Light-sensing intelligent recognition display stand and system

Abstract

The present application relates to light-sensing intelligent recognition display platform and system. The method comprises: acquiring live image with human image through camera, collecting ambient light-sensing data through first light-sensing sensor, collecting light-sensing data of multiple exhibit positions through multiple second light-sensing sensors, generating live face image through live image and ambient light-sensing data, determining moving state of exhibit, identifying age of live face image, determining correlation between live face image and moving state of exhibit within same data collection time period, playing advertisement of current moving exhibit when live face image moves exhibit, and playing advertisement according to age and interest database when there is no live face image moving exhibit. The present application identifies face and estimates age through light-sensing data, predicts interest degree of exhibit according to age, moving exhibit and moving frequency, and realizes active playing of exhibit video.

Description

Technical Field

[0001] This invention relates to a light-sensing intelligent recognition display stand and system, belonging to the field of display technology. Background Technology

[0002] The exhibit display stand not only has the function of displaying physical objects, but also the function of playing exhibit videos. It can stimulate the desire to buy through touch, sight and hearing at the same time, and achieve good publicity results. It is used for exhibit display in different occasions.

[0003] Traditional product display stands typically employ two video playback methods: one involves customers clicking the screen to play the video, a passive approach that hinders customers from quickly acquiring information about exhibits of interest; the other involves playing videos sequentially in a pre-set order, a less targeted approach where the playing video may contain exhibits of uninteresting items, thus diminishing the advertising effect. Therefore, predicting which exhibits visitors will be interested in and proactively playing videos of those items has become a crucial research direction for product display stand technology. Furthermore, the high mobility of product display stands, often requiring changes in location and environment, means that historical data may not be relevant to the current environment. Consequently, traditional data collection and sorting methods struggle to achieve highly accurate predictions of exhibits of interest to visitors.

[0004] Furthermore, the continuous development of age estimation technology based on facial recognition has made it possible to accurately estimate the age of individuals under ideal conditions. However, the unpredictable nature of the application scenarios for display booths leads to complex and variable detection environments, affecting the accuracy of age estimation. Moreover, the rapid turnover of personnel at display booths places higher demands on the system's operating speed. Summary of the Invention

[0005] This invention provides a light-sensing intelligent identification display stand and system, aiming to solve at least one of the technical problems existing in the prior art.

[0006] The technical solution of this invention is based on a light-sensing intelligent recognition display stand. The display stand includes a base, a screen, a camera for capturing images of the area in front of the display stand, a first light sensor for sensing ambient brightness, and a second light sensor for detecting the status of the exhibits. Exhibits are placed in designated positions on the base according to a preset schedule. The second light sensor detects the pressure applied to different exhibit positions on the base. The screen is located on the side of the base, the camera is mounted on the screen, and the first light sensor is mounted on either the screen or the base.

[0007] The first aspect of the technical solution of the present invention relates to a data processing method based on a display stand, used for light-sensing intelligent identification of the display stand, wherein the display stand includes a base, a second light sensor for detecting the placement and removal of exhibits, a screen, a camera for capturing faces, and a first light sensor for sensing ambient brightness. In this aspect, the method according to the present invention includes the following steps:

[0008] S100. During the same data acquisition period, the camera acquires on-site images with human figures, the first light sensor acquires ambient light data, and multiple second light sensors acquire light data of multiple exhibit locations.

[0009] S200 generates on-site facial images using on-site images and ambient light sensing data;

[0010] Based on ambient light sensing data, it is determined whether the illumination intensity of the on-site facial image is lower than a set illumination threshold. If so, an illumination enhancement operation is performed on the on-site facial image, which includes:

[0011] S210. Convert the on-site face image from RGB color space to YCbCr color space, extract the brightness component and construct an initial illumination image;

[0012] S220. Based on the initial illumination image, an enhanced illumination image is obtained through gamma correction;

[0013] S230. Based on the enhanced illumination image, the enhanced face image is obtained through calculation:

[0014] S(x,y)=L(x,y)·R(x,y),

[0015] R(x,y)=S(x,y) / L(x,y),

[0016] Where S(x,y) is the input face image; L(x,y) is the illumination image component; R(x,y) is the reflection image component; and (x,y) is any point in the face image.

