WO2020233055A1 - 基于活体检测的产品推广方法、装置、设备及存储介质 - Google Patents

基于活体检测的产品推广方法、装置、设备及存储介质 Download PDF

Info

Publication number
WO2020233055A1
WO2020233055A1 PCT/CN2019/120912 CN2019120912W WO2020233055A1 WO 2020233055 A1 WO2020233055 A1 WO 2020233055A1 CN 2019120912 W CN2019120912 W CN 2019120912W WO 2020233055 A1 WO2020233055 A1 WO 2020233055A1
Authority
WO
WIPO (PCT)
Prior art keywords
information
product
business person
business
living body
Prior art date
Application number
PCT/CN2019/120912
Other languages
English (en)
French (fr)
Inventor
陈佩
Original Assignee
深圳壹账通智能科技有限公司
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 深圳壹账通智能科技有限公司 filed Critical 深圳壹账通智能科技有限公司
Publication of WO2020233055A1 publication Critical patent/WO2020233055A1/zh

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/04Architecture, e.g. interconnection topology
    • G06N3/045Combinations of networks
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION 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
    • G06Q30/00Commerce
    • G06Q30/02Marketing; Price estimation or determination; Fundraising
    • G06Q30/0241Advertisements
    • G06Q30/0251Targeted advertisements
    • G06Q30/0269Targeted advertisements based on user profile or attribute
    • G06Q30/0271Personalized advertisement
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16Human faces, e.g. facial parts, sketches or expressions
    • G06V40/174Facial expression recognition
    • G06V40/176Dynamic expression

