍摄到的图片传输回来，对图片中的人脸进行识别。例如，某一顾客在9:05至9:10被第一摄像头采集到，则该顾客的一条人脸识别结果数据中被拍摄到的时间段相应地为9:05至9:10。
步骤120：根据顾客的ID对门店内顾客的人脸识别结果数据进行聚类。
一条人脸识别结果数据中包括顾客的ID以及顾客被拍摄到的时间段，当然，还可以有其他信息，例如，还可以包括顾客被拍摄到的区域，若存在有多个门店，可以包括被拍摄到的门店的信息。当多条人脸识别结果数据传来，根据人脸识别结果数据中顾客的ID进行聚类，获得每个顾客的ID对应的一个或者多个被拍摄到的时间段，从而实现人脸识别结果数据的聚类。通常情况下每个顾客的ID会具有多个被拍摄到的时间段，在门店中仅有一个摄像头，以及顾客一直被该摄像头拍摄到时，有可能仅存在一个被拍摄到的时间段。
步骤130：确定相邻两个时间段是否对应同一次到店行为。
若顾客的ID仅对应一个被拍摄到的时间段，则这个时间段认定为对应顾客的一次到店行为，若顾客的ID对应多个被拍摄到的时间段，则需要判断多个被拍摄到的时间段中相邻两个时间段之间的时间差，是否不大于预设的停留时长阈值。
举例说明，对于某一顾客来说，在门店内的区域很大时，该顾客可能在被拍摄到之后的某一时间段内位于店内的某些死角区域，摄像头并未拍摄到该顾客的人脸，因此在人脸识别结果数据中，顾客的多个被拍摄到的时间段中可能存在某一时间空白。例如，在9:05至9:10拍摄到该顾客，在9:20至9:40拍摄到该顾客，则这两个相邻的时间段的时间差为10分钟。在事先设定顾客身份对应的停留时长阈值，例如30分钟，也就是说，在该顾客两个相邻时间段的时间差为10分钟时，由于并未大于预设的停留时长阈值30分钟，判断这两条数据中被拍摄到的时间段对应同一次到店行为。
另需要说明的是，对于同一个顾客来说，在门店内购买商品时，可能同时被多个摄像头拍摄到。例如，顾客在9:05至9:20被第一摄像头拍摄到，在9:10至9:30被第二摄像头拍摄到，这种情况，由于顾客被拍摄到的时间段重叠，时间差显然不大于预设的停留时长阈值，同样视为一次到店行为。
本公开中，停留时长阈值可以灵活设定，例如，可以针对不同的门店设置不同的停留时长阈值。如根据各门店内的区域大小分别设置停留时长阈值，针对区域较大的门店所设定的停留时长阈值可以大于针对区域较小的门店所设定的停留时长阈值。又例如，可以针对不同的顾客的ID设置不同的停留时长阈值。如针对各顾客进行历史数据收集和分析，针对分析得出在门店内停留时间较长的顾客，所设定的停留时长阈值可以大于针对分析得出在门店内停留时间较短的顾客所设定的停留时长阈值。又例如，针对不同的门店和各顾客， 停留时长阈值可以设定为同一固定时长。
步骤140：确定每次到店行为的到店时间以及离店时间。
根据步骤130确定每个顾客的每一次到店行为后，则可以根据人脸识别结果数据中被拍摄到的时间段确定每次到店行为的到店时间以及离店时间，从而获得门店内每个顾客的到访信息。
通过上述步骤110至步骤140中所描述的方案能够针对某一门店对顾客到访行为进行分析。通过摄像头采集门店内的画面，在获得门店内的人脸识别结果数据后，可以分析出每个门店内每个顾客的到访信息，包括该顾客的到店时间以及离店时间。
但是，由于在全国存在有多个城市，每个城市中可能存在有多个分店，为了适用于这样大数据、大流量的情况，在步骤110之前，还可以包括：
将所有顾客的人脸识别结果数据按照门店进行聚类，获得每个门店内顾客的人脸识别结果数据。
例如，若某一个大型连锁超市，在全国各个城市设有多家门店，每个门店中的多个摄像头会将拍摄到的画面传回，并对画面中的人脸进行识别。在获取到所有门店的人脸识别结果数据后，按照门店确定每个门店内的顾客的人脸识别结果数据，并分别对每个门店内的人脸识别结果数据进行分析，从而能够有效提高分析效率，减少数据分析的时间。
可选地，为了对众多人脸识别结果数据进行处理，在将所有顾客的人脸识别结果数据按照门店进行聚类之前，需要对上层服务传来的人脸识别结果数据进行打包并发送至消息队列，在对人脸识别结果数据进行处理时，则从消息队列中读取数据。例如，摄像头中可以集成人脸识别器，摄像头采集到门店内的画面之后，人脸识别器对摄像头采集到的画面进行人脸识别处理，得到人脸识别结果数据。电子设备10的处理器12中包括缓存人脸识别结果数据的消息队列。人脸识别器得到人脸识别结果数据之后，将得到的人脸识别结果数据发送至处理器12，处理器12对人脸识别器传来的人脸识别结果数据进行打包并发送至消息队列。处理器12在对人脸识别结果数据进行处理时，按缓存的先后顺序，从消息队列中读取人脸识别结果数据并进行处理。从而确保数据分析可靠性。
