CN110135889A - Method, server and the storage medium of intelligent recommendation book list - Google Patents
Method, server and the storage medium of intelligent recommendation book list Download PDFInfo
- Publication number
- CN110135889A CN110135889A CN201910298660.5A CN201910298660A CN110135889A CN 110135889 A CN110135889 A CN 110135889A CN 201910298660 A CN201910298660 A CN 201910298660A CN 110135889 A CN110135889 A CN 110135889A
- Authority
- CN
- China
- Prior art keywords
- facial image
- default label
- verifying
- training
- image sample
- 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
- 238000000034 method Methods 0.000 title claims abstract description 55
- 230000001815 facial effect Effects 0.000 claims abstract description 128
- 238000012549 training Methods 0.000 claims abstract description 72
- 238000013527 convolutional neural network Methods 0.000 claims description 18
- 238000012706 support-vector machine Methods 0.000 claims description 17
- 238000000605 extraction Methods 0.000 claims description 14
- 239000012141 concentrate Substances 0.000 claims description 9
- 239000000284 extract Substances 0.000 claims description 7
- 238000013507 mapping Methods 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 abstract 1
- 238000004891 communication Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 5
- 238000001914 filtration Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 3
- 230000003044 adaptive effect Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000010224 classification analysis Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 230000002452 interceptive effect Effects 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000000611 regression analysis Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 239000009798 Shen-Fu Substances 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000010267 cellular communication Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06K—GRAPHICAL DATA READING; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
- G06K17/00—Methods or arrangements for effecting co-operative working between equipments covered by two or more of main groups G06K1/00 - G06K15/00, e.g. automatic card files incorporating conveying and reading operations
- G06K17/0022—Methods or arrangements for effecting co-operative working between equipments covered by two or more of main groups G06K1/00 - G06K15/00, e.g. automatic card files incorporating conveying and reading operations arrangements or provisious for transferring data to distant stations, e.g. from a sensing device
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0281—Customer communication at a business location, e.g. providing product or service information, consulting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Multimedia (AREA)
- Strategic Management (AREA)
- Human Computer Interaction (AREA)
- Marketing (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Finance (AREA)
- Tourism & Hospitality (AREA)
- Development Economics (AREA)
- Accounting & Taxation (AREA)
- Game Theory and Decision Science (AREA)
- General Engineering & Computer Science (AREA)
- Primary Health Care (AREA)
- Human Resources & Organizations (AREA)
- Entrepreneurship & Innovation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of methods of intelligent recommendation book list, are applied to server, and this method is by obtaining facial image, and after pre-processing to facial image, the feature vector of facial image is extracted according to predetermined characteristic vector pickup algorithm.Later, using the first identification model of training in advance, input feature value identifies the first default label of facial image, obtains second default label of the corresponding presupposed information of each books as each books.The the second default label for being greater than or equal to the first preset quantity of the first preset threshold with the first default label similarity value is matched from the multiple second default labels, is formed and is recommended book list.It is determined according to the match second default label and recommends book list, and generated two dimensional code corresponding with recommendation list and feed back to the user terminal.The present invention can be more humanized and targetedly recommends suitable books to user, improves the experience sense of user.
Description
Technical field
The present invention relates to technical field of face recognition more particularly to a kind of method of intelligent recommendation book list, server and deposit
Storage media.
Background technique
Large-scale library is often equipped with navigation system at present, retrieves Bibliographical Information, navigation system by input bibliography title
The surplus of books will be provided, position, borrow the information such as number recently, user is helped to find books, but can not be more humane
Recommend suitable books to different users in ground.The books in library recommend generally also only simply to recommend popular books or
Classical books, without specific aim, the effect of recommendation is not satisfactory.
Therefore, how more humanized and targetedly user is given to recommend suitable books, have become one urgently
Technical problem to be solved.
Summary of the invention
The main purpose of the present invention is to provide method, server and the storage mediums of a kind of intelligent recommendation book list.