[0017] S300: Determine the movement status of the exhibit based on the light-sensing data of the exhibit's location; identify the age of the on-site facial images, and determine the correlation between the on-site facial images and the movement status of the exhibit within the same data collection time period;

[0018] S400: When there is a moving exhibit with a live facial image, play an advertisement for the current moving exhibit; when there is no moving exhibit with a live facial image, play an advertisement based on the age and interest database.

[0019] Furthermore, step S300 includes:

[0020] S310. Input the on-site facial images into a convolutional neural network and calculate the output corresponding to each neuron:

[0021]

[0022] In the formula, h represents the number of neurons in the last layer, and l∈R n+1 θ represents the input vector of the last layer. d ∈R n+1 θ d The weights represent the connections between the d-th neuron and the input vector, where d = 1, 2, ..., h;

[0023] S320. Divide the age range into h age subsets A1, A2, A3, ... A h ; Set o d The age of the facial image at the scene belongs to age subset A. d The probability of being identified is calculated to determine the age of recognition.

[0024]

[0025] In the formula, a d Is the age value in A d The average age of all training images, o d It is the output of the d-th neuron.

[0026] Furthermore, it also includes the following steps:

[0027] S500: Based on the age of the facial images obtained from the identification, classify the currently moving exhibits and modify the total number of moves and the ranking of the corresponding products in the interest database.

[0028] Furthermore, step S500 includes:

[0029] S510. Establish an interest database, input the recognition age of the on-site facial images, and input the currently moving exhibits associated with the on-site facial images and their current number of moves;

[0030] S520. Set three interest categories, namely youth products, middle-aged products, and elderly products; set three age groups, namely youth area, middle-aged area, and elderly area; wherein, the youth products, the middle-aged products, and the elderly products correspond to the youth area, the middle-aged area, and the elderly area, respectively.

[0031] S530. Obtain the age range of the on-site facial image based on the identified age, thereby obtaining the interest category of the current mobile exhibit;

[0032] S540. Obtain the total number of moves corresponding to the currently moving exhibit in the interest category, and add the current number of moves to the total number of moves;

[0033] S550. Based on the total number of moves of exhibits in the interest-based numerical control library, sort the exhibits in each interest category.

[0034] Furthermore, step S550 includes:

[0035] S551. Based on the date the on-site facial image was captured, obtain the recording duration of the data; based on the recording duration, divide the data in the interest database into a first dataset and a second dataset;

[0036] S552. Rank the exhibits in the first dataset and the second dataset according to the number of times the exhibits are moved in each interest category, establish a first exhibit table and a second exhibit table, and obtain the first ranking of the exhibits in the first exhibit table and the second ranking of the exhibits in the second exhibit table.

[0037] S553. Calculate the ratio between the first ranking and the second ranking, and arrange the exhibits of each interest category in the original database in descending order according to the size of the ratio.

[0038] Furthermore, step S550 also includes:

[0039] A mapping function is established to map personnel interest levels to the interest categories corresponding to the personnel's age group, which include the exhibits. This function is used to calculate a quantitative measure of personnel's interest in moving exhibits. The calculation formula is as follows:

[0040] S K =segment(Q K )={t 1K ,t 2K ,…t nK}

[0041] g K (Q K )=g(t 1K ,t 2K ,…t nK )

[0042] In the formula, Q K t represents the set of all ages included within the age range to which a person belongs; k The table of exhibits included in the interest categories corresponding to this age group in the original data; g is the mapping function:

[0043]

[0044] Among them, v iKThis is the ratio of the first ranking to the second ranking.

[0045] Furthermore, it also includes the following steps:

[0046] S600: When no face is detected in the live image captured by the camera, a broadcast is played according to the preset settings.

[0047] The second aspect of the present invention relates to a computer-readable storage medium having program instructions stored thereon, which, when executed by a processor, implement the above-described method.

[0048] The third aspect of the technical solution of the present invention relates to a light-sensing intelligent identification display stand, comprising: a computer device, the computer device including the aforementioned computer-readable storage medium.