Definitions

  • This application relates to the field of artificial intelligence technology, and in particular to a product promotion method, device, equipment and storage medium based on living body detection.
  • Product promotion refers to the stage that a company's products (services) go through after they enter the market.
  • the product promotion methods adopted by enterprises are mainly a combination of online promotion and offline promotion, in order to better occupy the market and seize customers.
  • the company cannot monitor the business personnel who promote the product in an all-round way, and therefore cannot guarantee whether the product to be promoted assigned to the business personnel is suitable for the business personnel.
  • the business personnel cannot Accurately and vividly introduce the products to be promoted for product experiencers, which in turn leads to unsatisfactory product promotion effects.
  • the main purpose of this application is to provide a product promotion method, device, equipment and storage medium based on living body detection, which aims to solve the problem that the existing offline promotion method cannot fully monitor the business personnel and cannot guarantee the reasonable distribution of the products to be promoted. Sexual technical issues.
  • this application provides a product promotion method based on living body detection, and the method includes the following steps:
  • search for the second biometric information of the real holder of the user account from a pre-stored business personnel information database
  • the operation interface of the program displays the product information of the product to be promoted, so that the business personnel can promote the product to be promoted to experiencers according to the product information.
  • this application also proposes a product promotion device based on living body detection, which includes:
  • the receiving module is used to receive the login request of the product promotion application triggered by the business personnel;
  • An extraction module configured to extract a user account for logging in the product promotion application from the login request
  • the detection module is configured to perform a living body detection on the business person according to the login request to obtain the first biometric information of the business person;
  • the search module is configured to search for the second biometric information of the real holder of the user account from a pre-stored business personnel information database according to the user account;
  • the login module is used to compare the first biological characteristic information with the second biological characteristic information. If the first biological characteristic information matches the second biological characteristic information, perform a login operation, and The operation interface of the product promotion application program displays product information of the product to be promoted suitable for promotion by the business personnel, so that the business personnel can promote the product to be promoted to experiencers according to the product information.
  • this application also proposes a product promotion device based on living body detection, the device including: a memory, a processor, and computer-readable instructions stored on the memory and executable by the processor , Wherein when the computer-readable instructions are executed by the processor, the steps of the product promotion method based on living body detection as described above are realized.
  • this application also proposes a computer-readable storage medium with computer-readable instructions stored on the computer-readable storage medium.
  • the computer-readable instructions are executed by a processor, the implementation is as described above.
  • FIG. 1 is a schematic structural diagram of a product promotion device based on living body detection in a hardware operating environment involved in a solution of an embodiment of the present application;
  • FIG. 2 is a schematic flowchart of a first embodiment of a product promotion method based on living body detection in this application;
  • FIG. 3 is a schematic flowchart of a second embodiment of a product promotion method based on living body detection in this application;
  • FIG. 4 is a structural block diagram of a first embodiment of a product promotion device based on living body detection in this application.
  • FIG. 1 is a schematic structural diagram of a product promotion device based on living body detection in a hardware operating environment involved in a solution of an embodiment of the application.
  • the product promotion device based on living body detection may include a processor 1001, such as a central processing unit (Central Processing Unit, CPU), communication bus 1002, user interface 1003, network interface 1004, memory 1005.
  • the communication bus 1002 is used to implement connection and communication between these components.
  • the user interface 1003 may include a display screen (Display) and an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface.
  • the network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (WIreless-FIdelity, WI-FI) interface).
  • WIreless-FIdelity WI-FI
  • the memory 1005 may be a high-speed random access memory (Random Access Memory, RAM) memory, can also be a stable non-volatile memory (Non-Volatile Memory, NVM), such as disk storage.
  • RAM Random Access Memory
  • NVM Non-Volatile Memory
  • the memory 1005 may also be a storage device independent of the foregoing processor 1001.
  • FIG. 1 does not constitute a limitation on the product promotion equipment based on living body detection, and may include more or less components than shown, or a combination of certain components, or different Component arrangement.
  • a memory 1005 as a computer-readable storage medium may include an operating system, a network communication module, a user interface module, and computer-readable instructions.
  • the network interface 1004 is mainly used for data communication with a network server; the user interface 1003 is mainly used for data interaction with users; this application is used in the product promotion device based on living body detection
  • the processor 1001 and the memory 1005 may be set in a product promotion device based on living body detection.
  • the product promotion device based on living body detection calls the computer-readable instructions stored in the memory 1005 through the processor 1001 and executes the instructions provided in the embodiments of the present application.
  • the method of product promotion based on live detection is mainly used for data communication with a network server; the user interface 1003 is mainly used for data interaction with users; this application is used in the product promotion device based on living body detection
  • the processor 1001 and the memory 1005 may be set in a product promotion device based on living body detection.
  • the product promotion device based on living body detection calls the
  • the embodiment of the application provides a product promotion method based on living body detection.
  • FIG. 2 is a schematic flowchart of the first embodiment of a product promotion method based on living body detection of this application.
  • the product promotion method based on living body detection includes the following steps:
  • Step S10 receiving a login request for logging in to a product promotion application triggered by a business person, extracting a user account for logging in the product promotion application from the login request, and performing a live detection on the business person according to the login request, Obtain the first biometric information of the business person.
  • the executive body of this embodiment can be any mobile terminal installed with the above-mentioned product promotion application, such as the business staff’s own smart phone, tablet computer, personal computer, etc., or the aforementioned mobile terminal provided by the enterprise for the business staff. Terminal, there is no restriction here.
  • the product promotion application mentioned above can refer to the network application specially customized by the enterprise to assess the product promotion status of the business personnel, or it can be the office network application used for attendance, work communication and work arrangement within the enterprise. Program, there is no restriction here.
  • the first biological characteristic information mentioned in this embodiment may specifically be the face characteristic information, iris characteristic information, fingerprint characteristic information, voiceprint characteristic information, micro-expression characteristic information, etc. of the business personnel, and it will not be mentioned here. For a list, there are no restrictions on this.
  • the first biometric information is hereinafter referred to as the face Feature information, voiceprint feature information, and micro-expression feature information will be described as examples.
  • the detection module is configured to perform the following steps:
  • a living body detection instruction for collecting facial feature information is generated.
  • the generated living body detection instruction is specifically determined according to the predetermined type of the first biological feature information.
  • the generated life detection instruction is a life detection instruction for collecting facial feature information.
  • step (2) can be roughly as follows: according to the live detection instruction for collecting facial feature information, control the mobile terminal to turn on the camera, and display the preview on the user interface of the mobile terminal.
  • Set random content such as prompting the business personnel to read a displayed paragraph of text or a string of numbers, etc., and at the same time, record video information when the business personnel read the random preset content.
  • the reason why the business person is required to read the random preset content is to prevent others from using pre-photographed photos to impersonate.
  • the video information should include the facial image of the business person.
  • the facial feature information extraction model is specifically obtained by using a convolutional neural network algorithm to train the facial feature information in the face sample data obtained in advance.
  • the method of constructing the facial feature information extraction model can be roughly as follows:
  • a training model is constructed according to the facial feature information in the face sample data.
  • the training model is trained until a certain face image data is input, and the desired face feature information can be obtained, then the training of the training model can be completed.
  • the training model at this moment is the facial feature information extraction model.
  • the convolution kernel in the training model is a convolution kernel of size
  • the convolution kernel in order to increase the network depth of the training model and at the same time increase the training speed as much as possible, can be split into two Convolution kernel of size.
  • the face sample data can also be normalized to reduce the convolution kernel in each convolution layer and the fully connected layer as the output layer in the training process. The number of nodes to simplify various calculations in the training process.
  • the detection module is configured to perform the following steps:
  • the generated living body detection instruction is a living body detection instruction for collecting voiceprint feature information.
  • step (2) can be roughly as follows: according to the live detection instruction for collecting voiceprint feature information, control the mobile terminal to turn on the recording device, and display it on the user interface of the mobile terminal
  • the preset random content for example, prompts the business person to read a displayed paragraph of text or a string of numbers, etc., and at the same time, records audio information when the business person reads the random preset content.
  • the reason why the business person is required to read the random preset content is to prevent others from using pre-recorded audio information as a substitute.
  • the detection module is configured to perform the following steps:
  • micro-expression information may include not only facial expressions, but also body movements.
  • the micro-expression feature information in this embodiment is mainly for the facial expressions of the business person, so the generated life detection instructions are substantially the same as the life detection instructions for collecting the facial feature information.
  • the process of recording video information is roughly the same as when the first biometric information is facial feature information, the method of recording video information is roughly the same. I won't repeat it here.
  • the video information should include the face image of the business person.
  • the consecutive face images extracted from the video information based on the face detection model are P1, P2, P3.
  • the face detection model may also be obtained by training a pre-collected training model based on a convolutional neural network algorithm.
  • micro-expression recognition technology analyze and process each face image in order to obtain the micro-expression feature information of the business person.
  • the micro-expression feature information mentioned here refers to the actions taken by the business personnel according to the random preset content, for example, the random preset content is to ask the business staff to read "123", and When blinking at the same time, the extracted micro-expression feature information needs to include the mouth movement when the business person reads the random preset content "123" and the corresponding blinking movement.
  • micro-expression recognition operation since the main feature of micro-expression is short duration and difficult to recognize. Therefore, in order to ensure the accuracy of the extracted micro-expression feature information, when performing a micro-expression recognition operation, it is first necessary to perform a pre-processing operation on the micro-expression image, that is, to pre-process the personal face image.
  • micro-expression detection mentioned here refers to the use of an edge detection algorithm to perform an edge detection operation on the grayscale image after the preprocessing operation is performed, and then extract the facial features of the business person in the face image.
  • the extracted facial features can be marked with feature points. That is, the contours of the facial features are drawn with feature points.
  • micro-expression database can be an existing and already popular micro-expression database, such as CASME micro-expression database, Polikovsky database, SMIC database, etc., which will not be listed here, but can also be There is no restriction on the micro expression database built by the enterprise itself.
  • Step S20 According to the user account, search for the second biometric information of the real holder of the user account from a pre-stored business person information database.
  • the second biometric information is the biometric information collected by the mobile terminal when the holder registers the user account.
  • the first biometric information and the second biometric information in step S30 need to be of the same type, so that the feature comparison Will be consistent.
  • the above-mentioned collection process of the second characteristic information can refer to the several methods given in the above-mentioned step S10, which will not be repeated here and will not be limited.
  • Step S30 Perform feature comparison between the first biological feature information and the second biological feature information, and if the first biological feature information matches the second biological feature information, perform a login operation and log in to the
  • the operation interface of the product promotion application program displays product information of the product to be promoted suitable for the business person, so that the business person can promote the product to be promoted to the experiencer according to the product information.
  • the current location information of the machine when acquiring the current location information of the machine, it may be specifically by reading the global positioning system (Global Positioning System) built in the mobile terminal.
  • GPS Global Positioning System
  • collected location information such as latitude and longitude
  • the current address of the business person is determined according to the latitude and longitude and the map built into the mobile terminal.
  • the historical employment information mentioned here refers to the information of the arbitrary generation link of the product to be promoted that the business person previously participated in.
  • the historical employment information mentioned in this embodiment mainly includes the product information of the products promoted in the product promotion activities that the business person has participated in before, and the links that the business person is responsible for in the product promotion activities. , The promotion effect of the product promotion activities, etc., will not be listed here, and there will be no restrictions on this.
  • keywords are extracted from the historical employment information.
  • semantic analysis is performed on the keywords by using semantic analysis technology, and the characteristic information of the products to be promoted suitable for promotion by the business personnel is determined according to the semantic information of the keywords.
  • the feature information is matched with the product information of the product to be promoted currently launched by the enterprise.
  • pre-setting the weight ratios of the product features corresponding to various products to be promoted and then calculate the weight values of the features that are the same or similar to the product features in the feature information of the products to be promoted suitable for promotion by the business personnel according to the weight ratios, Obtain the matching degree between the business personnel and each product to be promoted launched by the enterprise. Then, it is determined whether the matching degree of each product to be promoted is greater than a certain preset threshold, and is greater than the preset threshold, then the product to be promoted is considered suitable for promotion by the business person.
  • the product promotion skills of the business personnel match the product information of the products to be promoted currently launched by the enterprise, the currently matched product to be promoted is determined as the product to be promoted suitable for promotion by the business personnel.
  • the extracted keywords are mainly It is a keyword that can identify the characteristics of product P.
  • the characteristic information of the product to be promoted suitable for promotion by the business personnel can be determined according to the semantic information.
  • the determined characteristic information includes characteristic A, characteristic B, and characteristic C.
  • (2-3) Use a pre-built product analysis model to analyze the product to be promoted and the resident address information, and determine the product promotion address that the business person needs to go to.
  • (2-32) Input the first input parameter into the product analysis model, and perform analysis processing through the product analysis model to obtain at least one first promotion address;
  • (2-34) Input the second input parameter into the product analysis model, and analyze and process the product analysis model to determine the product promotion address that the business person needs to go to.
  • the product promotion address that the business person needs to go to is one of the at least one first promotion address obtained above and the coordinate difference of the resident address information is less than the preset threshold.
