CN109816493A - The virtual accessories recommender system of new media and method based on liquid crystal light modulation film - Google Patents
The virtual accessories recommender system of new media and method based on liquid crystal light modulation film Download PDFInfo
- Publication number
- CN109816493A CN109816493A CN201910086516.5A CN201910086516A CN109816493A CN 109816493 A CN109816493 A CN 109816493A CN 201910086516 A CN201910086516 A CN 201910086516A CN 109816493 A CN109816493 A CN 109816493A
- Authority
- CN
- China
- Prior art keywords
- accessories
- liquid crystal
- light modulation
- crystal light
- user
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Present disclose provides a kind of virtual accessories recommender system of new media based on liquid crystal light modulation film and methods, and the disclosure is based on liquid crystal light modulation film and augmented reality, in conjunction with the virtual accessories recommender system of deep learning algorithm.Relative to the tube advertising of other forms, there is virtual accessories recommender system better interactivity, customization and the system can preferably attract the user's attention power.Meanwhile the application of liquid crystal light modulation membrane technology solves the functionality of subway shield door and as the contradiction between advertising media.And by using the posture information and dress ornament attribute information of depth learning technology identification user, different accessories can be recommended for each user, to reach preferably customization characteristic.Accessories are jumped to by two dimensional code in scanning system and buy interface selection purchase, have reached better interactivity.
Description
Technical field
This disclosure relates to a kind of virtual accessories recommender system of new media based on liquid crystal light modulation film and method.
Background technique
Only there is provided background technical informations relevant to the disclosure for the statement of this part, it is not necessary to so constitute first skill
Art.
Augmented reality, it is a kind of by " seamless " the integrated new technology of real world information and virtual world information,
It is entity information (visual information, sound, the taste that script is difficult to experience in the certain time spatial dimension of real world
Road, tactile etc.), it by science and technology such as computers, is superimposed again after analog simulation, by virtual Information application to real world, quilt
Human sensory is perceived, to reach the sensory experience of exceeding reality.True environment and virtual object are added in real time
Same picture or space exist simultaneously.
As soon as and as the application scenarios under augmented reality, virtual fitting realize user do not have to slough with clothes
It can complete to become a kind of technical application for filling view result.Virtual fitting have the function of it is very powerful, when shopper station trying
When before clothing mirror, virtual fitting can by the realtime graphic of personage in conjunction with 3D garment form after, effect is shown out.Body-sensing is virtual
The advantage of fitting is essentially consisted in when shopper station is before Kinect somatosensory virtual mirror, and fitting mirror will show fitting person automatically
The later fitting image of try-on garment.Using such equipment, consumer no longer needs to take numerously, takes off, changes, wearing each clothes,
The waiting of long period is not needed more, need to only pass through Kinect virtual mirror, so that it may to see oneself true fitting at once
Effect.Meanwhile virtual fitting is also used as the interactive advertisement platform in market, realizes that targetedly advertisement is launched.
In advertisement putting field, rail traffic advertisement occupies very big specific gravity.Especially in recent years, in some flourishing cities
City, such as Shanghai, Beijing, Guangzhou etc., subway has become the main force of urban transportation, and the development space of tier 2 cities subway is then more
For broadness, more wide space is provided to tube advertising.Simultaneously, it has been found that traditional subway new media has following three to lack
It falls into:
1. ad distribution lacks intention, it is difficult to obtain the attention rate of passenger.
2. media content is excessively subjective, lack the interaction with passenger, user is caused not participate well.
3. media content not enough customizes, it can not accomplish the demand for agreeing with each user.
How to solve problem above and the effect that preferably plays will be each advertising company and advertiser is most concerned asks
Topic.It is the new trend of future ads development that system based on augmented reality, which is applied in advertisement field,.It can promote use
The sense of participation at family, while user's feeling of immersion will be greatly promoted, therefore the effect for capableing of advertisement dispensing maximizes.
Existing tube advertising form have 12 bind lamp box, four envelope channel posters, special bit lamp box, staircase, in compartment
Poster etc..But as can most attract the subway shield door of user eyeball but to launch without any type of advertisement, this and subway
The particularity of shield door function has much relations.What it is as one of subway shield door function is exactly the operation that subway is checked for customer
Situation, i.e. subway shield door should keep pellucidity when subway goes out inbound.Therefore, subway shield door is solved as advertising media
It can be used as the breach that new advertisement is launched with the contradiction between its function particularity.