[0049] Furthermore, the photosensitive sensor is disposed on the screen or on the base; the second photosensitive sensor is disposed on the base; there are two or more bases, and the two bases are detachably connected by a snap-fit ​​mechanism.

[0050] The beneficial effects of this invention are as follows.

[0051] This invention relates to a data processing method based on an exhibition stand. It acquires facial images and identifies age using cameras and ambient light sensing data, determines the movement status of exhibits using light sensing data of exhibit locations, and predicts exhibit interest levels based on age, moved exhibits, and the number of moves, enabling the proactive playback of exhibit videos. Image processing operations based on ambient light sensing data and an improved convolutional neural network algorithm enhance the accuracy of age estimation and accelerate computation, adapting to scenarios with rapidly changing personnel. Interest categories are categorized based on age, and exhibit interest is ranked based on data recording duration. A quantitative measurement function for exhibit interest is introduced, improving the accuracy of predicting exhibits of interest to individuals and reducing the impact of changes in the geographical environment on prediction accuracy. Attached Figure Description

[0052] Figure 1 This is a basic flowchart of a data processing method based on a display stand according to an embodiment of the present invention.

[0053] Figure 2 This is a control flowchart of a data processing method based on a display stand according to an embodiment of the present invention.

[0054] Figure 3 This is a system flowchart for age estimation according to an embodiment of the present invention.

[0055] Figure 4 This is a flowchart illustrating the operation of face recognition and age estimation according to an embodiment of the present invention.

[0056] Figure 5 This is a schematic diagram of the network structure for age estimation according to an embodiment of the present invention.

[0057] Figure 6 This is a schematic diagram illustrating the prediction effect of the data processing method based on the display stand according to an embodiment of the present invention.

[0058] Figure 7 This is an exploded view of the overall structure of the display stand according to an embodiment of the present invention.

[0059] Figure 8 This is a schematic diagram of the overall structure of the display stand according to an embodiment of the present invention.

[0060] Figure 9 This is a structural schematic diagram of the display stand fastener according to an embodiment of the present invention.

[0061] Figure 10 This is a top view of the display stand according to an embodiment of the present invention.

[0062] Figure 11 This is a schematic diagram of the installation structure of the display stand according to an embodiment of the present invention.

[0063] Figure label:

[0064] 100. Display stand; 110. Base; 111. Snap-on hole; 120. Screen; 130. Camera; 140. First light sensor; 150. Second light sensor; 200. Snap-on component; 210. Snap-on groove. Detailed Implementation

[0065] The following will provide a clear and complete description of the concept, specific structure, and technical effects of the present invention in conjunction with the embodiments and accompanying drawings, so as to fully understand the purpose, solution, and effects of the present invention.

[0066] It should be noted that, unless otherwise specified, when a feature is referred to as "fixed" or "connected" to another feature, it can be directly fixed or connected to the other feature, or indirectly fixed or connected to the other feature. The singular forms "a," "described," and "the" used herein are also intended to include the plural forms, unless the context clearly indicates otherwise. Furthermore, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this specification is for the purpose of describing particular embodiments only and not for limiting the invention. The term "and / or" as used herein includes any combination of one or more of the associated listed items.

[0067] It should be understood that although the terms first, second, third, etc., may be used in this disclosure to describe various elements, these elements should not be limited to these terms. These terms are only used to distinguish elements of the same type from one another. For example, a first element may also be referred to as a second element without departing from the scope of this disclosure, and similarly, a second element may also be referred to as a first element. Any and all instances or exemplary language (“e.g.,” “such as,” etc.) provided herein are intended only to better illustrate embodiments of the invention and, unless otherwise required, do not impose a limitation on the scope of the invention.

[0068] Reference Figures 7 to 11 The technical solution of this invention is based on a light-sensing intelligent recognition display stand 100. The display stand 100 includes a base 110, a screen 120 for playing exhibit advertisements, a camera 130 for capturing the scene in front of the base, a first light sensor 140 for sensing ambient brightness, and a second light sensor 150 for detecting the movement of exhibits. The camera, the first light sensor 140, and the second light sensor 150 are connected to a main control board via a wired connection. The main control board plays corresponding exhibit advertisements on the screen according to system instructions.