  • the process of obtaining at least one first promotion address by inputting the first input parameters into the product analysis model for analysis processing may be to first determine the applicable population of the product to be promoted according to the first input parameter (ie the product information of the product to be promoted), and then determine the applicable population according to the applicable population The gathering address of the applicable crowd, so that the gathering address of the applicable crowd is determined as the first promotion address.
  • the first input parameter ie the product information of the product to be promoted
  • this embodiment provides a specific implementation method for constructing the product analysis model, which is roughly as follows:
  • (A) Receive data collection instructions and extract sample data to be collected from the data collection instructions (product information of products similar to the products to be promoted by the enterprise and the permanent address information of the business personnel participating in the promotion of the product)
  • Network address such as the network access address of a big data platform, or the product information storage access address of the product to be promoted by the company.
  • (B) Configure a web crawler according to the network address, and use the web crawler to obtain the sample data from a web page corresponding to the network address.
  • the product analysis model constructed is used to analyze the business person’s product to be promoted and the business person’s permanent address information at the time of determination. Therefore, the training data and test data obtained by the division must contain the product information of the product similar to the product to be promoted by the enterprise and the permanent address information of the business personnel participating in the promotion of the product.
  • sample data can also be cleaned according to preset processing rules, such as data Deduplication, format conversion, etc. will not be listed here, and there will be no restrictions on this.
  • (D) Mark the training data and the test data separately, and determine the input training data and the output training data, as well as the input test data and the output test data according to the marked training data.
  • the use of the input training data and the output training data to train the training model referred to here specifically refers to using the input training data as input parameters and inputting the input layer of the training model, and then Data processing is performed through the hidden layer, convolutional layer, and pooling layer in the training model, and the output layer of the training model outputs the processing result, and finally the output processing result is compared with the training output result.
  • the processing result matches the output training data, or the degree of matching meets a preset threshold, it can be considered that the training of the training model is completed, and the initial product analysis model is obtained.
  • the product to be promoted will be The product information of the product is used as the first input parameter, and the difference between the coordinates of the first promotion address and the coordinates of the resident address information determined based on the analysis of the first input parameter is used as the second input parameter, and then according to the second input parameter, the final process result.
  • the product information in the input training data is first input into the training model, and the output first promotion address The information is compared with the promotion address information corresponding to the product information in the output training data.
  • the coordinates corresponding to the first promotion address information and the coordinates of the resident address information in the input training data are made difference, and the difference is re-input to the training model to obtain the second promotion Address information, and compare the second promotion address information with the corresponding promotion address information in the output training data.
  • the initial product analysis model can be used after determining the product promotion address that the business person needs to go to, that is, the initial product analysis model can be used as the aforementioned pre-built product analysis model.
  • the login operation is performed.
  • the login operation needs to be executed when the first biometric information matches the second biometric information, and the current address of the business person matches the product promotion address.
  • the product promotion method based on living body detection when receiving a login request triggered by a business person to log in to a product promotion application, obtains all the information by performing a living body detection on the business person.
  • the first biometric information of the business person at the same time, extract the user account for logging in the product promotion application from the login request, and find the user account from the pre-stored business person information database according to the user account
  • the second biometric information of the true holder finally, by comparing the first biometric information with the second biometric information, the first biometric information is compared with the second biometric information.
  • the product promotion can accurately locate the actual participants, it is possible to better monitor the business personnel in the entire product promotion process, so as to better supervise the business personnel to actively participate in product promotion, thereby improving product promotion effect.
  • FIG. 3 is a schematic flowchart of a second embodiment of a product promotion method based on living body detection according to this application.
  • the product promotion method based on living body detection in this embodiment after the step S30, further includes:
  • Step S40 Collect the facial expressions of the experiencer and the audio information and video information of the business personnel in the process of describing the product to be promoted.
  • the method of collecting facial expressions of the experiencer in practical applications, you can refer to the method of collecting micro-expression feature information given in the first embodiment above, that is, it is necessary to describe to the business personnel to be promoted.
  • the product process first record the video information containing the face image of the experiencer; then, based on the preset face detection model, extract the continuous face of the experiencer at a certain moment from the video information.
  • the facial images are analyzed and processed in order to obtain the micro-expression feature information of the experiencer; finally, the entire product is introduced
  • the micro-expression feature information of the experiencer is used as the facial expression change of the experiencer.
  • the video information is specifically the recorded video information containing the face images, body movements, etc. of the business person when the business person explains the product to be promoted.
  • Step S50 Send the audio information, the video information and the facial expressions of the experiencer to a server, so that the server determines the facial expressions of the experiencer based on the audio information, the video information and the experiencer’s facial expressions. State the promotion effect of the product to be promoted, and adjust the promotion plan according to the promotion effect.
  • the server is essentially the server of the enterprise that launched the product to be promoted.
  • the server may be a traditional server occupying actual physical space, or a virtual cloud server.
  • the server may specifically analyze and process the information by using big data analysis technology to determine the product to be promoted The promotion effect.
  • the facial expression of the experiencer is Drowsiness; for example, moving upward at the characteristic point of the upper lip, marking the characteristic point of the lower lip and the characteristic point of the upper lip moving upward, causing the upper lip to lift up, and the lower lip and the upper lip are tightly closed, and the lower corner of the mouth is slightly raised;
  • the feature points that identify the inner corners of the eyebrows move toward the center of the eyebrows, causing the inner corners of the eyebrows to be wrinkled together, and the eyebrows are raised.
  • the experiencer s facial expression is considered doubtful; for example, the feature points that identify the corners of the lip move backward and above the cheek When the movement causes the corners of the lips to be pulled back and raised, and the characteristic points that mark the mouth move outwards, which causes the mouth to be widened, the facial expression of the experiencer can usually be considered as satisfactory.
  • the facial expression of the experiencer when determining the promotion effect of the product to be promoted according to the audio information, the video information, and the facial expression of the experiencer, specifically, each During the time period, the facial expression of the experiencer is combined with the audio information and video information describing the product to be promoted by the business person, so as to determine which content the experiencer is more satisfied with and which content is not explained by the business person If satisfied, the promotion effect of the product to be promoted is determined.
  • the promotion plan when adjusting the promotion plan according to the promotion effect, it may specifically be the place where the experiencer is dissatisfied with the audio information explained by the business personnel, by analyzing the content, tone, and all of the audio information in the audio information. The service attitude of the business personnel in the video information and then determine the specific adjustment plan.
  • the product promotion method based on living body detection performs the operation of logging in to the product promotion application after determining that the first biometric information matches the second biometric information Afterwards, by collecting the facial expressions of the experiencer and the audio information and video information of the business personnel in the process of describing the product to be promoted, and collecting the audio information, the video information and the facial expressions of the experiencer Sent to the server, so that the server can determine the promotion effect of the product to be promoted according to the audio information, the video information, and the facial expression of the experiencer, and adjust the promotion plan according to the promotion effect, so that the product The promotion plan can better meet the needs of users and further enhance the promotion effect.
  • the embodiments of the present application also provide a computer-readable storage medium, and the computer-readable storage medium may be a non-volatile readable storage medium.
  • the computer-readable storage medium of the present application stores computer-readable instructions, and when the computer-readable instructions are executed by a processor, the steps of the above-mentioned product promotion method based on living body detection are realized.
  • the method implemented when the computer-readable instruction is executed please refer to the various embodiments of the product promotion method based on living body detection in this application, which will not be repeated here.
  • FIG. 4 is a structural block diagram of a first embodiment of a product promotion device based on living body detection in this application.
  • the product promotion device based on living body detection proposed in the embodiment of the present application includes:
  • the receiving module 4001 is used to receive a login request triggered by a business person to log in to the product promotion application; the extraction module 4002 is used to extract the user account that logs in the product promotion application from the login request; the detection module 4003 is used to Perform a live detection on the business person according to the login request to obtain the first biometric information of the business person; the search module 4004 is used to search for the user from a pre-stored business person information database according to the user account The second biometric information of the real holder of the account; the login module 4005 is used to compare the first biometric information with the second biometric information.
  • the login operation is performed, and the product information of the product to be promoted is displayed on the operation interface of the product promotion application, so that the business person can promote the product to be promoted to the experiencer based on the product information product.
  • the virtual function modules of the above-mentioned living body detection-based product promotion device are stored in the memory 1005 of the living body detection-based product promotion device shown in FIG. 1, and are used to implement all the functions of computer-readable instructions; each module is used by the processor 1001 When executed, it can obtain the first and second biometric information of the business personnel, and compare and determine the products to be promoted according to the obtained biometric information, and complete the function of the whole process of product promotion based on living body detection.
  • the detection module includes: a first instruction generating unit configured to generate a living body detection instruction for collecting facial feature information according to the login request; first The collection unit is configured to collect video information when the business person reads random preset content according to the living body detection instruction, and the video information includes the face image of the business person; the first extraction unit is configured to The preset facial feature information extraction model performs facial feature extraction on the facial image in the video information to obtain the facial feature information of the business person.
  • the detection module further includes: a second instruction generating unit configured to generate a living body detection instruction for collecting voiceprint feature information according to the login request;
  • the second collection unit is used to collect the audio information when the business personnel read the random preset content according to the living body detection instruction;
  • the second extraction unit is used to extract the model based on the preset voiceprint feature information, Perform voiceprint feature extraction on the audio information to obtain voiceprint feature information of the business person.
  • the detection module further includes: a third instruction generating unit, configured to generate a living body detection instruction for collecting micro-expression feature information according to the login request;
  • the third collection unit is configured to collect the video information when the business person reads the randomly preset content according to the living body detection instruction, the video information contains the face image of the business person;
  • the sorting unit is used to The preset face detection model extracts the continuous face images of the business personnel at a certain moment from the video information, and sorts the face images in chronological order;
  • the analysis unit is used to analyze the facial images according to the micro-expression
  • the recognition technology analyzes and processes each face image in sequence to obtain the micro-expression feature information of the business person.
  • the product promotion device based on living body detection further includes: a first determining module configured to obtain the current location information of the machine, and determine the current address of the business person according to the location information; The second determining module is used to determine the product promotion address that the business person needs to go to according to the user account; the judgment module is used to determine whether the current address of the business person matches the product promotion address.
  • the login module is further configured to perform all the operations when the first biometric information matches the second biometric information, and the current address of the business person matches the product promotion address. Describe the login operation.
  • the second determination module includes: a search subunit, which is used to search for the business personnel’s historical practice information and resident address information from the business personnel information database according to the user account; the product to be promoted is determined The subunit is used to determine the product to be promoted suitable for promotion by the business personnel based on the historical employment information; the product promotion address determination subunit is used to use a pre-built product analysis model to analyze the product to be promoted and the product to be promoted. The resident address information is analyzed to determine the product promotion address that the business person needs to go to.
  • a search subunit which is used to search for the business personnel’s historical practice information and resident address information from the business personnel information database according to the user account
  • the product to be promoted is determined
  • the subunit is used to determine the product to be promoted suitable for promotion by the business personnel based on the historical employment information
  • the product promotion address determination subunit is used to use a pre-built product analysis model to analyze the product to be promoted and the product to be promoted.
  • the resident address information is analyzed to determine the product promotion address
  • the product promotion device based on living body detection further includes: a collecting unit for collecting the facial expressions of the experiencer and audio information and video information of the business personnel in the process of describing the product to be promoted; and a sending unit , Used to send the audio information, the video information and the facial expressions of the experiencer to a server, so that the server determines the facial expressions of the experiencer based on the audio information, the video information and the experiencer State the promotion effect of the product to be promoted, and adjust the promotion plan according to the promotion effect.
  • each module in the above-mentioned living body detection-based product promotion device corresponds to the steps in the above embodiment of the above-mentioned living body detection-based product promotion method, and the functions and realization processes are not repeated here.
  • the method of the embodiment can be implemented by means of software plus a necessary general hardware platform, of course, it can also be implemented by hardware, but the former is a better implementation in many cases.
  • the application s The essence of the technical solution or the part that contributes to the existing technology can be embodied in the form of a software product.
  • the computer software product is stored in a storage medium (such as Read Only Memory). Memory, ROM)/RAM, disk, optical
  • the disk includes several instructions to make a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) execute the method described in each embodiment of the present application.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • Accounting & Taxation (AREA)
  • Development Economics (AREA)
  • General Health & Medical Sciences (AREA)
  • Finance (AREA)
  • Strategic Management (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Marketing (AREA)
  • Economics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Game Theory and Decision Science (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • General Business, Economics & Management (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Human Computer Interaction (AREA)
  • Multimedia (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