Summary of the invention
The disclosure to solve the above-mentioned problems, proposes a kind of virtual accessories recommendation of the new media based on liquid crystal light modulation film system
System and method, the disclosure are based on liquid crystal light modulation film and augmented reality, recommend system in conjunction with the virtual accessories of deep learning algorithm
System.Relative to the tube advertising of other forms, virtual accessories recommender system has better interactivity, customization and the system
Power can preferably be attracted the user's attention.Meanwhile liquid crystal light modulation membrane technology application solve subway shield door functionality with
As the contradiction between advertising media.And believed by using the posture information and dress ornament attribute of depth learning technology identification user
Breath can recommend different accessories for each user, to reach preferably customization characteristic.It is jumped by two dimensional code in scanning system
The selection purchase of accessories purchase interface is gone to, better interactivity has been reached.
According to some embodiments, the disclosure is adopted the following technical scheme that
A kind of virtual accessories recommender system of new media based on liquid crystal light modulation film, comprising:
Display and interactive unit, are provided with liquid crystal light modulation film, are configured as passing through in primary server under display state
Kinect analyzes the three-dimensional bone information of color image and depth image prediction user, and virtual in the superposition of the bone site of prediction
Accessories；
Dress ornament style recognition unit is configured as the dress ornament style for the user that identification enters in system, while according to acquisition
The face of user identify and the dress ornament style of identification be sent to virtual accessories wearable unit；
Virtual accessories wearable unit is configured as the dress ornament of user's three-dimensional bone site information and user that fusion is extracted
The accessories to be selected that style information, simultaneous selection and user's gender and dress ornament style match are superimposed on user's body；
Accessories choose unit, are configured as buying two dimensional code by scanning accessories and jump to accessories and choose interface, are choosing
Interface provides accessories to be selected；
Liquid crystal light modulation film control unit, by receive subway Central Control Room send subway enter the station exit signal control liquid crystal tune
Light film from display state and primary server operating status.
It is limited as further, the display and interactive unit further include:
Overlapping display unit: for being superimposed virtual accessories with it to the user in color image；
Interactive unit: for realizing the selection of the gesture of user and the virtual accessories of gesture stability is passed through.
The virtual accessories recommended method of subway new media based on liquid crystal light modulation film, comprising the following steps:
Step 1: it builds subway circulation signal and receives primary server, connection primary server and subway Central Control Room, receive subway
Outbound check-in signal；
Step 2: connection subway circulation signal receives primary server and liquid crystal light modulation film, is received by subway circulation signal
Primary server controls liquid crystal light modulation film；
Step 3: the dress ornament style recognizer based on deep learning is deployed on server, while accessories being chosen
Interface is deployed to remote server；
Step 4: liquid crystal light modulation film is deployed on subway shield door, and kinect is arranged under liquid crystal light modulation film
Side；
Step 5: by accessories recommender system client deployment to primary server, pushing away for accessories is managed by primary server
It recommends and overlaying function；
Step 6: the access way of remote server is shown that on system main screen, user chooses corresponding accessories.
It is limited as further, the specific method of the step 1 includes:
(1-1) primary server is constantly in listening state, and load subway signal receives primary server and formed in local area network
One multicast domain；
(1-2) Central Control Room sends the IP address of oneself when the connection with primary server is established in request in the multicast domain
And it requests to connect；
(1-3) primary server is verified after receiving IP address, and Central Control Room is formally established with the connection of primary server；
(1-4) Central Control Room sends the check-in signal that goes out of subway to primary server, and requests to disconnect；
(1-5) primary server receives signal and disconnects, and server returns listening state；
It is limited as further, the step 3 method particularly includes:
Deep learning algorithm based on convolutional neural networks is trained by (3-1) locally, by trained mold portion
Affix one's name to primary server；
Accessories shopping servers are deployed in Tomcat frame by (3-2), and accessories shopping servers are built and are counted in distal end
Calculation machine；
Primary server and the accessories shopping servers of distal end are passed through socket connection by (3-3).
It is limited as further, deep learning algorithm training in the step (3-1) method particularly includes:
(31-1) loads label and attribute set；
(31-2) selects VGG-16 model, uses stochastic gradient descent algorithm and selects Adam optimizer with a fixed step size
It is trained on load label and attribute set；
Trained model is deployed on primary server by (31-3).
It is limited as further, the step 5 method particularly includes:
(5-1) connection Kinect and local client obtain the image and skeleton point of user using bone API；
The Kinect image asynchronous obtained is sent to by (5-2) unlatching sub thread has disposed dress ornament style recognizer
On server and waiting returns the result；
(5-3) recommends handbag according to the dress ornament style of the user identified, and handbag is superimposed to Kinect identification
User key point on.