由于对顾客进行到访分析对实时性的要求不高，而对众多数据进行分析计算较为耗时，本实施例引入了消息队列，作为数据处理的缓冲区。还可以引入异步机制，如引入多个消息队列，采用多线程的异步处理，对各消息队列分别进行处理，从而有效提高数据处理的能力。尤其是当数据流量较大的情况下，就可以结合消息队列，采用异步机制来对大数据进行处理，可以有效降低处理数据的负荷。本公开中，各数据可以存储在数据库中，相应地，采用异步机制来对大数据进行处理，可以有效降低数据库处理数据的负荷。
可选地，在步骤130中判断多个被拍摄到的时间段中相邻两个时间段之间的时间差是否不大于停留时长阈值的步骤，可以参阅图3实现。该步骤的一种详细的实施方式为：
步骤210：将每天划分为多个连续的预设时间段，并将每个顾客的ID对应的每个被拍摄到的时间段映射至多个连续的预设时间段。
可以将同一顾客的ID对应的每个被拍摄到的时间段中的连续的时间点投射到连续的预设时间段中。其中，多个连续的预设时间段中相邻的两个预设时间段的时间间隔长度可以与停留时长阈值相同。
步骤220：确定每个预设时间段中顾客的ID对应的最小时间点以及最大时间点。
通过聚合函数获得每个预设时间段中所有连续的时间点中的最小时间点以及最大时间点。
步骤230：判断映射有被拍摄到的时间段的预设时间段中，相邻两个预设时间段中在后的预设时间段的最小时间点与在前的预设时间段的最大时间点之间的差值是否不大于停留时长阈值。
通过上述算法分析获得每个顾客对应的每个被拍摄的时间段在每个预设时间段中最小时间点以及最大时间点，从而通过分析计算确定顾客的一次到店行为。
接下来再详细说明在大数据、大流量下整个顾客到访分析方法的实施方式。
电子设备10中的处理器12在获取到上层服务传来的人脸识别结果数据后，对人脸识别结果数据进行预处理，预处理可以包括完成数据的校验、整理数据结构以及压缩数据，将预处理后的人脸识别结果数据打包发送至消息队列中。
从消息队列中读取数据，并对读取到的数据进行后处理，后处理可以包括对人脸识别结果数据进行解压以及整理数据结构。进行后处理后的人脸识别结果数据按照一定的预设保存规则保存至时序数据库中，例如按照门店或按照拍摄区域进行保存，同时对这些数据进行顾客到访分析。本公开中，时序数据库可以理解为将上述数据库中的人脸识别结果数据按时间先后顺序存储而形成。这里的时序主要指每条数据都是有时间字段的，获取数据的时候，可以根据这个时间字段按照一定条件获取，例如按照进入时间去顺序获取数据，或者按照离开时间倒序获取数据。本公开中的时序数据库与前述数据库可以指同一个数据库，例如，可以为结构化查询语言(Structured Query Language，SQL)类型的数据库，也可以为非关系型数据库NoSQL等。
获取到的这些人脸识别结果数据为全国所有门店以及门店内各个拍摄区域各个摄像头拍摄到的画面的人脸识别结果数据，对这些数据区分出不同门店的数据，同时对于一些无用的数据进行过滤，无用的数据可以包括识别到的员工的人脸数据以及一些模糊脸部数据。
对每个门店内的人脸识别结果数据进行顾客的到访分析，构造数据分析条件。数据分析条件可以包括需要分析的顾客的ID、哪天的数据、顾客的停留时长阈值等字段。根据数据分析条件构造时序数据的查询聚合语句，向时序数据库发起查询聚合操作请求。
例如，若想要分析某日某个顾客的ID的到访行为，则从时序数据库中查询这一日该顾客的ID对应的多个被拍摄到的时间段，将这些时间段中的时间点进行排列并对应在每个预设的单位时间内，其单位时间间隔的长度为数据分析条件中顾客的停留时长阈值。若顾客的停留时长阈值为三十分钟，则将一天24小时按照三十分钟进行等分，通过聚合函数，获得该顾客被拍摄到的每个预设的单位时间内的最小时间点以及最大时间点。
举例说明，在停留时长阈值为三十分钟时，每个预设的单位时间间隔也为三十分钟，若顾客在9:05至9:20被第一摄像头拍摄到，在9:10至9:25被第二摄像头拍摄到，在9:45至9:50被第三摄像头拍摄到，则将这些时间段对应的时间点按照时间进行排列。