一种基于活体检测的产品推广方法、装置、设备及存储介质。该方法包括:接收业务人员触发的登录产品推广应用程序的登录请求,从登录请求中提取登录产品推广应用程序的用户账号,并根据登录请求对业务人员进行活体检测,得到业务人员的第一生物特征信息(S10);根据用户账号,从预存的业务人员信息库中查找用户账号的真实持有者的第二生物特征信息(S20);将第一生物特征信息与第二生物特征信息进行特征对比,若第一生物特征信息与第二生物特征信息匹配,则执行登录操作,并在产品推广应用程序的操作界面显示适合业务人员推广的待推广产品的产品信息,以使所述业务人员根据所述产品信息向体验者推广所述待推广产品(S30)。通过上述方法,既实现了对业务人员的监控,又保证了产品分配的合理性。

Description

基于活体检测的产品推广方法、装置、设备及存储介质
本申请要求于2019年5月21日提交中国专利局、申请号为201910432587.6、发明名称为“基于活体检测的产品推广方法、装置、设备及存储介质”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及人工智能技术领域,尤其涉及一种基于活体检测的产品推广方法、装置、设备及存储介质。
背景技术
产品推广是指企业产品(服务)问世后,进入市场所经过的一个阶段。目前,企业采取的产品推广方式主要是线上推广和线下推广相结合的方式,以更好的占领市场,抢占客户。
然而,传统的线下推广方式,企业不能对进行产品推广的业务人员进行全方位的监控,因而无法保证分配给业务人员的待推广产品是否适合业务人员,从而导致产品推广过程中,业务人员无法为产品体验者进行准确、形象的介绍待推广产品,进而导致产品推广效果不够理想。
所以,亟需提供一种既能够实现对业务人员的监控,又可以保证了待推广产品分配的合理性的产品推广方法,以保证产品推广效果。
发明内容
本申请的主要目的在于提供一种基于活体检测的产品推广方法、装置、设备及存储介质,旨在解决现有线下推广方式无法对对业务人员进行全方位监控,同时无法保证待推广产品分配合理性的技术问题。
为实现上述目的,本申请提供了一种基于活体检测的产品推广方法,所述方法包括以下步骤:
接收业务人员触发的登录产品推广应用程序的登录请求,从所述登录请求中提取登录所述产品推广应用程序的用户账号,并根据所述登录请求对所述业务人员进行活体检测,得到所述业务人员的第一生物特征信息;
根据所述用户账号,从预存的业务人员信息库中查找所述用户账号的真实持有者的第二生物特征信息;
将所述第一生物特征信息与所述第二生物特征信息进行特征对比,若所述第一生物特征信息与所述第二生物特征信息匹配,则执行登录操作,并在所述产品推广应用程序的操作界面显示待推广产品的产品信息,以使所述业务人员根据所述产品信息向体验者推广所述待推广产品。
此外,为实现上述目的,本申请还提出一种基于活体检测的产品推广装置,所述装置包括:
接收模块,用于接收业务人员触发的登录产品推广应用程序的登录请求;
提取模块,用于从所述登录请求中提取登录所述产品推广应用程序的用户账号;
检测模块,用于根据所述登录请求对所述业务人员进行活体检测,得到所述业务人员的第一生物特征信息;
查找模块,用于根据所述用户账号,从预存的业务人员信息库中查找所述用户账号的真实持有者的第二生物特征信息;
登录模块,用于将所述第一生物特征信息与所述第二生物特征信息进行特征对比,若所述第一生物特征信息与所述第二生物特征信息匹配,则执行登录操作,并在所述产品推广应用程序的操作界面显示适合所述业务人员推广的待推广产品的产品信息,以使所述业务人员根据所述产品信息向体验者推广所述待推广产品。
此外,为实现上述目的,本申请还提出一种基于活体检测的产品推广设备,所述设备包括:存储器、处理器及存储在所述存储器上并可被所述处理器执行的计算机可读指令,其中所述计算机可读指令被所述处理器执行时,实现如上文所述的基于活体检测的产品推广方法的步骤。
此外,为实现上述目的,本申请还提出一种计算机可读存储介质,所述计算机可读存储介质上存储有计算机可读指令,其中所述计算机可读指令被处理器执行时,实现如上文所述的基于活体检测的产品推广方法的步骤。
本申请的一个或多个实施例的细节在下面的附图和描述中提出。本申请的其他特征和优点将从说明书、附图以及权利要求书变得明显。
附图说明
图1是本申请实施例方案涉及的硬件运行环境的基于活体检测的产品推广设备的结构示意图;
图2为本申请基于活体检测的产品推广方法第一实施例的流程示意图;
图3为本申请基于活体检测的产品推广方法第二实施例的流程示意图;
图4为本申请基于活体检测的产品推广装置第一实施例的结构框图。
本申请目的的实现、功能特点及优点将结合实施例,参照附图做进一步说明。
具体实施方式
应当理解,此处所描述的具体实施例仅用以解释本申请,并不用于限定本申请。
参照图1,图1为本申请实施例方案涉及的硬件运行环境的基于活体检测的产品推广设备结构示意图。
如图1所示,该基于活体检测的产品推广设备可以包括:处理器1001,例如中央处理器(Central Processing Unit,CPU),通信总线1002、用户接口1003,网络接口1004,存储器1005。其中,通信总线1002用于实现这些组件之间的连接通信。用户接口1003可以包括显示屏(Display)、输入单元比如键盘(Keyboard),可选用户接口1003还可以包括标准的有线接口、无线接口。网络接口1004可选的可以包括标准的有线接口、无线接口(如无线保真(WIreless-FIdelity,WI-FI)接口)。存储器1005可以是高速的随机存取存储器(Random Access Memory,RAM)存储器,也可以是稳定的非易失性存储器(Non-Volatile Memory,NVM),例如磁盘存储器。存储器1005可选的还可以是独立于前述处理器1001的存储装置。
本领域技术人员可以理解,图1中示出的结构并不构成对基于活体检测的产品推广设备的限定,可以包括比图示更多或更少的部件,或者组合某些部件,或者不同的部件布置。
如图1所示,作为一种计算机可读存储介质的存储器1005中可以包括操作系统、网络通信模块、用户接口模块以及计算机可读指令。在图1所示的基于活体检测的产品推广设备中,网络接口1004主要用于与网络服务器进行数据通信;用户接口1003主要用于与用户进行数据交互;本申请基于活体检测的产品推广设备中的处理器1001、存储器1005可以设置在基于活体检测的产品推广设备中,所述基于活体检测的产品推广设备通过处理器1001调用存储器1005中存储的计算机可读指令,并执行本申请实施例提供的基于活体检测的产品推广方法。
本申请实施例提供了一种基于活体检测的产品推广方法,参照图2,图2为本申请一种基于活体检测的产品推广方法第一实施例的流程示意图。
本实施例中,所述基于活体检测的产品推广方法包括以下步骤:
步骤S10,接收业务人员触发的登录产品推广应用程序的登录请求,从所述登录请求中提取登录所述产品推广应用程序的用户账号,并根据所述登录请求对所述业务人员进行活体检测,得到所述业务人员的第一生物特征信息。
具体的说,本实施例的执行主体可以是任意安装有上述产品推广应用程序的移动终端,比如业务人员自己的智能手机、平板电脑、个人计算机等,也可以是企业为业务人员配备的上述移动终端,此处不做限制。
此外,上述所说的产品推广应用程序可以是指企业为了考核业务人员的产品推广情况而专门定制的网络应用程序,也可以是企业内部用于考勤、工作交流、工作安排于一体的办公网络应用程序,此处不做限制。
应当理解的是,上述所说的“产品推广应用程序”仅仅是为了表明该网络应用程序具备认证业务人员身份的功能。
此外,本实施例中所说的第一生物特征信息,具体可以是业务人员的人脸特征信息、虹膜特征信息、指纹特征信息、声纹特征信息、微表情特征信息等,此处不再一一列举,对此也不做任何限制。
为了便于理解上述步骤S10中所说的根据所述登录请求对所述业务人员进行活体检测,得到所述业务人员的第一生物特征信息的操作,以下以所述第一生物特征信息分别为人脸特征信息、声纹特征信息、微表情特征信息为例进行说明。
具体的说,在所述第一生物特征信息为人脸特征信息时,所述检测模块用于执行如下步骤:
(1)根据所述登录请求,生成采集人脸特征信息的活体检测指令。
具体的说,在实际应用中,生成的活体检测指令具体是根据预先规定的第一生物特征信息的类型确定的。
也就是说,在所述第一生物特征信息为人脸特征信息时,生成的活体检测指令便是采集人脸特征信息的活体检测指令。
(2)根据所述活体检测指令,采集所述业务人员读取随机预设内容时的视频信息。
具体的说,在实际应用中,步骤(2)的操作,大致可以是:根据采集人脸特征信息的活体检测指令,控制所述移动终端开启摄像头,并在所述移动终端的用户界面显示预设的随机内容,比如提示所述业务人员读取显示的一段文字或一串数字等,同时在所述业务人员读取所述随机预设内容时,录制视频信息。
应当理解的是,本实施中在采集所述业务人员的视频信息时,之所以要所述业务人员读取随机预设内容,是为了避免他人使用预先拍摄的照片冒名顶替。
此外,需要说明的是,为了保证后续步骤(3)中提取到的人脸特征信息较为精致,所述视频信息应当包含所述业务人员的人脸图像。
(3)基于预设的人脸特征信息提取模型,对所述视频信息中的人脸图像进行人脸特征提取,得到所述业务人员的人脸特征信息。
需要说明的是,在本实施例中,所述人脸特征信息提取模型具体是采用卷积神经网络算法对预先获取的人脸样本数据中的人脸特征信息训练获得的。
关于构建所述人脸特征信息提取模型的方式,大致可以如下所述:
首先,根据所述人脸样本数据中的人脸特征信息构建训练模型。
然后,基于所述卷积神经网络算法,对所述训练模型进行训练,直到输入某一人脸图像数据,可以得到想要的人脸特征信息为止,便可以完成对所述训练模型的训练。
相应地,此刻的训练模型便是所述人脸特征信息提取模型。
进一步地,在实际应用中,为了增加训练模型的网络深度,使得训练出的人脸特征信息提取模型的提取精度更加准确,在采用所述卷积神经网络算法,对所述训练模型进行训练之前,还可以先对所述训练模型中的初始卷积核进行拆分。
比如,在训练模型中的初始卷积核为一个尺寸为的卷积核时,为了增加训练模型的网络深度,同时尽可能的提升训练速度,可以将所述的卷积核拆分为两个尺寸为的卷积核。
进一步地,为了加速后续训练过程中人脸特征信息提取模型的收敛速度,并且在一定程度上提升人脸特征信息提取模型的泛化能力(机器学习算法对新鲜样本的适应能力),在根据所述人脸样本数据构建所述训练模型之前,还可以对所述人脸样本数据进行归一化处理,从而缩小训练过程中每层卷积层中卷积核以及作为输出层的全连接层中的节点数,进而简化训练过程中的各种计算。
此外,在所述第一生物特征信息为声纹特征信息时,所述检测模块用于执行如下步骤:
(1)根据所述登录请求,生成采集声纹特征信息的活体检测指令。
应当理解的是,在所述第一生物特征信息为声纹特征信息时,生成的活体检测指令便是采集声纹特征信息的活体检测指令。
(2)根据所述活体检测指令,采集所述业务人员读取随机预设内容时的音频信息。
具体的说,在实际应用中,步骤(2)的操作,大致可以是:根据采集声纹特征信息的活体检测指令,控制所述移动终端开启录音装置,并在所述移动终端的用户界面显示预设的随机内容,比如提示所述业务人员读取显示的一段文字或一串数字等,同时在所述业务人员读取所述随机预设内容时,录制音频信息。
应当理解的是,本实施中在采集所述业务人员的音频信息时,之所以要所述业务人员读取随机预设内容,是为了避免他人使用预先录制的音频信息冒名顶替。
(3)基于预设的声纹特征信息提取模型,对所述音频信息进行声纹特征提取,得到所述业务人员的声纹特征信息。
具体的说,关于所述声纹特征信息提取模型的构建,可以参照上述所说的人脸特征信息提取模型的构建,具体的构建过程,此处不再赘述。
此外,在所述第一生物特征信息为微表情特征信息时,所述检测模块用于执行如下步骤:
(1)根据所述登录请求,生成采集微表情特征信息的活体检测指令。
应当理解的是,在实际应用中,微表情信息不仅可以包括面部表情,还可以包括肢体动作等。
为了便于说明,本实施例中的微表情特征信息主要是针对所述业务人员的面部表情,因而生成的活体检测指令与采集所述人脸特征信息的活体检测指令大致相同。
(2)根据所述活体检测指令,采集所述业务人员读取随机预设内容时的视频信息。
由于此处采集的微表情特征信息主要是针对所述业务人员的面部表情,因而录制的视频信息的过程与所述第一生物特征信息为人脸特征信息时,录制视频信息的方式大致相同,此处不再赘述。
同样,所述视频信息应当包含所述业务人员的人脸图像。