[0069] Exhibits in pedestal 110 are placed in designated positions according to pre-set parameters. A second light sensor 150 scans and monitors the area, quickly acquiring the brightness of each exhibit's location. When an exhibit is moved, the brightness at its location increases. The second light sensor 150 detects the movement and obtains the information of the currently moved exhibit, feeding this information back to the system. If the same exhibit is moved and placed back multiple times under the same facial image, the number of times the exhibit has been moved is determined based on the changes in the detection data from the second light sensor 150, and this number is also fed back to the system.

[0070] The display stand 100 is equipped with a camera 130 and a first light sensor 140. The camera 130 captures images of the display stand 100, and the first light sensor 140 detects the ambient light intensity of the environment in which the display stand 100 is located, thus obtaining the ambient light intensity of the image. The first light sensor 140 can be mounted on the screen 120 or on the base 110. When a person stands in front of the display stand 100, the light-sensing intelligent recognition system preprocesses the facial image based on the feedback data from the first light sensor 140 to obtain the person's estimated age more quickly and accurately. Then, the light-sensing intelligent recognition system predicts the exhibits that the person is interested in and ranks the person's interest in all exhibits. The system then plays advertisement videos for the exhibits according to their interest ranking on the screen 120.

[0071] The stand 110 is located on the side of the screen 130, and a light-sensing module composed of multiple second light sensors 150 is mounted on the stand. Two or more stands can be installed, connected by a snap-fit ​​mechanism, facilitating installation and disassembly and improving the expandability of the display stand 100. Referring to Figure 9, the snap-fit ​​component 200 has a snap-fit ​​groove 210; correspondingly, referring to Figure 10, the bottom of the stand 110 has a snap-fit ​​hole 111. Referring to Figure 11, one sidewall of the snap-fit ​​groove 210 engages with the snap-fit ​​hole 111 of one stand 110, and the other sidewall engages with the snap-fit ​​hole 111 of the other stand 110, thus completing the connection and installation of the two stands 110.

[0072] Reference Figures 1 to 6 In some embodiments, the data processing method based on the display stand 100 according to the present invention includes at least the following steps:

[0073] S100: During the same data acquisition period, the camera acquires on-site images with human figures, the first light sensor 140 collects ambient light data, and multiple second light sensors 150 collect light data from multiple exhibit locations.

[0074] S200 generates on-site facial images using on-site images and ambient light sensing data;

[0075] Based on ambient light sensor data, determine whether the illumination intensity of the facial image is below a set illuminance threshold. If so, perform illumination enhancement on the facial image. The illumination enhancement operation includes:

[0076] S210. Convert the on-site face image from RGB color space to YCbCr color space, extract the luminance component and construct the initial illumination image;

[0077] S220. Based on the initial illumination image, an enhanced illumination image is obtained through gamma correction;

[0078] S230. Based on the enhanced illumination image, obtain the enhanced face image through calculation:

[0079] S(x,y)=L(x,y)·R(x,y),

[0080] R(x,y)=S(x,y) / L(x,y),

[0081] Where S(x,y) is the input face image; L(x,y) is the illumination image component; R(x,y) is the reflection image component; and (x,y) is any point in the face image.

[0082] S300: Determine the movement status of exhibits based on light-sensing data of exhibit locations; identify the age of people in on-site facial images and determine the correlation between on-site facial images and exhibit movement status within the same data collection time period;

[0083] S400: When there is a moving exhibit with a human face image on site, play the advertisement for the moving exhibit; when there is no moving exhibit with a human face image on site, that is, when there is no moving exhibit for the person in front of the display stand 100, obtain the interest category corresponding to the person's current age in the interest database based on the person's current age, and play the advertisements in order according to the exhibits in the interest category.

[0084] S500: Based on the age of the facial images obtained from the identification, classify the currently moving exhibits and modify the total number of moves and the ranking of the corresponding products in the interest database under that category.

[0085] S600: If no human face is detected in the on-site image acquired by the camera, it is assumed that there are no people in front of the display stand 100, and the broadcast is played according to the pre-set display order of the exhibits in the background.