It is limited as further, the step 6 method particularly includes:
Two-dimensional code generation module is integrated to accessories recommender system client by (6-1)；
Two-dimensional code generation module is linked to the IP address of dress ornament shopping servers by (6-2), passes through two-dimensional code generation module
It is two dimensional code by IP address conversion.
It is limited as further, it is raw that two dimensional code is added in the step (6-1) in accessories recommender system client
At module method particularly includes:
Two-dimension code generator SDK is integrated into accessories recommender system by (61-1)；
Two dimensional code icon UI is added in (61-2) in accessories recommender system, and the two dimensional code that generator generates is shown in UI
In.
Compared with prior art, the disclosure has the beneficial effect that
The disclosure solves the problems, such as that tube advertising attraction is low, and it is existing to joined enhancing in traditional tube advertising
The elements such as reality, interaction, one-button-to-buy, are greatly improved the enjoyment of advertisement, so that the attraction of advertisement significantly improves.Meanwhile
Also the problems such as solving general virtual accessories donning system recommends the specific aim of accessories insufficient, and repetitionization is serious.
The disclosure it is low in cost, relative to traditional tube advertising and virtual accessories donning system, this system is without making
It is interacted with touch screen with user.
It is convenient, easy to operate that the disclosure has many advantages, such as, does not need the additional learning training of user, user only needs by sweeping
Oneself favorite accessories can be selected by retouching the two dimensional code on screen, and can directly be bought in application, be eliminated many numerous
Trivial step.
The disclosure solve subway shield door can not dual-use material contradiction so that subway shield door can do display screen again
Passenger is not blocked to check subway operating condition.
Detailed description of the invention
The accompanying drawings constituting a part of this application is used to provide further understanding of the present application, and the application's shows
Meaning property embodiment and its explanation are not constituted an undue limitation on the present application for explaining the application.
Fig. 1 is that accessories dress interactive exhibition system hardware composite structural diagram.
Fig. 2 is that accessories dress interactive exhibition system overall construction drawing；
Fig. 3 is user's recognition subsystem flow chart；
Fig. 4 is accessories collocation subsystem flow chart；
Fig. 5 is information gathering subsystem structure chart；
Fig. 6 is that virtual accessories dress subsystem flow chart；
Fig. 7 is display and interactive subsystem flow chart；
Fig. 8 is virtual accessories stacked system schematic diagram；
Fig. 9 is that accessories dress interactive exhibition system operational flowchart；
Figure 10 example effect diagram；
Specific embodiment:
The disclosure is described further with embodiment with reference to the accompanying drawing.
It is noted that following detailed description is all illustrative, it is intended to provide further instruction to the application.Unless another
It indicates, all technical and scientific terms used herein has usual with the application person of an ordinary skill in the technical field
The identical meanings of understanding.
It should be noted that term used herein above is merely to describe specific embodiment, and be not intended to restricted root
According to the illustrative embodiments of the application.As used herein, unless the context clearly indicates otherwise, otherwise singular
Also it is intended to include plural form, additionally, it should be understood that, when in the present specification using term "comprising" and/or " packet
Include " when, indicate existing characteristics, step, operation, device, component and/or their combination.
In the disclosure, term for example "upper", "lower", "left", "right", "front", "rear", "vertical", "horizontal", " side ",
The orientation or positional relationship of the instructions such as "bottom" is to be based on the orientation or positional relationship shown in the drawings, only to facilitate describing this public affairs
The relative for opening each component or component structure relationship and determination, not refers in particular to either component or element in the disclosure, cannot understand
For the limitation to the disclosure.
In the disclosure, term such as " affixed ", " connected ", " connection " be shall be understood in a broad sense, and indicate may be a fixed connection,
It is also possible to be integrally connected or is detachably connected；It can be directly connected, it can also be indirectly connected through an intermediary.For
The related scientific research of this field or technical staff can determine the concrete meaning of above-mentioned term in the disclosure as the case may be,
It should not be understood as the limitation to the disclosure.