在9:00至9:30为一个预设单位时间，在9:30至10:00为相邻的另一个预设单位时间，通过聚合函数，获得9:00至9:30的最小时间点9:05以及最大时间点9:25，获得9:30至10:00的最小时间点9:45以及最大时间点9:50。
在通过聚合统计确定每个顾客的ID对应的每个预设单位时间内的最小时间点以及最大时间点后，判断在后的预设单位时间内的最小时间点与在前的预设单位时间内的最大时间点的差值是否不大于预设的停留时长阈值。若不大于，如9:45与9:25之间的差值为20分钟，不大于停留时长阈值30分钟，则确定为顾客的一次到访行为，并确定此次到访的到店时间和离店时间，也就是确定上述例子中的顾客9:05到店，9:50离店。若差值大于预设的停留时长阈值，则确定为另外一次到访行为。
在确定每个顾客的到店时间以及离店时间之后，进一步对数据进行分析计算，确定每次到访行为的停留时长，在顾客存在多次到访行为时，获得顾客的到访次数，以及这一天的总停留时长。
通过上述过程，可以对不同门店的每位顾客的到访行为进行分析，也可以在上述数据分析条件中增加拍摄区域，从而能够分区域分析获得每个区域内的顾客到访行为，例如生鲜区、服饰区、生活区等等。
在分析计算完成后，能够获得顾客每次到访的停留时长、到访次数以及当日总停留时长。进一步地，也可以根据每日分析的数据，获得每位顾客每周、每月、每年的总到访次数、每次到访的停留时长以及总停留时长。
将整个分析过程中获得的数据进行整合，如对顾客的ID、门店、拍摄区域、日期、每次到访行为的到店时间、离店时间、停留时长阈值、停留时长、到访次数以及总停留时长 等数据进行整合，将这些数据按照预设的保存条件保存至数据库中，例如按照门店或者按照区域进行保存。
通过上述步骤，能够有效地利用被人脸识别后的数据来分析出每位顾客的到访记录，且整个数据处理流程适用于大流量、大数据的场景，同时做到了高可用。
第二实施例
本实施例提供一种顾客到访分析装置，参阅图4，包括：
数据获取模块100，用于获取门店内顾客的人脸识别结果数据，所述人脸识别结果数据包括所述顾客的ID以及顾客被拍摄到的时间段。
关于数据获取模块100的具体实现方式可以参阅对图2中步骤S110的相关描述，因而在此不作赘述。
顾客分析模块101，用于根据顾客的ID对所述门店内顾客的人脸识别结果数据进行聚类，获得每个顾客的ID对应的至少一个所述被拍摄到的时间段。
关于顾客分析模块101的具体实现方式可以参阅对图2中步骤S120的相关描述，因而在此不作赘述。
数据处理模块102，用于在顾客的ID对应多个所述被拍摄到的时间段时，判断多个所述被拍摄到的时间段中相邻两个时间段之间的时间差，是否不大于停留时长阈值；以及在所述时间差不大于所述停留时长阈值时，确定所述相邻两个时间段对应同一次到店行为。
关于数据处理模块102的具体实现方式可以参阅对图2中步骤S130的相关描述，因而在此不作赘述。
到访确定模块103，用于确定每次到店行为的到店时间以及离店时间。
关于到访确定模块103的具体实现方式可以参阅对图2中步骤S140的相关描述，因而在此不作赘述。
可选地，所述装置还包括：数据归类模块，用于将所有顾客的人脸识别结果数据按照门店进行聚类，获得每个门店内顾客的人脸识别结果数据。
可选地，所述数据归类模块还配置为，在将所有顾客的人脸识别结果数据按照门店进行聚类之前，对上层服务传来的人脸识别结果数据进行打包并发送至消息队列；从所述消息队列中读取人脸识别结果数据进行处理。
可选地，所述数据归类模块配置为，在获取到上层服务传来的人脸识别结果数据后，对该人脸识别结果数据进行预处理，所述预处理包括完成数据的校验、整理数据结构以及压缩数据；将预处理后的人脸识别结果数据打包发送至消息队列中；将进行处理后的人脸识别结果数据按照预设保存规则保存至时序数据库中。
可选地，所述数据处理模块包括：时间映射模块，用于将每天划分为多个连续的预设时间段，并将每个顾客的ID对应的每个所述被拍摄到的时间段映射至所述多个连续的预设时间段；聚合统计模块，用于确定每个预设时间段中所述顾客的ID对应的最小时间点以及最大时间点；阈值判断模块，用于判断映射有所述被拍摄到的时间段的预设时间段中，相邻两个预设时间段中在后的预设时间段的最小时间点与在前的预设时间段的最大时间点之间的差值是否不大于所述停留时长阈值。