(3)基于预设的人脸检测模型,从所述视频信息中提取出所述业务人员在某一时刻的连续人脸图像,并按照时间顺序为各人脸图像进行排序。
比如说,在T1、T2、T3,三个连续时刻,基于人脸检测模型从所述视频信息中提取出的连续人脸图像为P1、P2、P3。
需要说明的是,所述人脸检测模型同样可以是基于卷积神经网络算法对预先收集的训练模型训练获得的。
关于人脸检测模型的训练过程,可以参照上述人脸特征信息提取模型的训练过程,此处不再赘述。
(4)根据微表情识别技术,按序对各人脸图像进行分析处理,得到所述业务人员的微表情特征信息。
具体的说,此处所述的微表情特征信息是指所述业务人员在根据所述随机预设内容做出的动作,比如随机预设内容为要求所述业务人员读出“123”,并同时眨眼睛,则提取到的微表情特征信息需要包括所述业务人员读取随机预设内容“123”时的嘴巴动作,以及配合的眨眼睛动作。
为了更好的理解上述步骤(4)中所说的微表情特征信息的提取,以下进行具体说明,其大致过程可以分为如下4个子步骤:
(4-1)微表情图像预处理。
应当理解的是,由于微表情的主要特征是持续时间短且难以识别。因而为了保证提取到的微表情特征信息的精确度,在进行微表情识别操作时,首先需要对微表情图像进行预处理操作,即对个人脸图像进行预处理。
比如,去噪、增强对比度、灰度处理等,此处不再一一列举,对此也不做任何限制。
(4-2)微表情检测。
具体的说,此处所说的微表情检测,是指采用边缘检测算法对上述进行预处理操作后的灰度图像进行边缘检测操作,进而提取所述人脸图像中所述业务人员的五官轮廓。
(4-3)微表情序列特征提取。
具体的说,为了便于后续微表情识别,可以对提取出的五官轮廓进行特征点标记。即,将五官轮廓用特征点勾勒出。
(4-4)微表情识别。
具体的说,由于连续时间内,各人脸图像中提取出的五官轮廓已经进行了特征点标记。因而,在进行微表情识别时,只需识别标识同一位置的特征点,在不同人脸图像中的位置变化,然后记录同一特征点从开始帧(即第一张人脸图像)到结束帧(最后一张人脸图像)中位置的变化。然后,将得到各个标记点的位置变化,与预先构建的微表情数据库中不同微表情中相应特征的变化趋势进行对比即可。
比如,在标识眼内角的上眼皮的特征点向下运动,导致眼内角的上眼皮降低,标识嘴巴的特征点向外运动,导致出嘴巴张大,根据这些信息与微表情数据库中的微表情进行匹配,最终可以确定所述体验者的面部表情为困倦。
需要说明的是,上述所说的微表情数据库可以是现有,且已经较为普及的微表情数据库,比如CASME微表情库、Polikovsky数据库、SMIC数据库等,此处不再一一列举,也可以是企业自己构建的微表情数据库,此处不做限制。
应当理解的是,以上给出的仅为几种具体的生物特征信息的获取方式,对本申请的技术方案并不构成任何限定,在具体实现中,本领域的技术人员可以根据需要进行设置,此处不做限制。
此外,值得一提的是,上述用到的各种生物特征提取模型,均为采用机器学习算法,对从各大数据平台收集到的样本数据进行训练获得的,具体的构建方式本领域的技术人员可以参考其选取的具体机器学习算法的文档,此处不再赘述。
步骤S20,根据所述用户账号,从预存的业务人员信息库中查找所述用户账号的真实持有者的第二生物特征信息。
应当理解的是,此处所说的所述用户账号的真实持有者,即为注册所述用户账号的业务人员。
相应地,所述第二生物特征信息便是所述持有者在注册所述用户账号时,移动终端采集到的生物特征信息。
此外,为了保证步骤S30中,第一生物特征信息和第二生物特征信息的对比操作有对比性,在实际应用中,第一生物特征信息和第二生物特征信息需要为同一类型,这样特征对比才会一致。
此外,上述第二特征信息的采集过程,可以参照上述步骤S10中给出的几种方式,此处不再赘述,对此也不做限制。
步骤S30,将所述第一生物特征信息与所述第二生物特征信息进行特征对比,若所述第一生物特征信息与所述第二生物特征信息匹配,则执行登录操作,并在所述产品推广应用程序的操作界面显示适合所述业务人员的待推广产品的产品信息,以使所述业务人员根据所述产品信息向体验者推广所述待推广产品。
具体的说,在进行特征对比时,可以是通过将两个生物特征信息进行遍历,并将遍历到的同一位置的特征点进行单独对比,如果在将两个生物特征信息包含的所有特征点全部进行对比后,匹配度满足某一预设阈值就认为两者匹配,则可以认为进行登录操作的业务人员即为所述用户账号的真实持有者。
进一步地,为了确保参与到产品推广的业务人员的实际推广情况,即业务人员确实前往了企业规定的产品推广地址,而并非在任意地点进行的登录操作。在确定所述第一生物特征信息与所述第二生物特征信息匹配,执行登录操作之前,还可以先判断一下所述业务人员当前所处的位置是否是规定的产品推广地址。
为了便于理解,以下给出一种具体的实现方式,大致如下:
(1)获取本机当前所处的位置信息,并根据所述位置信息确定所述业务人员当前所处的地址。
应当理解的是,在获取本机当前所处的位置信息时,具体可以是通过读取所述移动终端内置的全球定位系统(Global Positioning System,GPS)采集到的位置信息,比如经纬度;然后,根据所述经纬度及所述移动终端内置的地图确定所述业务人员当前所处的地址。
(2)根据所述用户账号,确定所述业务人员需要前往的产品推广地址。
关于步骤(2)中根据所述用户账号,确定所述业务人员需要前往的产品推广地址的操作,大致可以通过如下方式实现:
(2-1)根据所述用户账号,从所述业务人员信息库中查找所述业务人员的历史从业信息和常住地址信息。
具体的说,此处所说的历史从业信息是指所述业务人员之前参与到的待推广产品任意生成环节的信息。
(2-2)根据所述历史从业信息,确定适合所述业务人员推广的所述待推广产品。
具体的说,本实施例中所说的历史从业信息,主要包括所述业务人员之前参与过的产品推广活动中负责推广的产品的产品信息,以及所述业务人员在产品推广活动中负责的环节、该产品推广活动的推广效果等,此处不再一一列举,对此也不做任何限制。
相应地,在根据所述历史从业信息,确定适合所述业务人员推广的所述待推广产品时,具体可以按照如下方式实现:
首先,基于关键词提取技术,从所述历史从业信息中提取关键词。
然后,采用语义分析技术,对所述关键词进行语义分析,根据所述关键词的语义信息确定适合所述业务人员推广的待推广产品具备的特征信息。
接着,将所述特征信息与企业当前推出的待推广产品的产品信息进行匹配。
比如,预先设置各种待推广产品对应的产品特征的权重比例,然后根据权重比例计算适合所述业务人员推广的待推广产品具备的特征信息中,与产品特征相同或相似的特征的权重值,得到所述业务人员与企业推出的各个待推广产品的匹配度。然后,判断各个待推广产品的匹配度是否大于某一预设阈值,大于预设的阈值,则认为所述待推广产品适合所述业务人员推广。
最后,若所述业务人员具备的产品推广技能与企业当前推出的待推广产品的产品信息匹配,则将当前匹配的待推广产品确定为所述适合所述业务人员推广的待推广产品。
为了便于理解,以下结合实例说明:比如说,在所述业务人员参与过产品P的推广工作,且负责了产品P的主要推广工作,并取得了良好的推广效果,则提取到的关键词主要是能够标识产品P特征的关键词。
相应地,通过对提取到的关键词进行语义分析,便可以根据语义信息确定适合所述业务人员推广的待推广产品具备的特征信息,比如确定的特征信息包括特征A、特征B和特征C三个特征。
这样,只需将上述特征A、特征B和特征C与企业当前推出的各个待推广产品的产品信息进行匹配,按照上述匹配规则,匹配出一个满足要求的产品,即可完成确定适合所述业务人员推广的所述待推广产品的操作。
应当理解的是,以上给出的仅为一种根据历史从业信息,确定适合所述业务人员推广的待推广产品的具体实现方式,对本申请的技术方案并不构成任何限定,在实际应用中,本领域的技术人员可以根据需要进行设置,此处不做限制。
(2-3)采用预先构建的产品分析模型对所述待推广产品和所述常住地址信息进行分析,确定所述业务人员需要前往的产品推广地址。
具体的说,由于在实际应用中,某些推广活动,需要持续多天,业务人员往往需要从家直接前往产品推广地址,因而在确定适合业务人员进行推广的待推广产品之后,通过根据待推广产品和业务人员的常住地址来确定业务人员需要前往的产品推广地址,不仅可以使得产品推广地址适合所述待推广产品的推广,也方便所述业务人员前往。
此外,关于上述步骤中所说的采用预先构建的产品分析模型对所述待推广地址和所述常住地址信息进行分析,进而确定所述业务人员需要前往的产品推广地址的操作,在实际应用中大致如下:
(2-31)获取所述待推广产品的产品信息,将所述产品信息作为第一输入参数;
(2-32)将所述第一输入参数,输入所述产品分析模型,经所述产品分析模型进行分析处理,得到至少一个第一推广地址;
(2-33)分别将所述常住地址信息对应的坐标与各第一推广地址对应的坐标做差,得到第二输入参数;
(2-34)将所述第二输入参数,输入所述产品分析模型,经所述产品分析模型进行分析处理,确定所述业务人员需要前往的产品推广地址。
应当理解的是,最终确定所述业务人员需要前往的产品推广地址是上述得到的至少一个第一推广地址中与所述常住地址信息的坐标差值小于预设阈值的一个。
此外,上述步骤(2-32)中所说的通过将所述第一输入参数输入所述产品分析模型进行分析处理,获得至少一个第一推广地址的过程,在实际应用中,所述产品分析模型在进行分析处理时,内部执行的操作可以是先根据所述第一输入参数(即待推广产品的产品信息)确定所述待推广产品的适用人群,然后根据所述适用人群,确定所述适用人群的聚集地址,从而将所述适用人群的聚集地址确定为第一推广地址。
此外,为了便于理解所述产品分析模型的,本实施例中给出一种构建所述产品分析模型的具体实现方式,大致如下所述:
(a)接收数据采集指令,从所述数据采集指令中提取待采集的样本数据(与企业推出的待推广产品类似的产品的产品信息以及参与所述产品推广的业务人员的常住地址信息)的网络地址,比如某一大数据平台的网络访问地址,或者企业推出的待推广产品的产品信息存储访问地址。
(b)根据所述网络地址对网络爬虫进行配置,利用所述网络爬虫从所述网络地址对应的网页中获取所述样本数据。
(c)对所述样本数据进行划分,得到训练数据和测试数据。
应当理解的是,由于上述确定所述业务人员需要前往的产品推广地址的时候,是利用构建的产品分析模型对确定的时候业务人员的待推广产品和所述业务人员的常住地址信息进行分析得到的,因而划分得到的训练数据和测试数据中均需包含企业推出的待推广产品类似的产品的产品信息以及参与所述产品推广的业务人员的常住地址信息。
此外,值得一提的是,在实际应用中,为了保证样本数据的有效性、完整性,在进行上述操作之前,还可以根据预设处理规则,对所述样本数据进行数据清洗操作,比如数据去重、格式转换等,此处不再一一列举,对此也不做任何限制。
(d)分别对所述训练数据和测试数据进行标记,并根据标记后的训练数据确定输入训练数据和输出训练数据,以及输入测试数据和输出测试数据。
(e)采用机器学习算法,比如卷积神经网络算法,构建训练模型。
(f)采用所述输入训练数据和所述输出训练数据对所述训练模型进行训练。
具体的说,此处所说的采用所述输入训练数据和所述输出训练数据对所述训练模型进行训练,具体是将所述输入训练数据作为输入参数,输入所述训练模型的输入层,然后经所述训练模型中的隐藏层、卷积层、池化层进行数据处理,并由所述训练模型的输出层将处理结果输出,最终将输出的处理结果与所述训练输出结果进行对比。
相应地,若通过对比,处理结果和所述输出训练数据匹配,或者匹配度满足预设阈值,则可以认为对所述训练模型的训练完成,得到初始产品分析模型。
此外,应当理解的是,由于在使用最终构建的产品分析模型对所述待推广产品和所述常住地址信息进行分析,确定所述业务人员需要前往的产品推广地址的过程中,会将待推广产品的产品信息作为第一输入参数,将根据第一输入参数分析确定的第一推广地址的坐标与常住地址信息的坐标的差值作为第二输入参数,进而根据第二输入参数,得到最终的处理结果。
因而,在采用所述输入训练数据和所述输出训练数据对所述训练模型进行训练时,具体是先将输入训练数据中的产品信息输入到所述训练模型中,将输出的第一推广地址信息与所述输出训练数据中所述产品信息对应的推广地址信息对比。
相应地,若匹配,则将所述第一推广地址信息对应的坐标与所述输入训练数据中的常住地址信息坐标做差,并将所述差值再次输入所述训练模型,获得第二推广地址信息,并将第二推广地址信息与所述输出训练数据中对应的推广地址信息对比。
(g)采用所述输入测试数据和所述输出测试数据对所述初始产品分析模型进行测试,评估所述初始产品分析模型是否具备泛化性。