To achieve the above object, the method for a kind of intelligent recommendation book list provided by the invention is applied to server, this method
Include:
Obtaining step: the facial image of user terminal acquisition is obtained, the facial image is pre-processed；
Extraction step: feature is extracted from the pretreated facial image using preset characteristic vector pickup algorithm
Vector；
Identification step: by described eigenvector input training in advance and the first default label (such as age, the property of user
Not and character trait) corresponding first identification model, identify the first default label of the facial image；
Establishment step: second default label (such as each books of the corresponding presupposed information of each books as each books are obtained
History borrows the age distribution of user, gender, borrows duration, character trait)；
Comparison step: it matches from multiple second default labels and is greater than with the described first default label similarity value
Or the second default label of the first preset quantity equal to the first preset threshold；And
Generation step: it is determined according to the second default label of first preset quantity matched and recommends book list, generated
Two dimensional code corresponding with the recommendation list simultaneously feeds back to the user terminal.
Preferably, the training process of corresponding first identification model of the described first default label includes:
The facial image sample of the second preset quantity is obtained, is that each facial image sample distributes one unique first
Default label；
The facial image sample is divided into the first training set according to the first preset ratio and the first verifying collects, described first
Facial image sample size in training set is greater than the facial image sample size that first verifying is concentrated；
According to the predetermined characteristic vector pickup algorithm, extracts first training set and the first verifying is concentrated
The feature vector of every facial image sample；
Facial image sample in first training set is inputted the support vector machines to be trained, every default week
Phase verifies the support vector machines using the first verifying collection, concentrates each facial image using first verifying
The feature vector of sample and corresponding first default label verify the accuracy rate of first identification model；And
When the accuracy rate of verifying is greater than the second preset threshold, terminates training, obtain first identification model.
Preferably, described eigenvector extraction algorithm includes:
Corresponding with the face-image region of the facial image the second of facial image input training in advance is known
Other model identifies the face-image region of the facial image；
It is the elementary area of the second preset quantity by each face-image region cutting, utilizes predetermined calculating
The gradient-norm and gradient direction of each pixel in described image unit is calculated in rule；
The histogram of gradients of each described image unit is established according to the calculated gradient-norm and gradient direction；And
Each histogram of gradients is formed into every people in the position in the face-image region according to correspondence image unit
The feature vector of face image.
Preferably, the second preset kind identification model is convolutional neural networks model, the convolutional neural networks mould
The training process of type is as follows:
The facial image sample of third preset quantity is obtained, is labeled with face-image region in every facial image sample；
The facial image sample is divided into the second training set according to the second preset ratio and the second verifying collects, described second
Image pattern quantity in training set is greater than the image pattern quantity that second verifying is concentrated；
Facial image sample in second training set is inputted the convolutional neural networks model to be trained, every
Predetermined period verifies the convolutional neural networks model using the second verifying collection, is concentrated using second verifying
The accuracy rate of second identification model is verified in each facial image and corresponding face-image region；And
When the accuracy rate of verifying is greater than third predetermined threshold value, terminates training, obtain second identification model.
Preferably, the predetermined computation rule includes:
Wherein, Gx(x, y) and Gy(x, y) respectively indicates the ladder horizontally and vertically at current pixel point (x, y)
Angle value,Gradient-norm and the direction at current pixel point (x, y) are respectively indicated with θ (x, y).
Preferably, before the comparison step, this method is further comprising the steps of:
Establish the mapping relations between the described second default label and book categories.
To achieve the above object, the present invention furthermore provides a kind of server, and the server includes memory and place
Device is managed, is stored with intelligent recommendation book list program on the memory, the intelligent recommendation book list program is executed by the processor
Shi Shixian following steps:
Obtaining step: the facial image of user terminal acquisition is obtained, the facial image is pre-processed；
Extraction step: feature is extracted from the pretreated facial image using preset characteristic vector pickup algorithm
Vector；
Identification step: by described eigenvector input training in advance and the first default label (such as age, the property of user
Not and character trait) corresponding first identification model, identify the first default label of the facial image；
Establishment step: second default label (such as each books of the corresponding presupposed information of each books as each books are obtained
History borrows the age distribution of user, gender, borrows duration, character trait)；
Comparison step: it matches from multiple second default labels and is greater than with the described first default label similarity value
Or the second default label of the first preset quantity equal to the first preset threshold；And
Generation step: it is determined according to the second default label of first preset quantity matched and recommends book list, generated
Two dimensional code corresponding with the recommendation list simultaneously feeds back to the user terminal.