可选地，所述时间映射模块配置为，将同一顾客的ID对应的每个被拍摄到的时间段中的连续的时间点投射到连续的预设时间段中，其中，多个连续的预设时间段中相邻的两个预设时间段的时间间隔长度与所述停留时长阈值相同。
可选地，所述到访确定模块103还配置为，确定每个顾客的ID的每次到店行为的停留时长。
可选地，所述人脸识别结果数据中包括日期，所述到访确定模块103还配置为，在所述确定每次到店行为的到店时间以及离店时间之后，确定每个顾客的ID在所述日期对应的到访次数以及总停留时长。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的装置的具体工作过程，可以参考前述方法中的对应过程，在此不再过多赘述。
本公开还提供一种电子设备，包括：处理器、存储器和总线，所述存储器存储有所述处理器可执行的机器可读指令，当所述电子设备运行时，所述处理器与所述存储器之间通过总线通信，所述机器可读指令被所述处理器执行时执行第一实施例所述的方法。
本公开还提供一种存储介质，该存储介质上存储有计算机程序，该计算机程序被处理器运行时执行第一实施例所述的方法。
本公开还提供一种计算机程序产品，所述计算机程序产品在计算机上运行时，使得计算机执行第一实施例所述的方法。
本公开所提供的实施例中，应该理解到，所揭露的装置和方法，也可以通过其它的方式实现。以上所描述的装置实施例仅仅是示意性的，例如，附图中的流程图和框图显示了根据本公开的多个实施例的装置、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段或代码的一部分，所述模块、程序段或代码的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。也应当注意，在有些作为替换的实现方式中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个连续的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方 框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或动作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。
另外，在本公开各个实施例中的各功能模块可以集成在一起形成一个独立的部分，也可以是各个模块单独存在，也可以两个或两个以上模块集成形成一个独立的部分。
所述功能如果以软件功能模块的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本公开的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本公开各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质。需要说明的是，在本文中，诸如第一和第二等之类的关系术语仅仅用来将一个实体或者操作与另一个实体或操作区分开来，而不一定要求或者暗示这些实体或操作之间存在任何这种实际的关系或者顺序。而且，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者设备不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者设备所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括所述要素的过程、方法、物品或者设备中还存在另外的相同要素。
以上所述仅为本公开的可选实施例而已，并不用于限制本公开，对于本领域的技术人员来说，本公开可以有各种更改和变化。凡在本公开的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本公开的保护范围之内。
工业实用性
本公开提供的顾客到访分析方法、装置及存储介质，能够基于人脸识别结果对顾客的到访行为进行分析，进而分析出门店内顾客的到访信息，包括该顾客的到店时间以及离店时间，实现对顾客到访信息的准确分析。