[0086] Detailed Implementation of Step S200

[0087] Reference Figure 3 and Figure 4 The system acquires live images of the display stand 100 meters away using a camera. Face detection is performed on these images; if a face is detected, its coordinates are recorded, and the face image is cropped from the live image. Preliminary preprocessing of the face image includes face alignment, standardization, and normalization. Because the cropped face may be slightly tilted, affine transformations are used to adjust the facial geometry, bringing the face to a uniform state and completing the face alignment operation. Then, mean removal is used to center the image, ensuring a more uniform data distribution and achieving image standardization. Finally, the face image size is normalized to the same pixel count.

[0088] Illumination intensity affects feature extraction from facial images; in high-light environments, facial image details are clearer. Therefore, further preprocessing of facial images is necessary. First, an illumination threshold is set, and the ambient illumination detected by the first light sensor 140 is acquired. When the ambient illumination is greater than the photo threshold, facial image feature extraction can be performed directly. When the ambient illumination is less than the photo threshold, the facial image is determined to have low illumination, requiring a light enhancement preprocessing operation before feature extraction.

[0089] In the illumination enhancement process, the low-light face image S(x,y) is first converted from the RGB color space to the YCbCr color space, the luminance component is extracted, and an initial illumination image L1(x,y) is constructed. Then, gamma correction is applied to obtain the enhanced illumination image L2(x,y). Finally, the enhanced face image R(x,y) is obtained using the Retinex algorithm, which is as follows:

[0090] S(x,y)=L(x,y)·R(x,y),

[0091] R(x,y)=S(x,y) / L(x,y),

[0092] In the formula, S(x,y) is the input face image; L(x,y) is the illumination image component; E(x,y) is the reflection image component; and (x,y) is any point in the face image.

[0093] Detailed implementation of step S300

[0094] The method in this embodiment of the invention employs a convolutional neural network structure, see [link to relevant documentation]. Figure 5 The network consists of 13 convolutional layers and 3 fully connected layers, and uses the Softmax activation function. The last layer of the network is modified so that each neuron corresponds to an age.

[0095] Divide the age range into h age subsets A1, A2, A3, ... A h If the minimum age contained therein is and the maximum age is , then the number of neurons in the last layer of the network is set to h, l∈R. n+1 Using the input vector representing the last layer (Softmax layer), θ d ∈R n+1 ,d=1,2,…,h represent the weights connecting the d-th neuron to the input vector. Using the Softmax activation function, the network output is:

[0096]

[0097] Thus, the output of the d-th neuron can be obtained:

[0098]

[0099] Set o d The age of the facial image at the scene belongs to age subset A. d The probability of being identified is calculated to determine the age:

[0100]

[0101] In the formula, a dIs the age value in A d The average age of all training images, o d It is the output of the d-th neuron.

[0102] Detailed Implementation of Step S500

[0103] An interest database was established, taking the recognized age of the obtained on-site facial images as input, along with the corresponding moving exhibit and the number of times the exhibit moved. Three interest categories were set: youth exhibits, middle-aged exhibits, and senior exhibits. The interest dataset was used to record the number of times each exhibit moved within each interest category. Correspondingly, age was divided into three age groups: youth (0-25 years), middle-aged (26-45 years), and senior (46-80 years). The number of exhibit moves in the youth, middle-aged, and senior categories was recorded in the corresponding exhibit move counts for the youth, middle-aged, and senior categories, respectively. The same exhibit could exist in different interest categories simultaneously.

[0104] When a person moves an exhibit on the platform, a camera captures their facial image and estimates their age to determine their age group. For example, if the estimated age is 30, the person is considered to be in the middle-aged category. Then, based on data collected by the second light sensor 150, the exhibit moved by the person is identified. The number of moves for that exhibit is then increased within the corresponding interest category for that age group. For instance, if the 30-year-old person moves exhibit A once, the number of moves for exhibit A in the middle-aged category increases from 10 to 11. Based on changes in data from the second light sensor 150, the current number of moves for an exhibit is accumulated. For example, if the person picks up exhibit A, puts it back on the platform, and then picks it up again, the current number of moves for exhibit A in the middle-aged category increases from 10 to 12.