Preferably, the training process of corresponding first identification model of the described first default label includes:
The facial image sample of the second preset quantity is obtained, is that each facial image sample distributes one unique first
Default label；
The facial image sample is divided into the first training set according to the first preset ratio and the first verifying collects, described first
Facial image sample size in training set is greater than the facial image sample size that first verifying is concentrated；
According to the predetermined characteristic vector pickup algorithm, extracts first training set and the first verifying is concentrated
The feature vector of every facial image sample；
Facial image sample in first training set is inputted the support vector machines to be trained, every default week
Phase verifies the support vector machines using the first verifying collection, concentrates each facial image using first verifying
The feature vector of sample and corresponding first default label verify the accuracy rate of first identification model；And
When the accuracy rate of verifying is greater than the second preset threshold, terminates training, obtain first identification model.
Preferably, described eigenvector extraction algorithm includes:
Corresponding with the face-image region of the facial image the second of facial image input training in advance is known
Other model identifies the face-image region of the facial image；
It is the elementary area of the second preset quantity by each face-image region cutting, utilizes predetermined calculating
The gradient-norm and gradient direction of each pixel in described image unit is calculated in rule；
The histogram of gradients of each described image unit is established according to the calculated gradient-norm and gradient direction；And
Each histogram of gradients is formed into every people in the position in the face-image region according to correspondence image unit
The feature vector of face image.
To achieve the above object, the present invention further provides a kind of computer readable storage mediums, described computer-readable
Intelligent recommendation book list program is stored on storage medium, the intelligent recommendation book list program can be held by one or more processor
Row, to realize such as the step of above-mentioned intelligent recommendation book list method.
Method, server and the storage medium of intelligent recommendation book list proposed by the present invention, are identified by face recognition technology
The first of active user default label out, and borrow information using big data methods analysis chart book shop books history and generate each books
The second default label, generated according to the matching result of the first default label of active user and the second default label of each books
Recommend book list, and generates two dimensional code corresponding with recommendation list and show user.Compared to more existing traditional books way of recommendation,
Method provided by the invention can borrow information targetedly to user's recommended book according to user property and history, promote book
The validity and user experience that nationality is recommended.
Detailed description of the invention
Fig. 1 is the applied environment figure of server preferred embodiment of the present invention；
Fig. 2 is the program module schematic diagram of intelligent recommendation book list program preferred embodiment in Fig. 1；
Fig. 3 is the flow diagram of invoice method of inspection preferred embodiment of the present invention.
The realization, the function and the advantages of the object of the present invention will in conjunction with the embodiments, and Shenfu figure is described further.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that described herein, specific examples are only used to explain the present invention, not
For limiting the present invention.Based on the embodiments of the present invention, those of ordinary skill in the art are not before making creative work
Every other embodiment obtained is put, shall fall within the protection scope of the present invention.
It should be noted that the description for being related to " first ", " second " etc. in the present invention is used for description purposes only, and cannot
It is interpreted as its relative importance of indication or suggestion or implicitly indicates the quantity of indicated technical characteristic.Define as a result, " the
One ", the feature of " second " can explicitly or implicitly include at least one of the features.In addition, the skill between each embodiment
Art scheme can be combined with each other, but must be based on can be realized by those of ordinary skill in the art, when technical solution
Will be understood that the combination of this technical solution is not present in conjunction with there is conflicting or cannot achieve when, also not the present invention claims
Protection scope within.
The present invention provides a kind of server 1.Shown in referring to Fig.1, server 1 identifies current use by face recognition technology
The default label of the first of family, and borrow information using big data methods analysis chart book shop books history and generate the second pre- of each books
Bidding label generate recommendation according to the matching result of the first default label of active user and the second default label of each books
It is single, and generate two dimensional code corresponding with recommendation list and show user.Compared to more existing traditional books way of recommendation, the present invention
The method of offer can borrow information targetedly to user's recommended book according to user property and history, promote books and recommend
Validity and user experience.