Claims (20)

  1. 一种顾客到访分析方法，其特征在于，包括：
    获取门店内顾客的人脸识别结果数据，所述人脸识别结果数据包括所述顾客的ID以及顾客被拍摄到的时间段；
    根据顾客的ID对所述门店内顾客的人脸识别结果数据进行聚类，获得每个顾客的ID对应的至少一个所述被拍摄到的时间段；
    若顾客的ID对应多个所述被拍摄到的时间段，则判断多个所述被拍摄到的时间段中相邻两个时间段之间的时间差，是否不大于停留时长阈值；
    若不大于，则确定所述相邻两个时间段对应同一次到店行为；
    确定每次到店行为的到店时间以及离店时间。
  2. 根据权利要求1所述的顾客到访分析方法，其特征在于，在所述获取门店内顾客的人脸识别结果数据之前，所述方法还包括：
    将所有顾客的人脸识别结果数据按照门店进行聚类，获得每个门店内顾客的人脸识别结果数据。
  3. 根据权利要求2所述的顾客到访分析方法，其特征在于，在将所有顾客的人脸识别结果数据按照门店进行聚类之前，所述方法还包括：
    对上层服务传来的人脸识别结果数据进行打包并发送至消息队列；
    从所述消息队列中读取人脸识别结果数据进行处理。
  4. 根据权利要求3所述的顾客到访分析方法，其特征在于，所述对上层服务传来的人脸识别结果数据进行打包并发送至消息队列，包括：
    在获取到上层服务传来的人脸识别结果数据后，对该人脸识别结果数据进行预处理，所述预处理包括完成数据的校验、整理数据结构以及压缩数据；
    将预处理后的人脸识别结果数据打包发送至消息队列中；
    所述从所述消息队列中读取人脸识别结果数据进行处理，包括：对人脸识别结果数据进行解压以及整理数据结构；
    将进行处理后的人脸识别结果数据按照预设保存规则保存至时序数据库中。
  5. 根据权利要求1至4任意一项所述的顾客到访分析方法，其特征在于，所述判断多个所述被拍摄到的时间段中相邻两个时间段之间的时间差，是否不大于停留时长阈值，包括：
    将每天划分为多个连续的预设时间段，并将每个顾客的ID对应的每个所述被拍摄到 的时间段映射至所述多个连续的预设时间段；
    确定每个预设时间段中所述顾客的ID对应的最小时间点以及最大时间点；
    判断映射有所述被拍摄到的时间段的预设时间段中，相邻两个预设时间段中在后的预设时间段的最小时间点与在前的预设时间段的最大时间点之间的差值是否不大于所述停留时长阈值。
  6. 根据权利要求5所述的顾客到访分析方法，其特征在于，所述将每天划分为多个连续的预设时间段，并将每个顾客的ID对应的每个所述被拍摄到的时间段映射至所述多个连续的预设时间段，包括：
    将同一顾客的ID对应的每个被拍摄到的时间段中的连续的时间点投射到连续的预设时间段中，其中，多个连续的预设时间段中相邻的两个预设时间段的时间间隔长度与所述停留时长阈值相同。
  7. 根据权利要求1至6任意一项所述的顾客到访分析方法，其特征在于，在所述确定每次到店行为的到店时间以及离店时间之后，所述方法还包括：
    确定每个顾客的ID的每次到店行为的停留时长。
  8. 根据权利要求7所述的顾客到访分析方法，其特征在于，所述人脸识别结果数据中包括日期，在所述确定每次到店行为的到店时间以及离店时间之后，所述方法还包括：
    确定每个顾客的ID在所述日期对应的到访次数以及总停留时长。
  9. 根据权利要求8所述的顾客到访分析方法，其特征在于，所述人脸识别结果数据还包括顾客被拍摄到的拍摄区域，所述方法还包括：
    将每个顾客的ID对应的所述到店时间、所述离店时间、所述停留时长、所述总停留时长以及所述到访次数中的至少一个进行整合并按照预设保存条件保存至数据库，所述预设保存条件包括门店和/或拍摄区域。
  10. 根据权利要求1至9任意一项所述的顾客到访分析方法，其特征在于，所述获取门店内顾客的人脸识别结果数据，包括：
    在实时的人脸识别结果数据传来时，判断该人脸识别结果数据对应的脸部数据是否与预先存储的员工的脸部数据相同；
    若该人脸识别结果数据对应的脸部数据与预先存储的员工的脸部数据相同，则判定该人脸识别结果数据对应的为员工的人脸识别结果数据；