[0105] Furthermore, the interest database is a set S consisting of all ages within each age group. K Based on the interest categories of age groups, obtain the set T of all exhibits contained within each interest category. K When K is 1, S1 represents the set of all ages included in the youth section of the interest database, and T1 represents the set of exhibits in the youth section corresponding to the youth products; when K is 2, S2 represents the set of all ages included in the middle-aged section of the interest database, and T2 represents the set of exhibits in the middle-aged section corresponding to the middle-aged products; when K is 3, S3 represents the set of all ages included in the elderly section of the interest database, and T3 represents the set of exhibits in the elderly section corresponding to the elderly products.

[0106] Set S K Divided into the first set S 1K The second set S 2K Based on the date the face image was acquired, the data recording duration is determined. If the recording duration is less than or equal to 100 days, the face image data is assigned to set S. 1K If not, then divide its face image data into set S. 2K In the first database, records are recent, so exhibits within its interest categories are defined as exhibits that the current customer is actually interested in. In the second database, records are older, so exhibits within its interest categories are defined as exhibits that are not currently of interest to the customer.

[0107] For set S 1K Obtain set T 1K The number of times each exhibit was moved was used to rank the exhibits, forming the first exhibit list L. 1k Get each exhibit t K In set T 1K rank(t) K L 1k For set S 2K Obtain set T 1K The number of times each exhibit was moved was used to rank the exhibits, forming the second exhibit list L. 2k Get each exhibit t K In set T 2K rank(t) K L 2k Then, calculate the ratio of the two rankings:

[0108] v k (t k ) = rank(t K L 1k ) / rank(t K L 2k (2-1),

[0109] Based on the magnitude of the ratio, the exhibit set T K The exhibits t in the table are arranged in descending order to obtain the exhibit list L. k .

[0110] Based on the current scene images acquired by the camera, and the age estimation results of the facial images in the scene images, according to the exhibit list L kThe exhibits are arranged in a specific order, and their corresponding videos are played sequentially. For example, in the "Middle-Aged" section, if exhibits C, D, and A are listed in order, then when camera 130 detects a person aged 30 approaching display stand 100, since this person belongs to the "Middle-Aged" section and is interested in the "Middle-Aged" exhibits, screen 120 will actively play advertisements for the exhibits within the "Middle-Aged" section. The advertisements will be played in the order of exhibits within the "Middle-Aged" section: first, the advertisement for exhibit C; then, the advertisement for exhibit D; and finally, the advertisement for exhibit A. A sorting threshold is set; if an exhibit's order is higher than the threshold, its advertisement will be played; otherwise, it will not be played.

[0111] Furthermore, given any scene image captured by the camera, a facial image and its estimated age are obtained. Information on the number of exhibits the person picked up and moved is obtained through a second light sensor 150. Based on the estimated age, the age group Q of the facial image is determined. K Based on the number of exhibits taken and moved, the age group Q is obtained. K The collection of exhibits included in S K Then define a mapping g, that is:

[0112] S K =segment(Q K )={t 1K ,t 2K ,…t nK}

[0113] g K (Q K )=g(t 1K ,t 2K ,…t nK )

[0114] In the formula, when K is 1, it indicates that the estimated age of the face image belongs to the youth region, and Q1 represents the set of all ages included in the youth region; when K is 2, it indicates that the estimated age of the face image belongs to the middle-aged region, and Q2 represents the set of all ages included in the middle-aged region; when K is 3, it indicates that the estimated age of the face image belongs to the elderly region, and Q3 represents the set of all ages included in the elderly region. k For exhibits L k The exhibits in the text. `g` is a function that maps the degree of interest of people in an exhibit to the corresponding interest category containing that exhibit within that age group. It represents a quantitative measure of the degree of interest of people in taking the exhibit within that age group. The `g` function can take various forms, such as `g...` K (sum), g K (max), g K (avg), etc. Among them, according to the ranking ratio vk(t) in formula (2-1) k ),get:

[0115]

[0116] g K (max)=max(v iK )

[0117]

[0118] Thus, a quantitative measure g(Q) of the interest level of individuals in exhibits across a complete age group Q, consisting of several exhibits, is obtained. An interest threshold is set. If g(Q) is greater than the interest threshold, it is determined that the exhibit being moved by an individual belongs to the exhibits of interest to individuals in that age group, and the number of moves for that exhibit is increased in the corresponding interest category. Conversely, if g(Q) is greater than the interest threshold, it is considered that the exhibit being moved by an individual does not belong to the exhibits of interest to individuals in that age group, and the number of moves for that exhibit in the corresponding interest category does not need to be increased.