As shown in Figure 1, covering liquid crystal light modulation film on the glass of subway shield door, Kinect device is fixed thereunder,
And be liquid crystal light modulation film access controller, projector is fixed in the other end appropriate location of subway tunnel and calculating is provided
Server.When subway enters the station, controller is that liquid crystal light modulation film accesses electric current, and liquid crystal light modulation film is converted to transparent state, makes subway
Shield door is transparent；When subway is outbound, controller stops being that liquid crystal light modulation film accesses electric current, and liquid crystal light modulation film restores impermeable again
Bright state；
As shown in Fig. 2, accessories wearing interactive exhibition system is made of five parts, user's recognition subsystem, accessories collocation
Subsystem, information gathering subsystem, virtual accessories purchase subsystem, display and interactive subsystem.Information gathering subsystem will be adopted
The color image information and deep image information collected is sent to other subsystems.User's recognition subsystem passes through the coloured silk received
Color image information identifies age, gender and the dress ornament style of user.Accessories arrange in pairs or groups subsystem according to age, gender and dress ornament wind
Lattice recommend several accessories for virtual wearing, and the accessory information of selection is sent to virtual accessories and buys subsystem.Virtual accessories
Purchase subsystem picks out suitable accessories model according to the gender and dress ornament style of the user received from database, is rendered to
Image is sent to display and interactive subsystem.Display and interactive subsystem are by the accessories received rendering image and collected user
Color image be superimposed to form virtual accessories wearing effect.
As shown in figure 3, user's recognition subsystem has invoked recognition of face and dress ornament identifies two modules.
User's recognition subsystem is acquired color image using kinect and face recognition module is called to be calculated using deep learning
The facial information of method identification user.Face recognition module using Baidu's recognition of face SDK as supporting, and by Baidu's recognition of face
SDK is integrated to local.In remotely creation face database user for searching face while calling recognition of face SDK.If retrieval
Then continue to identify user's face information to user's face information, while analyzing gender and the age of user.Then by gender and year
Age information is sent to accessories and recommends and select subsystem.
User's registration and identification submodule registered face are called if not retrieving user's face information, register perfect person's face
Information calls user's dress ornament Attribute Recognition module.
User dress ornament Attribute Recognition submodule using VGG16 as inference pattern, this is because for the classification of dress ornament
Fine identification needs to extract finer feature, and effect is very in the fine feature for extracting bottom by Inception, Resnet
Well, but both the above model is higher for the performance requirement of machine to be applied in practical situations, therefore this system is selected
VGG16 is selected as inference pattern, the last softmax output layer of VGG16 is changed to 50 kinds of classifications by system, finally again by class in 50
It Ju Lei be not 6 kinds of styles, use Tensorflow as rear end, it is big using the dress ornament identification module inference time of VGG16 model
About 2S.User's dress ornament Attribute Recognition submodule is used as according to the key point that bone identification module identifies cuts foundation first,
By the left shoulder of user, right shoulder, head, image cropping within foot, and the image cut out is sent to backstage dress ornament identification clothes
Business device.50 kinds of dress ornament attributes are divided into six kinds of wind such as sweet maiden's style, sports style by the dress ornament attribute of server analysis user
Lattice.And dress ornament style information is sent to clothing matching subsystem, clothing matching subsystem is automatic according to the dress ornament style of user
Recommend suitable accessories
As shown in figure 4, accessories collocation subsystem mainly includes the processes such as accessories selection and accessories purchase；
Accessories selecting module obtains the gender and age information of user, is divided to gender to (male/female) two values, character classification by age
For values such as (10-20 years old/20-30 years old/30-40 years old/40-50 years old/50 years old or more) and use the information and dress ornament style information
It is retrieved in accessories collocation database, accessories arrange in pairs or groups database selection Mysql as support.It in the database include an accessories
Table.Wherein accessories table includes accessories id, accessories style, the affiliated age bracket of accessories, the affiliated gender field of accessories, and accessories table is with accessories
Id is as major key.Accessories selecting module is examined in accessories collocation database according to gender, age, dress ornament style as search field
The corresponding accessories id of rope, and the accessories of retrieval are sent to accessories selection and purchase module.
The accessories recommended in accessories purchase module are updated according to the accessories of accessories selecting module selection, user buys in accessories
Module completes the replacement and purchase of accessories, and accessories id is sent to virtual accessories and dresses subsystem.
As shown in figure 5, information gathering subsystem includes Color Image Acquisition module, depth image acquisition module.
Color Image Acquisition module obtains the color data stream that Kinect is captured and is sent to display and interactive subsystem, together
When according to whether bone information can be recognized and judge whether there is user enter monitoring range.If user enters within the scope of system monitoring
Color image is then sent to face recognition module, face recognition module is by calling the recognition of face under Baidu AI open platform
Color image is sent to cloud recognition of face server and made inferences by sdk.
Depth image acquisition module is responsible for acquiring the depth information in scene, and Kinect is by calling depth image to acquire mould
The skeleton point position of the prediction user of block and Color Image Acquisition module cooperative.
As shown in fig. 6, virtual accessories wearing subsystem includes identification user's bone, filtration system one skilled in the art, selection accessories
And the processes such as rending model.