    若该人脸识别结果数据对应的脸部数据与预先存储的员工的脸部数据不相同，则判定该人脸识别结果数据对应的为门店内顾客的人脸识别结果数据，并获取该人脸识别结果数据。
  11. 根据权利要求1至10任意一项所述的顾客到访分析方法，其特征在于，所述方法还包括：
    若顾客的ID对应一个所述被拍摄到的时间段，则该时间段认定为对应顾客的一次到店行为。
  12. 一种顾客到访分析装置，其特征在于，包括：
    数据获取模块，配置为获取门店内顾客的人脸识别结果数据，所述人脸识别结果数据包括所述顾客的ID以及顾客被拍摄到的时间段；
    顾客分析模块，配置为根据顾客的ID对所述门店内顾客的人脸识别结果数据进行聚类，获得每个顾客的ID对应的至少一个所述被拍摄到的时间段；
    数据处理模块，配置为在顾客的ID对应多个所述被拍摄到的时间段时，判断多个所述被拍摄到的时间段中相邻两个时间段之间的时间差，是否不大于停留时长阈值；以及在所述时间差不大于所述停留时长阈值时，确定所述相邻两个时间段对应同一次到店行为；
    到访确定模块，配置为确定每次到店行为的到店时间以及离店时间。
  13. 根据权利要求12所述的顾客到访分析装置，其特征在于，所述装置还包括：
    数据归类模块，配置为将所有顾客的人脸识别结果数据按照门店进行聚类，获得每个门店内顾客的人脸识别结果数据。
  14. 根据权利要求13所述的顾客到访分析装置，其特征在于，所述数据归类模块还配置为，在将所有顾客的人脸识别结果数据按照门店进行聚类之前，对上层服务传来的人脸识别结果数据进行打包并发送至消息队列；从所述消息队列中读取人脸识别结果数据进行处理。
  15. 根据权利要求14所述的顾客到访分析装置，其特征在于，所述数据归类模块配置为，在获取到上层服务传来的人脸识别结果数据后，对该人脸识别结果数据进行预处理，所述预处理包括完成数据的校验、整理数据结构以及压缩数据；将预处理后的人脸识别结果数据打包发送至消息队列中；将进行处理后的人脸识别结果数据按照预设保存规则保存至时序数据库中。
  16. 根据权利要求12至15任意一项所述的顾客到访分析装置，其特征在于，所述数据处理模块包括：
    时间映射模块，配置为将每天划分为多个连续的预设时间段，并将每个顾客的ID对应的每个所述被拍摄到的时间段映射至所述多个连续的预设时间段；
    聚合统计模块，配置为确定每个预设时间段中所述顾客的ID对应的最小时间点以及最大时间点；
    阈值判断模块，配置为判断映射有所述被拍摄到的时间段的预设时间段中，相邻两个预设时间段中在后的预设时间段的最小时间点与在前的预设时间段的最大时间点之间的差值是否不大于所述停留时长阈值。
  17. 根据权利要求16所述的顾客到访分析装置，其特征在于，所述时间映射模块配置为，将同一顾客的ID对应的每个被拍摄到的时间段中的连续的时间点投射到连续的预设时间段中，其中，多个连续的预设时间段中相邻的两个预设时间段的时间间隔长度与所述停留时长阈值相同。
  18. 根据权利要求12至17任意一项所述的顾客到访分析装置，其特征在于，所述到访确定模块还配置为，确定每个顾客的ID的每次到店行为的停留时长。
  19. 根据权利要求18所述的顾客到访分析装置，其特征在于，所述人脸识别结果数据中包括日期，所述到访确定模块还配置为，在所述确定每次到店行为的到店时间以及离店时间之后，确定每个顾客的ID在所述日期对应的到访次数以及总停留时长。
  20. 一种存储介质，其特征在于，所述存储介质上存储有计算机程序，所述计算机程序被处理器运行时执行如权利要求1至11任一项所述的方法。
PCT/CN2018/117286 2018-09-18 2018-11-23 顾客到访分析方法、装置及存储介质 WO2020056915A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
EP18934444.3A EP3855343A4 (en) 2018-09-18 2018-11-23 APPARATUS AND PROCESS FOR CUSTOMER VISIT ANALYSIS, AND INFORMATION SUPPORT