[0119] This invention validates a data processing method based on a display stand 100, predicting whether exhibits included in the interest categories corresponding to each age group are exhibits of interest to that age group, and evaluating performance such as prediction accuracy and recall based on actual exhibit movement by users. See Table 1 and... Figure 5 If no prediction is made and advertisements are played for every exhibit that a person moves through, the precision is 0.07, the recall is 1.00, and the F1-metric is 0.10. Compared with other mapping functions, using the mapping function g(avg) achieves a precision of 0.35, a recall of 0.51, and an F1-metric of 0.48 for advertisement playback (exhibit sorting), which is better than the other two mapping functions.

[0120] Table 1 Performance test results of each mapping function

[0121]

[0122] It should be understood that the method steps in the embodiments of the present invention can be implemented or carried out by computer hardware, a combination of hardware and software, or by computer instructions stored in a non-transitory computer-readable storage medium. The method can use standard programming techniques. Each program can be implemented in a high-level procedural or object-oriented programming language to communicate with the computer system. However, if necessary, the program can be implemented in assembly or machine language. In any case, the language can be a compiled or interpreted language. Furthermore, for this purpose, the program can run on a programmed application-specific integrated circuit (ASIC).

[0123] Furthermore, the procedures described herein may be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by the context. The procedures described herein (or variations and / or combinations thereof) may be executed under the control of one or more computer systems configured with executable instructions, and may be implemented by hardware or a combination thereof as code (e.g., executable instructions, one or more computer programs, or one or more applications) that commonly executes on one or more processors. The computer program comprises a plurality of instructions executable by one or more processors.

[0124] Furthermore, the method can be implemented in any suitable type of computing platform, including but not limited to personal computers, minicomputers, mainframes, workstations, networked or distributed computing environments, standalone or integrated computer platforms, or in communication with charged particle tools or other imaging devices. Aspects of the invention can be implemented as machine-readable code stored on a non-transitory storage medium or device, whether removable or integrated into a computing platform, such as a hard disk, optical read and / or write storage medium, RS1M, ROM, etc., such that it is readable by a programmable computer, and when the storage medium or device is read by the computer, it can be used to configure and operate the computer to perform the processes described herein. Furthermore, the machine-readable code, or portions thereof, can be transmitted via wired or wireless networks. The invention described herein includes these and other different types of non-transitory computer-readable storage media when such media comprises instructions or programs that implement the steps described above in conjunction with a microprocessor or other data processor. When programmed according to the methods and techniques described in the invention, the invention may also include the computer itself.

[0125] A computer program can be applied to input data to perform the functions described herein, thereby transforming the input data to generate output data stored in non-volatile memory. The output information can also be applied to one or more output devices, such as a display. In a preferred embodiment of the invention, the transformed data represents physical and tangible objects, including specific visual depictions of physical and tangible objects generated on the display.

[0126] The above description is merely a preferred embodiment of the present invention. The present invention is not limited to the above-described embodiments. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention, as long as they achieve the technical effects of the present invention by the same means, should be included within the scope of protection of the present invention. Within the scope of protection of the present invention, the technical solutions and / or implementation methods can have various modifications and variations.