应当理解的是,采用所述输入测试数据和所述输出测试数据对所述初始产品分析模型的测试过程,与上述采用所述输入训练数据和所述输出训练数据对所述训练模型进行训练的过程大致相似,此处不再赘述。
相应地,若通过将所述输入测试数据输入所述初始产品分析模型后得到的测试结果与所述测试输出数据匹配,或者匹配度满足预设阈值,则可以认为所述初始产品分析模型可以用于后续确定所述业务人员需要前往的产品推广地址,即所述初始产品分析模型便可以作为上述所说的预先构建的产品分析模型。
应当理解的是,以上给出的仅为一种构建所述产品分析模型的具体实现方式,对本申请的技术方案并不构成任何限定,在实际应用中,本领域的技术人员可以根据需要进行设置,此处不做限制。
(3)判断所述业务人员当前所处的地址与所述产品推广地址是否匹配。
相应地,若通过判断确定所述业务人员当前所处的地址与所述产品推广地址匹配,则执行所述登录操作。
也就是说,所述登录操作需要在所述第一生物特征信息与所述第二生物特征信息匹配,且所述业务人员当前所处的地址与所述产品推广地址匹配时,才执行。
应当理解的是,以上给出的仅为一种具体的实现方式,对本申请的技术方案并不构成任何限定,在具体应用中,本领域的技术人员可以根据需要进行设置,本申请对此不做限制。
通过上述描述不难发现,本实施例中提供的基于活体检测的产品推广方法,在接收到业务人员触发的登录产品推广应用程序的登录请求时,通过对所述业务人员进行活体检测,获得所述业务人员的第一生物特征信息;同时,从所述登录请求中提取登录所述产品推广应用程序的用户账号,并根据所述用户账号从预存的业务人员信息库中查找到所述用户账号的真实持有者的第二生物特征信息;最后,通过将所述第一生物特征信息与所述第二生物特征信息进行特征对比,在所述第一生物特征信息与所述第二生物特征信息匹配时,才登录所述产品推广应用程序,并在所述产品推广应用程序的操作界面显示适合所述业务人推广的待推广产品的产品信息,以使所述业务人员根据所述产品信息向体验者推广所述待推广产品。通过这种产品推广方式,保证了分配给业务人员的待推广产品为适合所述业务人员的,从而可以使得业务人员能够在产品推广过程中更好的为体验者进行产品介绍,提升业务推广效率。
此外,由于产品的推广能够精确定位到实际参与人员,因而可以在整个产品推广过程中,更好的对业务人员进行全方位监控,从而更好的督促业务人员积极参与产品推广,进而提升产品推广效果。
参考图3,图3为本申请一种基于活体检测的产品推广方法第二实施例的流程示意图。
基于上述第一实施例,本实施例基于活体检测的产品推广方法在所述步骤S30之后,还包括:
步骤S40,采集所述体验者的面部表情以及所述业务人员在描述待推广产品过程中的音频信息和视频信息。
具体的说,关于所述体验者的面部表情的采集方式,在实际应用中,可以参照上述第一实施例中给出的微表情特征信息的采集方式,即需要在所述业务人员描述待推广产品的过程中,先录制包含所述体验者人脸图像的视频信息;然后,基于预设的人脸检测模型,从所述视频信息中提取出所述体验者在某一时刻的连续人脸图像,并按照时间顺序为各人脸图像进行排序;接着,根据微表情识别技术,按序对各人脸图像进行分析处理,得到所述体验者的微表情特征信息;最后,将整个产品介绍过程中体验者的微表情特征信息作为所述体验者的面部表情变化。
此外,应当理解的是,上述采集到的音频信具体为所述业务人员讲解所述待推广产品的音频信息。
相应地,所述视频信息具体为所述业务人员讲解所述待推广产品时、录制的包含所述业务人员人脸图像、肢体动作等内容的视频信息。
步骤S50,将所述音频信息、所述视频信息和所述体验者的面部表情发送至服务器,以使所述服务器根据所述音频信息、所述视频信息和所述体验者的面部表情确定所述待推广产品的推广效果,并根据所述推广效果调整推广方案。
应当理解的是,所述服务器实质为推出所述待推广产品的企业的服务器。在实际应用中,所述服务器可以是占用实际物理空间的传统服务器,也可以是虚拟云服务器。
相应地,所述服务器在接收到所述音频信息、所述视频信息和所述体验者的面部表情后,具体可以采用大数据分析技术对所述信息进行分析处理,进而确定所述待推广产品的推广效果。
比如,在标识眼内角的上眼皮的特征点向下运动,导致眼内角的上眼皮降低,标识嘴巴的特征点向外运动,导致出嘴巴张大时,通常可以认为所述体验者的面部表情为困倦;还比如,在标识上唇的特征点向上运动,标识下唇的特征点跟谁上唇的特征点向上运动,导致上唇抬起,且下唇与上唇紧闭,嘴角下端,唇轻微凸起;标识眉毛内角的特征点向眉心运动,导致眉毛内角皱在一起,且眉毛抬高是,通常认为所述体验者的面部表情为疑惑;还比如,在标识唇角的特征点向后脸颊后上方运动,导致唇角向后拉并抬高,标识嘴巴的特征点向外运动,导致出嘴巴张大时,通常可以认为所述体验者的面部表情为满意。
相应地,在确定所述体验者的面部表情之后,在根据所述音频信息、所述视频信息和所述体验者的面部表情确定所述待推广产品的推广效果的时候,具体可以将每一个时间段所述体验者的面部表情与所述业务人员描述所述待推广产品的音频信息和视频信息结合,从而确定所述体验者对所述业务人员讲解的哪些内容比较满意,哪些内容不太满意,进而确定所述待推广产品的推广效果。
相应地,在根据所述推广效果调整推广方案时,具体可以是将所述体验者对所述业务人员讲解的音频信息不满意的地方,通过分析所述音频信息中的内容、语气,以及所述视频信息中所述业务人员的服务态度进而确定具体的调整方案。
比如,通过对分析识别所述视频信息中所述业务人员的微表情,所述音频信息中所述业务人员讲解的内容,可以确定所述业务人员服务是否到位,如果大部分业务人员都表现出相同的问题,则可以根据存在的问题,制定专门的培训计划,对存在相关问题的业务人员进行培训。
此外,值得一提的是,在实际应用中,为了督促所述业务人员更好的进行产品的推广,还可以将所述业务人员的音频信息、视频信息,以及体验者对其讲解内容的满意情况进行记录,以便企业对所述业务人员进行后续绩效考核。
通过上述描述不难发现,本实施例中提供的基于活体检测的产品推广方法,在确定所述第一生物特征信息与所述第二生物特征信息匹配,执行登录所述产品推广应用程序的操作之后,通过采集体验者的面部表情以及所述业务人员在描述待推广产品过程中的音频信息和视频信息,并将采集到的所述音频信息、所述视频信息和所述体验者的面部表情发送至服务器,从而可以使所述服务器根据所述音频信息、所述视频信息和所述体验者的面部表情确定所述待推广产品的推广效果,并根据所述推广效果调整推广方案,使得产品的推广方案能够更好的贴近用户需求,进一步提升推广效果。
此外,本申请实施例还提供一种计算机可读存储介质,所述计算机可读存储介质可以为非易失性可读存储介质。本申请计算机可读存储介质上存储有计算机可读指令,其中所述计算机可读指令被处理器执行时,实现如上述的基于活体检测的产品推广方法的步骤。其中,该计算机可读指令被执行时所实现的方法可参照本申请基于活体检测的产品推广方法的各个实施例,此处不再赘述。
参照图4,图4为本申请基于活体检测的产品推广装置第一实施例的结构框图。
如图4所示,本申请实施例提出的基于活体检测的产品推广装置包括:
接收模块4001,用于接收业务人员触发的登录产品推广应用程序的登录请求;提取模块4002,用于从所述登录请求中提取登录所述产品推广应用程序的用户账号;检测模块4003,用于根据所述登录请求对所述业务人员进行活体检测,得到所述业务人员的第一生物特征信息;查找模块4004,用于根据所述用户账号,从预存的业务人员信息库中查找所述用户账号的真实持有者的第二生物特征信息;登录模块4005,用于将所述第一生物特征信息与所述第二生物特征信息进行特征对比,若所述第一生物特征信息与所述第二生物特征信息匹配,则执行登录操作,并在所述产品推广应用程序的操作界面显示待推广产品的产品信息,以使所述业务人员根据所述产品信息向体验者推广所述待推广产品。
其中,上述基于活体检测的产品推广装置的各虚拟功能模块存储于图1所示基于活体检测的产品推广设备的存储器1005中,用于实现计算机可读指令的所有功能;各模块被处理器1001执行时,可实现获取业务人员第一、第二生物特征信息,并根据获取的生物特征信息进行比对确定待推广产品等操作,完成基于活体检测的产品推广的全流程的功能。
进一步地,在所述第一生物特征信息为人脸特征信息时,所述检测模块包括:第一指令生成单元,用于根据所述登录请求,生成采集人脸特征信息的活体检测指令;第一采集单元,用于根据所述活体检测指令,采集所述业务人员读取随机预设内容时的视频信息,所述视频信息包含所述业务人员的人脸图像;第一提取单元,用于基于预设的人脸特征信息提取模型,对所述视频信息中的人脸图像进行人脸特征提取,得到所述业务人员的人脸特征信息。
进一步地,在所述第一生物特征信息为声纹特征信息时,所述检测模块还包括:第二指令生成单元,用于根据所述登录请求,生成采集声纹特征信息的活体检测指令;第二采集单元,用于根据所述活体检测指令,采集所述业务人员读取随机预设内容时的音频信息;第二提取单元,用于基于预设的声纹特征信息提取模型,对所述音频信息进行声纹特征提取,得到所述业务人员的声纹特征信息。
进一步地,在所述第一生物特征信息为微表情特征信息时,所述检测模块还包括:第三指令生成单元,用于根据所述登录请求,生成采集微表情特征信息的活体检测指令;第三采集单元,用于根据所述活体检测指令,采集所述业务人员读取随机预设内容时的视频信息,所述视频信息包含所述业务人员的人脸图像;排序单元,用于基于预设的人脸检测模型,从所述视频信息中提取出所述业务人员在某一时刻的连续人脸图像,并按照时间顺序为各人脸图像进行排序;分析单元,用于根据微表情识别技术,按序对各人脸图像进行分析处理,得到所述业务人员的微表情特征信息。
进一步地,所述基于活体检测的产品推广装置,还包括:第一确定模块,用于获取本机当前所处的位置信息,并根据所述位置信息确定所述业务人员当前所处的地址;第二确定模块,用于根据所述用户账号,确定所述业务人员需要前往的产品推广地址;判断模块,用于判断所述业务人员当前所处的地址与所述产品推广地址是否匹配。
相应地,所述登录模块,还用于在所述第一生物特征信息与所述第二生物特征信息匹配,且所述业务人员当前所处的地址与所述产品推广地址匹配时,执行所述登录操作。
进一步地,所述第二确定模块,包括:查找子单元,用于根据所述用户账号,从所述业务人员信息库中查找所述业务人员的历史从业信息和常住地址信息;待推广产品确定子单元,用于根据所述历史从业信息,确定适合所述业务人员推广的所述待推广产品;产品推广地址确定子单元,用于采用预先构建的产品分析模型对所述待推广产品和所述常住地址信息进行分析,确定所述业务人员需要前往的产品推广地址。
进一步地,所述基于活体检测的产品推广装置,还包括:采集单元,用于采集所述体验者的面部表情以及所述业务人员在描述待推广产品过程中的音频信息和视频信息;发送单元,用于将所述音频信息、所述视频信息和所述体验者的面部表情发送至服务器,以使所述服务器根据所述音频信息、所述视频信息和所述体验者的面部表情确定所述待推广产品的推广效果,并根据所述推广效果调整推广方案。
其中,上述基于活体检测的产品推广装置中各个模块的功能实现与上述基于活体检测的产品推广方法实施例中各步骤相对应,其功能和实现过程在此处不再一一赘述。
此外,需要说明的是,在本文中,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、物品或者系统不仅包括那些要素,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、物品或者系统所固有的要素。在没有更多限制的情况下,由语句“包括一个……”限定的要素,并不排除在包括该要素的过程、方法、物品或者系统中还存在另外的相同要素。上述本申请实施例序号仅仅为了描述,不代表实施例的优劣。通过以上的实施方式的描述,本领域的技术人员可以清楚地了解到上述 实施例方法可借助软件加必需的通用硬件平台的方式来实现,当然也可以通 过硬件,但很多情况下前者是更佳的实施方式。基于这样的理解,本申请的 技术方案本质上或者说对现有技术做出贡献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(如只读存储器(Read Only Memory,ROM)/RAM、磁碟、光 盘)中,包括若干指令用以使得一台终端设备(可以是手机,计算机,服务器,或者网络设备等)执行本申请各个实施例所述的方法。
以上仅为本申请的优选实施例,并非因此限制本申请的专利范围,凡是利用本申请说明书及附图内容所作的等效结构或等效流程变换,或直接或间接运用在其他相关的技术领域,均同理包括在本申请的专利保护范围内。