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811093386.XA CN109190586B (zh) 2018-09-18 2018-09-18 顾客到访分析方法、装置及存储介质
CN201811093386.X 2018-09-18

Publications (1)

Publication Number Publication Date
WO2020056915A1 true WO2020056915A1 (zh) 2020-03-26

Family

ID=64908263

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/117286 WO2020056915A1 (zh) 2018-09-18 2018-11-23 顾客到访分析方法、装置及存储介质

Country Status (3)

Country Link
EP (1) EP3855343A4 (zh)
CN (1) CN109190586B (zh)
WO (1) WO2020056915A1 (zh)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111695449A (zh) * 2020-05-27 2020-09-22 北京华夏电通科技有限公司 基于图像检测的行为识别方法及装置
CN112036957A (zh) * 2020-09-08 2020-12-04 广州图普网络科技有限公司 一种访客留存数确定方法、装置、电子设备和存储介质
CN112085534A (zh) * 2020-09-11 2020-12-15 中德（珠海）人工智能研究院有限公司 一种关注度分析方法、系统及存储介质
CN114842414A (zh) * 2022-04-26 2022-08-02 中国电信股份有限公司 图像处理方法、装置、电子设备及存储介质

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112036919A (zh) * 2019-06-03 2020-12-04 北京市商汤科技开发有限公司 数据处理方法、装置及存储介质
CN110450175A (zh) * 2019-08-05 2019-11-15 深圳市三宝创新智能有限公司 一种前台机器人的访客接待方法
CN110471963B (zh) * 2019-08-14 2022-04-05 北京市商汤科技开发有限公司 数据处理方法、装置及存储介质
CN110659615A (zh) * 2019-09-26 2020-01-07 上海依图信息技术有限公司 基于人脸识别的客群流量及结构化分析系统及方法
CN111340569A (zh) * 2020-03-27 2020-06-26 上海钧正网络科技有限公司 基于跨境追踪的门店人流分析方法、装置、系统、终端和介质
CN112183380B (zh) * 2020-09-29 2021-08-10 新疆爱华盈通信息技术有限公司 基于人脸识别的客流量分析方法和系统、电子设备
CN113763051A (zh) * 2021-09-08 2021-12-07 北京房江湖科技有限公司 门店到访信息确定方法、装置及存储介质
CN115862100A (zh) * 2022-11-28 2023-03-28 成都比特至简科技有限公司 一种基于录像的人脸回溯方法和系统
CN116320322B (zh) * 2023-05-12 2023-08-22 美宜佳控股有限公司 门店云值守视频推流方法、装置、系统、设备及介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2015043328A1 (zh) * 2013-09-27 2015-04-02 杭州海康威视数字技术股份有限公司 基于移动信息的视频录像智能回放方法及其装置
CN107622673A (zh) * 2017-09-22 2018-01-23 广东方纬科技有限公司 一种基于卡口数据的停车需求分析方法、系统和装置
CN107679613A (zh) * 2017-09-30 2018-02-09 同观科技（深圳）有限公司 一种人员信息的统计方法、装置、终端设备和存储介质
CN107820209A (zh) * 2017-11-07 2018-03-20 锐捷网络股份有限公司 兴趣推荐方法、装置及服务器

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040078260A1 (en) * 2002-10-22 2004-04-22 Michael Milgramm System for sales optimization utilizing biometric customer recognition technique
US9245276B2 (en) * 2012-12-12 2016-01-26 Verint Systems Ltd. Time-in-store estimation using facial recognition
CN105518734A (zh) * 2013-09-06 2016-04-20 日本电气株式会社 顾客行为分析系统、顾客行为分析方法、非暂时性计算机可读介质和货架系统
CN106776781B (zh) * 2016-11-11 2018-08-24 深圳云天励飞技术有限公司 一种人群关系网络分析方法及装置
CN108470157A (zh) * 2018-03-06 2018-08-31 杜潇 一种客户信息获取方法及系统