Claims

1. A data processing method based on a display stand, wherein, The display stand includes a base, a screen, a camera, a first light sensor, and a second light sensor; the base is used to place exhibits; the first light sensor is used to detect ambient light intensity; and the second light sensor is used to detect whether the exhibits are on the base. The method is characterized by comprising the following steps: S100. During the same data acquisition period, the camera acquires on-site images with human figures, the first light sensor acquires ambient light data, and multiple second light sensors acquire light data of multiple exhibit locations. S200 generates on-site facial images using on-site images and ambient light sensing data; Based on ambient light sensing data, it is determined whether the illumination intensity of the on-site facial image is lower than a set illumination threshold. If so, an illumination enhancement operation is performed on the on-site facial image, which includes: S210. Convert the on-site face image from RGB color space to YCbCr color space, extract the brightness component and construct an initial illumination image; S220. Based on the initial illumination image, an enhanced illumination image is obtained through gamma correction; S230. Based on the enhanced illumination image, the enhanced face image is obtained through calculation: ， ， in, The input is a face image; For the illumination image components; For the reflected image components; Any point in a face image; S300: Determine the movement status of the exhibit based on the light-sensing data of the exhibit's location; identify the age of the on-site facial images, and determine the correlation between the on-site facial images and the movement status of the exhibit within the same data collection time period; S400: When there is a moving exhibit with a live facial image, play an advertisement for the current moving exhibit; when there is no moving exhibit with a live facial image, play an advertisement based on the age and interest database.

2. The method according to claim 1, wherein step S300 comprises: S310. Input the on-site facial images into a convolutional neural network and calculate the output corresponding to each neuron: ； In the formula, This indicates the number of neurons in the last layer. This represents the input vector of the last layer. , Indicates the first The weights of each neuron connected to the input vector, ; S320, Divide the age range into Age subset ; set up The age of the facial image at the scene belongs to the age subset. The probability of being identified is calculated to determine the age of recognition. ； In the formula, Is it age value? The average age of all training images It is the first The output of each neuron.

3. The method according to claim 2, further comprising the following step: S500: Based on the age of the facial images obtained from the identification, classify the currently moving exhibits and modify the total number of moves and the ranking of the corresponding products in the interest database.

4. The method according to claim 3, wherein step S500 comprises: S510. Establish an interest database, input the recognition age of the on-site facial images, and input the currently moving exhibits associated with the on-site facial images and their current number of moves; S520. Set three interest categories, namely youth products, middle-aged products, and elderly products; set three age groups, namely youth area, middle-aged area, and elderly area; wherein, the youth products, the middle-aged products, and the elderly products correspond to the youth area, the middle-aged area, and the elderly area, respectively. S530. Obtain the age range of the on-site facial image based on the identified age, thereby obtaining the interest category of the current mobile exhibit; S540. Obtain the total number of moves corresponding to the currently moving exhibit in the interest category, and add the current number of moves to the total number of moves; S550. Based on the total number of moves of exhibits in the interest-based numerical control library, sort the exhibits in each interest category.

5. The method according to claim 4, wherein step S550 comprises: S551. Obtain the recording duration of the data based on the date the on-site facial image was captured; Based on the recorded duration, the data in the interest database is divided into a first dataset and a second dataset; S552. Rank the exhibits in the first dataset and the second dataset according to the number of times the exhibits are moved in each interest category, establish a first exhibit table and a second exhibit table, and obtain the first ranking of the exhibits in the first exhibit table and the second ranking of the exhibits in the second exhibit table. S553. Calculate the ratio between the first ranking and the second ranking, and sort the exhibits of each interest category in the original database in descending order according to the size of the ratio.

6. The method according to claim 5, wherein, Step S550 further includes: A mapping function is established to map personnel interest levels to the interest categories corresponding to the personnel's age group, which include the exhibits. This function is used to calculate a quantitative measure of personnel's interest in moving exhibits. The calculation formula is as follows: ； ； In the formula, t represents the set of all ages included within the age range to which a person belongs; k The table of exhibits included in the interest categories corresponding to this age group in the original data; g is the mapping function: ； in, This is the ratio of the first ranking to the second ranking.

7. The method according to claim 1, further comprising the following step: S600: When no face is detected in the live image captured by the camera, a broadcast is played according to the preset settings.

8. A computer-readable storage medium having stored thereon program instructions that, when executed by a processor, perform the method as described in any one of claims 1 to 7.

9. A product display stand, characterized in that, include: A computer device comprising the computer-readable storage medium of claim 8.

10. The product display stand according to claim 9, characterized in that, A photosensitive sensor is disposed on the screen or on the base; a second photosensitive sensor is disposed on the base; there are two or more bases, and the two bases are detachably connected by a snap-fit ​​mechanism.

Images