Claims (20)

  1. 一种基于活体检测的产品推广方法,其中,所述方法包括以下步骤:
    接收业务人员触发的登录产品推广应用程序的登录请求,从所述登录请求中提取登录所述产品推广应用程序的用户账号,并根据所述登录请求对所述业务人员进行活体检测,得到所述业务人员的第一生物特征信息;
    根据所述用户账号,从预存的业务人员信息库中查找所述用户账号的真实持有者的第二生物特征信息;
    将所述第一生物特征信息与所述第二生物特征信息进行特征对比,若所述第一生物特征信息与所述第二生物特征信息匹配,则执行登录操作,并在所述产品推广应用程序的操作界面显示适合所述业务人员推广的待推广产品的产品信息,以使所述业务人员根据所述产品信息向体验者推广所述待推广产品。
  2. 如权利要求1所述的方法,其中,所述第一生物特征信息为人脸特征信息;所述根据所述登录请求对所述业务人员进行活体检测,得到所述业务人员的第一生物特征信息的步骤,包括:
    根据所述登录请求,生成采集人脸特征信息的活体检测指令;
    根据所述活体检测指令,采集所述业务人员读取随机预设内容时的视频信息,所述视频信息包含所述业务人员的人脸图像;
    基于预设的人脸特征信息提取模型,对所述视频信息中的人脸图像进行人脸特征提取,得到所述业务人员的人脸特征信息。
  3. 如权利要求1所述的方法,其中,所述第一生物特征信息为声纹特征信息;所述根据所述登录请求对所述业务人员进行活体检测,得到所述业务人员的第一生物特征信息的步骤,包括:
    根据所述登录请求,生成采集声纹特征信息的活体检测指令;
    根据所述活体检测指令,采集所述业务人员读取随机预设内容时的音频信息;
    基于预设的声纹特征信息提取模型,对所述音频信息进行声纹特征提取,得到所述业务人员的声纹特征信息。
  4. 如权利要求1所述的方法,其中,所述第一生物特征信息为微表情特征信息;所述根据所述登录请求对所述业务人员进行活体检测,得到所述业务人员的第一生物特征信息的步骤,包括:
    根据所述登录请求,生成采集微表情特征信息的活体检测指令;
    根据所述活体检测指令,采集所述业务人员读取随机预设内容时的视频信息,所述视频信息包含所述业务人员的人脸图像;
    基于预设的人脸检测模型,从所述视频信息中提取出所述业务人员在某一时刻的连续人脸图像,并按照时间顺序为各人脸图像进行排序;
    根据微表情识别技术,按序对各人脸图像进行分析处理,得到所述业务人员的微表情特征信息。
  5. 如权利要求1所述的方法,其中,所述若所述第一生物特征信息与所述第二生物特征信息匹配,则执行登录操作的步骤之前,所述方法还包括以下步骤:
    获取本机当前所处的位置信息,并根据所述位置信息确定所述业务人员当前所处的地址;
    根据所述用户账号,确定所述业务人员需要前往的产品推广地址;
    判断所述业务人员当前所处的地址与所述产品推广地址是否匹配;
    其中,所述若所述第一生物特征信息与所述第二生物特征信息匹配,则执行登录操作的步骤,包括:
    若所述第一生物特征信息与所述第二生物特征信息匹配,且所述业务人员当前所处的地址与所述产品推广地址匹配,则执行所述登录操作。
  6. 如权利要求5所述的方法,其中,所述根据所述用户账号,确定所述业务人员需要前往的产品推广地址的步骤,包括:
    根据所述用户账号,从所述业务人员信息库中查找所述业务人员的历史从业信息和常住地址信息;
    根据所述历史从业信息,确定适合所述业务人员推广的所述待推广产品;
    采用预先构建的产品分析模型对所述待推广产品和所述常住地址信息进行分析,确定所述业务人员需要前往的产品推广地址。
  7. 如权利要求1所述的方法,其中,所述在所述产品推广应用程序的操作界面显示待推广产品的产品信息,以使所述业务人员根据所述产品信息向体验者推广所述待推广产品的步骤之后,所述方法还包括以下步骤:
    采集所述体验者的面部表情以及所述业务人员在描述待推广产品过程中的音频信息和视频信息;
    将所述音频信息、所述视频信息和所述体验者的面部表情发送至服务器,以使所述服务器根据所述音频信息、所述视频信息和所述体验者的面部表情确定所述待推广产品的推广效果,并根据所述推广效果调整推广方案。
  8. 一种基于活体检测的产品推广装置,其中,所述装置包括:
    接收模块,用于接收业务人员触发的登录产品推广应用程序的登录请求;
    提取模块,用于从所述登录请求中提取登录所述产品推广应用程序的用户账号;
    检测模块,用于根据所述登录请求对所述业务人员进行活体检测,得到所述业务人员的第一生物特征信息;
    查找模块,用于根据所述用户账号,从预存的业务人员信息库中查找所述用户账号的真实持有者的第二生物特征信息;
    登录模块,用于将所述第一生物特征信息与所述第二生物特征信息进行特征对比,若所述第一生物特征信息与所述第二生物特征信息匹配,则执行登录操作,并在所述产品推广应用程序的操作界面显示适合所述业务人员的待推广产品的产品信息,以使所述业务人员根据所述产品信息向体验者推广所述待推广产品。
  9. 如权利要求8所述的基于活体检测的产品推广装置,其中,所述第一生物特征信息为人脸特征信息;所述检测模块包括:
    第一指令生成单元,用于根据所述登录请求,生成采集人脸特征信息的活体检测指令;
    第一采集单元,用于根据所述活体检测指令,采集所述业务人员读取随机预设内容时的视频信息,所述视频信息包含所述业务人员的人脸图像;
    第一提取单元,用于基于预设的人脸特征信息提取模型,对所述视频信息中的人脸图像进行人脸特征提取,得到所述业务人员的人脸特征信息。
  10. 如权利要求8所述的基于活体检测的产品推广装置,其中,所述第一生物特征信息为声纹特征信息;所述检测模块还包括:
    第二指令生成单元,用于根据所述登录请求,生成采集声纹特征信息的活体检测指令;
    第二采集单元,用于根据所述活体检测指令,采集所述业务人员读取随机预设内容时的音频信息;
    第二提取单元,用于基于预设的声纹特征信息提取模型,对所述音频信息进行声纹特征提取,得到所述业务人员的声纹特征信息。
  11. 如权利要求8所述的基于活体检测的产品推广装置,其中,所述第一生物特征信息为微表情特征信息;所述检测模块还包括:
    第三指令生成单元,用于根据所述登录请求,生成采集微表情特征信息的活体检测指令;
    第三采集单元,用于根据所述活体检测指令,采集所述业务人员读取随机预设内容时的视频信息,所述视频信息包含所述业务人员的人脸图像;
    排序单元,用于基于预设的人脸检测模型,从所述视频信息中提取出所述业务人员在某一时刻的连续人脸图像,并按照时间顺序为各人脸图像进行排序;
    分析单元,用于根据微表情识别技术,按序对各人脸图像进行分析处理,得到所述业务人员的微表情特征信息。
  12. 如权利要求8所述的基于活体检测的产品推广装置,其中,所述基于活体检测的产品推广装置,还包括:
    第一确定模块,用于获取本机当前所处的位置信息,并根据所述位置信息确定所述业务人员当前所处的地址;
    第二确定模块,用于根据所述用户账号,确定所述业务人员需要前往的产品推广地址;
    判断模块,用于判断所述业务人员当前所处的地址与所述产品推广地址是否匹配;
    相应地,所述登录模块,还用于在所述第一生物特征信息与所述第二生物特征信息匹配,且所述业务人员当前所处的地址与所述产品推广地址匹配时,执行所述登录操作。
  13. 一种基于活体检测的产品推广设备,其中,所述设备包括:存储器、处理器及存储在所述存储器上并可被所述处理器执行的计算机可读指令,其中所述计算机可读指令被所述处理器执行时,实现如下步骤:
    接收业务人员触发的登录产品推广应用程序的登录请求,从所述登录请求中提取登录所述产品推广应用程序的用户账号,并根据所述登录请求对所述业务人员进行活体检测,得到所述业务人员的第一生物特征信息;
    根据所述用户账号,从预存的业务人员信息库中查找所述用户账号的真实持有者的第二生物特征信息;
    将所述第一生物特征信息与所述第二生物特征信息进行特征对比,若所述第一生物特征信息与所述第二生物特征信息匹配,则执行登录操作,并在所述产品推广应用程序的操作界面显示适合所述业务人员推广的待推广产品的产品信息,以使所述业务人员根据所述产品信息向体验者推广所述待推广产品。
  14. 如权利要求13所述的基于活体检测的产品推广设备,其中,所述第一生物特征信息为人脸特征信息;所述根据所述登录请求对所述业务人员进行活体检测,得到所述业务人员的第一生物特征信息的步骤,包括:
    根据所述登录请求,生成采集人脸特征信息的活体检测指令;
    根据所述活体检测指令,采集所述业务人员读取随机预设内容时的视频信息,所述视频信息包含所述业务人员的人脸图像;
    基于预设的人脸特征信息提取模型,对所述视频信息中的人脸图像进行人脸特征提取,得到所述业务人员的人脸特征信息。
  15. 如权利要求13所述的基于活体检测的产品推广设备,其中,所述第一生物特征信息为声纹特征信息;所述根据所述登录请求对所述业务人员进行活体检测,得到所述业务人员的第一生物特征信息的步骤,包括:
    根据所述登录请求,生成采集声纹特征信息的活体检测指令;
    根据所述活体检测指令,采集所述业务人员读取随机预设内容时的音频信息;
    基于预设的声纹特征信息提取模型,对所述音频信息进行声纹特征提取,得到所述业务人员的声纹特征信息。
  16. 如权利要求13所述的基于活体检测的产品推广设备,其中,所述第一生物特征信息为微表情特征信息;所述根据所述登录请求对所述业务人员进行活体检测,得到所述业务人员的第一生物特征信息的步骤,包括:
    根据所述登录请求,生成采集微表情特征信息的活体检测指令;
    根据所述活体检测指令,采集所述业务人员读取随机预设内容时的视频信息,所述视频信息包含所述业务人员的人脸图像;
    基于预设的人脸检测模型,从所述视频信息中提取出所述业务人员在某一时刻的连续人脸图像,并按照时间顺序为各人脸图像进行排序;
    根据微表情识别技术,按序对各人脸图像进行分析处理,得到所述业务人员的微表情特征信息。
  17. 如权利要求13所述的基于活体检测的产品推广设备,其中,所述若所述第一生物特征信息与所述第二生物特征信息匹配,则执行登录操作的步骤之前,所述方法还包括以下步骤:
    获取本机当前所处的位置信息,并根据所述位置信息确定所述业务人员当前所处的地址;
    根据所述用户账号,确定所述业务人员需要前往的产品推广地址;
    判断所述业务人员当前所处的地址与所述产品推广地址是否匹配;
    其中,所述若所述第一生物特征信息与所述第二生物特征信息匹配,则执行登录操作的步骤,包括:
    若所述第一生物特征信息与所述第二生物特征信息匹配,且所述业务人员当前所处的地址与所述产品推广地址匹配,则执行所述登录操作。
  18. 一种计算机可读存储介质,其中,所述计算机可读存储介质上存储有计算机可读指令,其中所述计算机可读指令被处理器执行时,实现如下步骤:
    接收业务人员触发的登录产品推广应用程序的登录请求,从所述登录请求中提取登录所述产品推广应用程序的用户账号,并根据所述登录请求对所述业务人员进行活体检测,得到所述业务人员的第一生物特征信息;
    根据所述用户账号,从预存的业务人员信息库中查找所述用户账号的真实持有者的第二生物特征信息;
    将所述第一生物特征信息与所述第二生物特征信息进行特征对比,若所述第一生物特征信息与所述第二生物特征信息匹配,则执行登录操作,并在所述产品推广应用程序的操作界面显示适合所述业务人员推广的待推广产品的产品信息,以使所述业务人员根据所述产品信息向体验者推广所述待推广产品。
  19. 如权利要求18所述的计算机可读存储介质,其中,所述第一生物特征信息为人脸特征信息;所述根据所述登录请求对所述业务人员进行活体检测,得到所述业务人员的第一生物特征信息的步骤,包括:
    根据所述登录请求,生成采集人脸特征信息的活体检测指令;
    根据所述活体检测指令,采集所述业务人员读取随机预设内容时的视频信息,所述视频信息包含所述业务人员的人脸图像;
    基于预设的人脸特征信息提取模型,对所述视频信息中的人脸图像进行人脸特征提取,得到所述业务人员的人脸特征信息。
  20. 如权利要求18所述的计算机可读存储介质,其中,,所述第一生物特征信息为声纹特征信息;所述根据所述登录请求对所述业务人员进行活体检测,得到所述业务人员的第一生物特征信息的步骤,包括:
    根据所述登录请求,生成采集声纹特征信息的活体检测指令;
    根据所述活体检测指令,采集所述业务人员读取随机预设内容时的音频信息;
    基于预设的声纹特征信息提取模型,对所述音频信息进行声纹特征提取,得到所述业务人员的声纹特征信息。
PCT/CN2019/120912 2019-05-21 2019-11-26 基于活体检测的产品推广方法、装置、设备及存储介质 WO2020233055A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910432587.6A CN110264243A (zh) 2019-05-21 2019-05-21 基于活体检测的产品推广方法、装置、设备及存储介质
CN201910432587.6 2019-05-21