Bone identification module merges the color data stream obtained from information gathering subsystem and depth data stream point
Analyse and be calculated 21 skeleton point information of user.Due to subway station, one skilled in the art is numerous, while being superimposed accessories for multiple users
Meeting is so that system interface is all in tumble.Therefore, system can distribute a unique id when identifying user for user and be used to
User is marked, and id is stored in a temporary table, system no longer identifies other use when the id number in list is equal to 3
Family.The maximum number of user amount that system can identify is 3.After recognizing user, then left shoulder, right shoulder, right finesse, hip are therefrom extracted
Then bone site is sent to accessories rendering module by the skeleton points such as bone center.
Accessories model in accessories model database is model of the user in accessories purchase module selection, and user passes through scanning
Two dimensional code jumps to accessories purchase interface in system, and accessories id is inserted into accessories model if user selects favorite accessories
Database.If the accessories id for the recommendation that system is obtained from accessories selecting module is inserted into accessories mould by the non-selected accessories of user
Type database.
Accessories rendering module obtains user's root position, right finesse position and the untiy scene in bone identification module
The position of middle main phase machine, and calculate the angle between root-camera normal and root-right finesse normal.Accessories rendering module root
Accessories superposed positions and the rotation angle of user are obtained according to user's right finesse position and calculated angle, while requesting accessories mould
Type database obtains the accessories model that will be dressed, and is then rendered.And Overlapping display module is sent by rendering result.
As shown in fig. 7, the accessories received are rendered image to Overlapping display module and the color image of collected user is folded
Add virtual accessories wearing effect out (stacked system is as shown in Figure 8).And final effect is pushed on liquid crystal light modulation film and is shown.It hands over
The two dimensional code of accessories purchase module is rendered on liquid crystal film by mutual module, and accessories purchase module choosing is entered after scanning input two dimensional code
It selects and purchase accessories.Whether liquid crystal light modulation film and its control module will enter the station according to subway controls the transparence of liquid crystal light modulation film
State.
As shown in figure 8, the projection pattern of camera is rectangular projection, there is a virtual projection screen immediately ahead of camera.It throws
The color image of display user in shadow curtain.Virtual accessories are inserted between virtual projection curtain and camera.And it is arrived according to user
The position Kinect adjusts the size of virtual accessories.Under the support of the above technology, the illusion of accessories superposition is caused to user.
As shown in figure 9, system running when the moment detection subway whether inbound, if subway does not have inbound, liquid crystal light modulation film
In for opaque state, projector starting waits user to enter system scope；After user enters system scope, system identification
Gender, age and the dress ornament attribute of user recommends accessories, shows virtual accessories wearing effect in subway shield door (liquid crystal light modulation
Film)；If user needs to buy accessories, the two dimensional code using mobile phone scanning accessories purchase module enters accessories purchase interface
Buy accessories；If subway inbound, liquid crystal light modulation film receives check-in signal and screen is switched to pellucidity, and projector is closed.
The virtual accessories recommendation process of subway new media based on liquid crystal light modulation film, comprising the following steps:
Step 1: it builds subway circulation signal and receives primary server, connection server and subway Central Control Room.Subway is received to go out
It stands check-in signal；
Step 2: connection subway circulation signal receives server and liquid crystal light modulation film, is received and is taken by subway circulation signal
Business device controls liquid crystal light modulation film；
Step 3: the dress ornament style recognizer based on deep learning is deployed on server, while accessories being chosen
Interface is deployed to remote server；
Step 4: liquid crystal light modulation film is deployed on subway shield door, and kinect is arranged under liquid crystal light modulation film
Side；
Step 5: by accessories recommender system client deployment to primary server, pushing away for accessories is managed by primary server
It recommends and overlaying function.
Step 6: the access way of distal end accessories shopping servers is shown by way of two dimensional code in system main screen
On.User chooses accessories by scanning the two-dimensional code.