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2015043328A1 (zh) * 2013-09-27 2015-04-02 杭州海康威视数字技术股份有限公司 基于移动信息的视频录像智能回放方法及其装置
CN107622673A (zh) * 2017-09-22 2018-01-23 广东方纬科技有限公司 一种基于卡口数据的停车需求分析方法、系统和装置
CN107679613A (zh) * 2017-09-30 2018-02-09 同观科技（深圳）有限公司 一种人员信息的统计方法、装置、终端设备和存储介质
CN107820209A (zh) * 2017-11-07 2018-03-20 锐捷网络股份有限公司 兴趣推荐方法、装置及服务器

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3855343A4 *

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111695449A (zh) * 2020-05-27 2020-09-22 北京华夏电通科技有限公司 基于图像检测的行为识别方法及装置
CN112036957A (zh) * 2020-09-08 2020-12-04 广州图普网络科技有限公司 一种访客留存数确定方法、装置、电子设备和存储介质
CN112036957B (zh) * 2020-09-08 2023-11-28 广州图普网络科技有限公司 一种访客留存数确定方法、装置、电子设备和存储介质
CN112085534A (zh) * 2020-09-11 2020-12-15 中德（珠海）人工智能研究院有限公司 一种关注度分析方法、系统及存储介质
CN112085534B (zh) * 2020-09-11 2023-01-06 中德（珠海）人工智能研究院有限公司 一种关注度分析方法、系统及存储介质
CN114842414A (zh) * 2022-04-26 2022-08-02 中国电信股份有限公司 图像处理方法、装置、电子设备及存储介质
CN114842414B (zh) * 2022-04-26 2023-11-21 中国电信股份有限公司 图像处理方法、装置、电子设备及存储介质

Also Published As

Publication number Publication date
CN109190586B (zh) 2019-06-11
EP3855343A4 (en) 2021-11-17
CN109190586A (zh) 2019-01-11
EP3855343A1 (en) 2021-07-28

Similar Documents

Publication Publication Date Title
WO2020056915A1 (zh) 顾客到访分析方法、装置及存储介质
US20220092881A1 (en) Method and apparatus for behavior analysis, electronic apparatus, storage medium, and computer program
WO2020056914A1 (zh) 人群热力图获得方法、装置、电子设备及可读存储介质
WO2022001097A1 (zh) 一种用于人证核验设备性能检测的算法测评系统及检测方法
CN110941978B (zh) 一种未识别身份人员的人脸聚类方法、装置及存储介质
CN110738779A (zh) 一种人员轨迹检测方法、系统及装置
CN111415241A (zh) 欺诈人员识别方法、装置、设备和存储介质
CN111104841A (zh) 暴力行为检测方法及系统
CN110991231B (zh) 活体检测方法、装置、服务器和人脸识别设备
US11657548B2 (en) Information processing device, display method, and program storage medium for monitoring object movement
WO2021043089A1 (zh) 一种客流量统计方法、装置、设备及计算机可读存储介质
US20160125404A1 (en) Face recognition business model and method for identifying perpetrators of atm fraud
CN109902550A (zh) 行人属性的识别方法和装置
JP2015090579A (ja) 行動分析システム
CN114973436B (zh) 考勤方法、装置、智能考勤终端及存储介质
CN109558773A (zh) 信息识别方法、装置及电子设备
CN112183380B (zh) 基于人脸识别的客流量分析方法和系统、电子设备
CN112687044A (zh) 滞留人员的管理方法、系统、介质及设备
CN115063852A (zh) 用于确定员工考勤信息的方法、装置、存储介质及处理器
CN111723761B (zh) 异常人脸图像的确定方法、装置及存储介质
WO2021135933A1 (zh) 一种目标的识别方法及装置、存储介质及电子装置
AU2020202221A1 (en) Privacy preserving visitor recognition and movement pattern analysis based on computer vision
CN114913600A (zh) 一种电表窃电检测方法、装置、电子设备及存储介质
CN112600846B (zh) 基于物联网的考勤管理方法和系统
US20220215525A1 (en) Information processing device, information processing program, and information processing method

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18934444

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2018934444

Country of ref document: EP

Effective date: 20210419