Publications (1)

Publication Number Publication Date
WO2020233055A1 true WO2020233055A1 (zh) 2020-11-26

Family

ID=67915232

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/120912 WO2020233055A1 (zh) 2019-05-21 2019-11-26 基于活体检测的产品推广方法、装置、设备及存储介质

Country Status (2)

Country Link
CN (1) CN110264243A (zh)
WO (1) WO2020233055A1 (zh)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110264243A (zh) * 2019-05-21 2019-09-20 深圳壹账通智能科技有限公司 基于活体检测的产品推广方法、装置、设备及存储介质
CN110765434A (zh) * 2019-10-23 2020-02-07 上海商汤智能科技有限公司 身份验证方法、装置、电子设备和存储介质
CN113453531B (zh) * 2021-06-23 2023-05-30 广州佳帆计算机有限公司 供料检测方法及装置
CN113486317A (zh) * 2021-07-06 2021-10-08 中国工商银行股份有限公司 身份验证方法、身份验证装置、电子设备以及存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080255933A1 (en) * 2007-04-16 2008-10-16 Jeffrey Leventhal Method and system for recommending a product over a computer network
CN106411856A (zh) * 2016-09-06 2017-02-15 北京交通大学 基于移动终端人脸识别的认证方法和装置
CN109658126A (zh) * 2018-11-12 2019-04-19 平安科技(深圳)有限公司 基于产品推广的数据处理方法、装置、设备及存储介质
CN110264243A (zh) * 2019-05-21 2019-09-20 深圳壹账通智能科技有限公司 基于活体检测的产品推广方法、装置、设备及存储介质

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106303599B (zh) * 2016-08-11 2021-04-06 腾讯科技(深圳)有限公司 一种信息处理方法、系统及服务器
CN106897067A (zh) * 2017-02-26 2017-06-27 广州衡昊数据科技有限公司 一种基于人机交互技术建模的方法和专家系统
CN107864118B (zh) * 2017-08-14 2020-03-17 深圳壹账通智能科技有限公司 登录验证方法、系统及计算机可读存储介质
CN109712001A (zh) * 2018-11-29 2019-05-03 平安科技(深圳)有限公司 信息推荐方法、装置、计算机设备和存储介质

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080255933A1 (en) * 2007-04-16 2008-10-16 Jeffrey Leventhal Method and system for recommending a product over a computer network
CN106411856A (zh) * 2016-09-06 2017-02-15 北京交通大学 基于移动终端人脸识别的认证方法和装置
CN109658126A (zh) * 2018-11-12 2019-04-19 平安科技(深圳)有限公司 基于产品推广的数据处理方法、装置、设备及存储介质
CN110264243A (zh) * 2019-05-21 2019-09-20 深圳壹账通智能科技有限公司 基于活体检测的产品推广方法、装置、设备及存储介质

Also Published As

Publication number Publication date
CN110264243A (zh) 2019-09-20

Similar Documents

Publication Publication Date Title
WO2020233055A1 (zh) 基于活体检测的产品推广方法、装置、设备及存储介质
WO2020213996A1 (en) Method and apparatus for interrupt detection
WO2020073495A1 (zh) 基于人工智能的复审方法、装置、设备及存储介质
WO2020207035A1 (zh) 骚扰电话拦截方法、装置、设备及存储介质
WO2019201215A1 (zh) 课堂评测方法、装置及计算机可读存储介质
WO2018223520A1 (zh) 面向儿童的学习方法、学习设备及存储介质
WO2020159288A1 (ko) 전자 장치 및 그 제어 방법
WO2015178600A1 (en) Speech recognition method and apparatus using device information
WO2015005679A1 (ko) 음성 인식 방법, 장치 및 시스템
WO2015115681A1 (ko) 표정 동작사전을 이용한 표정인식 방법 및 장치
WO2019125082A1 (en) Device and method for recommending contact information
WO2020253115A1 (zh) 基于语音识别的产品推荐方法、装置、设备和存储介质
EP3545487A1 (en) Electronic apparatus, controlling method of thereof and non-transitory computer readable recording medium
WO2018182201A1 (ko) 사용자의 음성 입력에 대한 답변을 제공하는 방법 및 장치
WO2018101671A1 (en) Apparatus and method for providing sentence based on user input
WO2021112631A1 (en) Device, method, and program for enhancing output content through iterative generation
WO2021010744A1 (ko) 음성 인식 기반의 세일즈 대화 분석 방법 및 장치
EP3652925A1 (en) Device and method for recommending contact information
WO2020207038A1 (zh) 基于人脸识别的人数统计方法、装置、设备及存储介质
WO2022039450A1 (ko) 가상 피팅 서비스 제공 방법, 장치 및 그 시스템
WO2020233076A1 (zh) 基于身份验证的物品入库方法、装置、设备及存储介质
WO2022139155A1 (ko) 컨텐츠 기반 케어 서비스를 제공하는 전자 장치 및 그 제어 방법
WO2021006482A1 (en) Apparatus and method for generating image
WO2019107674A1 (en) Computing apparatus and information input method of the computing apparatus
WO2019190142A1 (en) Method and device for processing image

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: 19929660

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 02/03/2022)

122 Ep: pct application non-entry in european phase

Ref document number: 19929660

Country of ref document: EP

Kind code of ref document: A1