The specific method of step 1 includes:
(1-1) server is constantly in listening state, and load subway signal receives server and forms one in local area network
Multicast domain；
(1-2) Central Control Room sends the IP address of oneself simultaneously in the multicast domain when the connection with server is established in request
Request connection；
(1-3) server is verified after receiving IP address, and Central Control Room is formally established with the connection of server；
(1-4) Central Control Room sends the check-in signal that goes out of subway to server, and requests to disconnect；
(1-5) primary server receives signal and disconnects, and server returns listening state；
Step 3 method particularly includes:
Deep learning algorithm based on convolutional neural networks is trained by (3-1) locally, by trained mold portion
Affix one's name to primary server；
Accessories shopping servers are deployed in Tomcat frame by (3-2), and the server is built in far-end computer；
Primary server and the accessories shopping servers of distal end are passed through socket connection by (3-3)；
Deep learning algorithm training in step (3-1) method particularly includes:
(31-1) loads the label and attribute in the large-scale dataset Deep Fashion opened by Hong Kong Chinese University
Subset (data set includes 50 kinds of different labels)；
(31-2) selects VGG-16 model, uses stochastic gradient descent algorithm and selects Adam optimizer with the step of 1e-4
It grows and is trained on Deep Fashion, BATCH SIZE is set as 32 and enhances data set at random during the training period.Model
Training using Tensorflow as rear end, whole flow process is trained four days on Nvidia1070 video card.The TOP of model
5error reaches 95%.；
Trained model is deployed on primary server by (31-3)；
The step 5 method particularly includes:
(5-1) connection Kinect and local client are obtained using the bone API in Kinect for Windows SDK
Take the image and skeleton point at family；
The Kinect image asynchronous obtained is sent to by (5-2) unlatching sub thread has disposed dress ornament style recognizer
On server and waiting returns the result；
(5-3) recommends handbag according to the dress ornament style of the user identified, and handbag is superimposed to Kinect identification
User key point on；
The step 6 method particularly includes:
Two-dimensional code generation module is integrated to accessories recommender system client by (6-1)；
Two-dimensional code generation module is linked to the IP address of dress ornament shopping servers by (6-2), passes through two-dimensional code generation module
It is two dimensional code by IP address conversion；
Two-dimensional code generation module is added in the step (6-1) in accessories recommender system client method particularly includes:
Zxing two-dimension code generator SDK is integrated into accessories recommender system by (61-1)；
Two dimensional code icon UI is added in (61-2) in accessories recommender system, and the two dimensional code that generator generates is shown in UI
In.
It should be understood by those skilled in the art that, embodiments herein can provide as method, system or computer program
Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the application
Apply the form of example.Moreover, it wherein includes the computer of computer usable program code that the application, which can be used in one or more,
The computer program implemented in usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) produces
The form of product.
The application is referring to method, the process of equipment (system) and computer program product according to the embodiment of the present application
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
The foregoing is merely preferred embodiment of the present application, are not intended to limit this application, for the skill of this field
For art personnel, various changes and changes are possible in this application.Within the spirit and principles of this application, made any to repair
Change, equivalent replacement, improvement etc., should be included within the scope of protection of this application.
Although above-mentioned be described in conjunction with specific embodiment of the attached drawing to the disclosure, model not is protected to the disclosure
The limitation enclosed, those skilled in the art should understand that, on the basis of the technical solution of the disclosure, those skilled in the art are not
Need to make the creative labor the various modifications or changes that can be made still within the protection scope of the disclosure.
Claims (9)
1. the virtual accessories recommender system of a kind of new media based on liquid crystal light modulation film, it is characterized in that: including:
Display and interactive unit, are provided with liquid crystal light modulation film, are configured as passing through in primary server under display state
Kinect analyzes the three-dimensional bone information of color image and depth image prediction user, and virtual in the superposition of the bone site of prediction
Accessories；
Dress ornament style recognition unit is configured as the dress ornament style for the user that identification enters in system, while according to the use of acquisition
The face at family identify and the dress ornament style of identification is sent to virtual accessories wearable unit；
Virtual accessories wearable unit is configured as the dress ornament style of user's three-dimensional bone site information and user that fusion is extracted
The accessories to be selected that information, simultaneous selection and user's gender and dress ornament style match are superimposed on user's body；
Accessories choose unit, are configured as buying two dimensional code by scanning accessories and jump to accessories and choose interface, are choosing interface
Accessories to be selected are provided；
Liquid crystal light modulation film control unit, by receive subway Central Control Room send subway enter the station exit signal control liquid crystal light modulation film
From display state and the operating status of primary server.
2. the virtual accessories recommender system of a kind of new media based on liquid crystal light modulation film as described in claim 1, it is characterized in that: institute
State display and interactive unit further include:
Overlapping display unit: for being superimposed virtual accessories with it to the user in color image；
Interactive unit: for realizing the selection of the gesture of user and the virtual accessories of gesture stability is passed through.
3. the virtual accessories recommended method of subway new media based on liquid crystal light modulation film, it is characterized in that: the following steps are included:
Step 1: it builds subway circulation signal and receives primary server, connection primary server and subway Central Control Room, it is outbound to receive subway
Check-in signal；
Step 2: connection subway circulation signal receives primary server and liquid crystal light modulation film, receives main clothes by subway circulation signal
Business device controls liquid crystal light modulation film；
Step 3: the dress ornament style recognizer based on deep learning is deployed on server, while accessories are chosen into interface
It is deployed to remote server；
Step 4: liquid crystal light modulation film is deployed on subway shield door, and kinect is arranged into the lower section of liquid crystal light modulation film；
Step 5: by accessories recommender system client deployment to primary server, by primary server manage accessories recommendation and
Overlaying function；
Step 6: the access way of remote server is shown that on system main screen, user chooses corresponding accessories.
4. the virtual accessories recommended method of subway new media as claimed in claim 3 based on liquid crystal light modulation film, it is characterized in that: institute
The specific method for stating step 1 includes:
(1-1) primary server is constantly in listening state, and load subway signal receives primary server and forms one in local area network
Multicast domain；
(1-2) Central Control Room sends the IP address of oneself when the connection with primary server is established in request and is asked in the multicast domain
Ask connection；
(1-3) primary server is verified after receiving IP address, and Central Control Room is formally established with the connection of primary server；
(1-4) Central Control Room sends the check-in signal that goes out of subway to primary server, and requests to disconnect；
(1-5) primary server receives signal and disconnects, and server returns listening state；
5. the virtual accessories recommended method of subway new media as claimed in claim 3 based on liquid crystal light modulation film, it is characterized in that: institute
State step 3 method particularly includes:
Deep learning algorithm based on convolutional neural networks is trained by (3-1) locally, and trained model is deployed to
Primary server；
Accessories shopping servers are deployed in Tomcat frame by (3-2), and accessories shopping servers are built and are calculated in distal end
Machine；
Primary server and the accessories shopping servers of distal end are passed through socket connection by (3-3).
6. the virtual accessories recommended method of subway new media as claimed in claim 5 based on liquid crystal light modulation film, it is characterized in that: institute
State deep learning algorithm training in step (3-1) method particularly includes:
(31-1) loads label and attribute set；
(31-2) selects VGG-16 model, uses stochastic gradient descent algorithm and Adam optimizer is selected to add with a fixed step size
It carries and is trained on label and attribute set；
Trained model is deployed on primary server by (31-3).
7. the virtual accessories recommended method of subway new media as claimed in claim 3 based on liquid crystal light modulation film, it is characterized in that: institute
State step 5 method particularly includes:
(5-1) connection Kinect and local client obtain the image and skeleton point of user using bone API；
(5-2) opens sub thread and the Kinect image asynchronous obtained is sent to the service for having disposed dress ornament style recognizer
On device and waiting returns the result；
(5-3) recommends handbag according to the dress ornament style of the user identified, and handbag is superimposed to the use of Kinect identification
In the key point at family.
8. the virtual accessories recommended method of subway new media as claimed in claim 3 based on liquid crystal light modulation film, it is characterized in that: institute
State step 6 method particularly includes:
Two-dimensional code generation module is integrated to accessories recommender system client by (6-1)；
Two-dimensional code generation module is linked to the IP address of dress ornament shopping servers by (6-2), by two-dimensional code generation module by IP
Address conversion is two dimensional code.
9. the virtual accessories recommended method of subway new media as claimed in claim 8 based on liquid crystal light modulation film, it is characterized in that: institute
It states and two-dimensional code generation module is added in step (6-1) in accessories recommender system client method particularly includes:
Two-dimension code generator SDK is integrated into accessories recommender system by (61-1)；
Two dimensional code icon UI is added in (61-2) in accessories recommender system, and the two dimensional code that generator generates is shown in UI.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910086516.5A CN109816493A (en) | 2019-01-29 | 2019-01-29 | The virtual accessories recommender system of new media and method based on liquid crystal light modulation film |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910086516.5A CN109816493A (en) | 2019-01-29 | 2019-01-29 | The virtual accessories recommender system of new media and method based on liquid crystal light modulation film |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109816493A true CN109816493A (en) | 2019-05-28 |
Family
ID=66605738
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910086516.5A Pending CN109816493A (en) | 2019-01-29 | 2019-01-29 | The virtual accessories recommender system of new media and method based on liquid crystal light modulation film |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109816493A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140035913A1 (en) * | 2012-08-03 | 2014-02-06 | Ebay Inc. | Virtual dressing room |
| CN105047108A (en) * | 2015-08-27 | 2015-11-11 | 姜旭东 | Projection imaging device for subway glass shielded gate |
| CN105843386A (en) * | 2016-03-22 | 2016-08-10 | 宁波元鼎电子科技有限公司 | Virtual fitting system in shopping mall |
| US20170053158A1 (en) * | 2015-08-18 | 2017-02-23 | Samsung Electronics Co., Ltd. | Large format display apparatus and control method thereof |
| CN107862277A (en) * | 2017-11-02 | 2018-03-30 | 北京奇虎科技有限公司 | Live dress ornament, which is dressed up, recommends method, apparatus, computing device and storage medium |
| CN108256975A (en) * | 2018-01-23 | 2018-07-06 | 喻强 | Wearing for 3-D effect is provided for virtual fitting person take system and method based on artificial intelligence |
| CN207718772U (en) * | 2018-01-02 | 2018-08-10 | 珠海兴业新材料科技有限公司 | Advertisement delivery system suitable for subway electronic-control dimming glass shield door |
-
2019
- 2019-01-29 CN CN201910086516.5A patent/CN109816493A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140035913A1 (en) * | 2012-08-03 | 2014-02-06 | Ebay Inc. | Virtual dressing room |
| US20170053158A1 (en) * | 2015-08-18 | 2017-02-23 | Samsung Electronics Co., Ltd. | Large format display apparatus and control method thereof |
| CN105047108A (en) * | 2015-08-27 | 2015-11-11 | 姜旭东 | Projection imaging device for subway glass shielded gate |
| CN105843386A (en) * | 2016-03-22 | 2016-08-10 | 宁波元鼎电子科技有限公司 | Virtual fitting system in shopping mall |
| CN107862277A (en) * | 2017-11-02 | 2018-03-30 | 北京奇虎科技有限公司 | Live dress ornament, which is dressed up, recommends method, apparatus, computing device and storage medium |
| CN207718772U (en) * | 2018-01-02 | 2018-08-10 | 珠海兴业新材料科技有限公司 | Advertisement delivery system suitable for subway electronic-control dimming glass shield door |
| CN108256975A (en) * | 2018-01-23 | 2018-07-06 | 喻强 | Wearing for 3-D effect is provided for virtual fitting person take system and method based on artificial intelligence |
Non-Patent Citations (3)
| Title |
|---|
| 王锡林: "《计算机安全》", 31 January 1995, 人民邮电出版社 * |
| 矫春龙: ""基于Kinect的3D虚拟试衣系统"", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 高志强等: "《深度学习从入门到实战》", 30 June 2018, 中国铁道出版社 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11790589B1 (en) | System and method for creating avatars or animated sequences using human body features extracted from a still image | |
| CN104461006A (en) | Internet intelligent mirror based on natural user interface | |
| CN204576413U (en) | A kind of internet intelligent mirror based on natural user interface | |
| US20200257121A1 (en) | Information processing method, information processing terminal, and computer-readable non-transitory storage medium storing program | |
| CN103200181B (en) | A kind of network virtual method based on user's real identification | |
| CN1932799A (en) | System and method for simulating real three-dimensional virtual network travel | |
| JP2012181688A (en) | Information processing device, information processing method, information processing system, and program | |
| CN103533449A (en) | Method and system for realizing three-dimensional fitting based on intelligent three-dimensional television | |
| CN108074278A (en) | Video presentation method, device and equipment | |
| CN107688717A (en) | A kind of Design System of Garment and method based on user itself | |
| CN111862866A (en) | Image display method, device, equipment and computer readable storage medium | |
| CN106447786A (en) | Parallel space establishing and sharing system based on virtual reality technologies | |
| CN108648061A (en) | image generating method and device | |
| KR20190094874A (en) | Digital signage system for providing mixed reality content comprising three-dimension object and marker and method thereof | |
| CN205540924U (en) | Intelligence fitting device | |
| CN114779946A (en) | Wisdom exhibition room management system based on VR technique | |
| CN109887099A (en) | A kind of interaction display method that AR is combined with guideboard | |
| CN108958491A (en) | Long-range experiential method and its system | |
| CN112684893A (en) | Information display method and device, electronic equipment and storage medium | |
| CN109816493A (en) | The virtual accessories recommender system of new media and method based on liquid crystal light modulation film | |
| CN108958571B (en) | Three-dimensional session data display method and device, storage medium and computer equipment | |
| Stojchevska et al. | Modelling the impact of the hotel facilities on online hotel review score for city of Skopje | |
| CN113206868A (en) | Multimedia information processing method and system | |
| CN207396966U (en) | Immersion 3D banquet hall devices | |
| KR20200081102A (en) | Monorail smart tourism information system based on VR technology |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190528 |
|
| RJ01 | Rejection of invention patent application after publication |