The server 1 can be rack-mount server, blade server, tower server or Cabinet-type server etc.
One or more.The server is include but are not limited to, memory 11, processor 12 and network interface 13.
Wherein, memory 11 include at least a type of readable storage medium storing program for executing, the readable storage medium storing program for executing include flash memory,
Hard disk, multimedia card, card-type memory (for example, SD or DX memory etc.), magnetic storage, disk, CD etc..Memory 11
It can be the internal storage unit of server, such as the hard disk of the server in some embodiments.Memory 11 is at other
It is also possible to the plug-in type hard disk being equipped on the External memory equipment of server 1, such as the server 1 in embodiment, intelligently deposits
Card storage (Smart Media Card, SMC), secure digital (Secure Digital, SD) card, flash card (Flash Card)
Deng.
Further, memory 11 can also both including server 1 internal storage unit and also including External memory equipment.
Memory 11 can be not only used for the application software and Various types of data that storage is installed on server 1, such as intelligent recommendation book list journey
The code etc. of sequence 10 can be also used for temporarily storing the data that has exported or will export.
Processor 12 can be in some embodiments a central processing unit (Central Processing Unit,
CPU), controller, microcontroller, microprocessor or other data processing chips, the program for being stored in run memory 11
Code or processing data, such as execute intelligent recommendation book list program 10 etc..
Network interface 13 optionally may include standard wireline interface and wireless interface (such as WI-FI interface), be commonly used in
Communication connection is established between the server 1 and other electronic equipments.
User terminal 14 can be camera or other have the function of obtain facial image and can by network 15 and clothes
The terminal installation that business device is communicated.
Network 15 can be internet, cloud network, Wireless Fidelity (Wi-Fi) network, personal net (PAN), local area network (LAN)
And/or Metropolitan Area Network (MAN) (MAN).Various equipment in network environment can be configured as to be connected according to various wired and wireless communication protocols
It is connected to communication network.The example of such wired and wireless communication protocol can include but is not limited at least one of the following:
Transmission control protocol and Internet protocol (TCP/IP), User Datagram Protocol (UDP), hypertext transfer protocol (HTTP), text
Part transport protocol (FTP), ZigBee, EDGE, IEEE 802.11, light fidelity (Li-Fi), 802.16, IEEE 802.11s,
IEEE 802.11g, multi-hop communication, wireless access point (AP), equipment are to equipment communication, cellular communication protocol and/or bluetooth
(BlueTooth) communication protocol or combinations thereof.
Optionally, which can also include user interface, and user interface may include display (Display), input
Unit such as keyboard (Keyboard), optional user interface can also include standard wireline interface and wireless interface.It is optional
Ground, in some embodiments, display can be light-emitting diode display, liquid crystal display, touch-control liquid crystal display and OLED
(Organic Light-Emitting Diode, Organic Light Emitting Diode) touches device etc..Wherein, display is referred to as showing
Display screen or display unit, for showing the information handled in the server 1 and for showing visual user interface.
Fig. 1 illustrates only the server 1 with component 11-15 and intelligent recommendation book list program 10, those skilled in the art
Member may include less or more than illustrating it is understood that structure shown in fig. 1 does not constitute the restriction to server 1
More components perhaps combines certain components or different component layouts.
In one embodiment, it when the intelligent recommendation book list program 10 of Fig. 1 is executed by processor 12, performs the steps of
Obtaining step: the facial image of user terminal acquisition is obtained, the facial image is pre-processed；
Extraction step: feature is extracted from the pretreated facial image using preset characteristic vector pickup algorithm
Vector；
Identification step: by described eigenvector input training in advance and the first default label (such as age, the property of user
Not and character trait) corresponding first identification model, identify the first default label of the facial image；
Establishment step: second default label (such as each books of the corresponding presupposed information of each books as each books are obtained
History borrows the age distribution of user, gender, borrows duration, character trait)；
Comparison step: it matches from multiple second default labels and is greater than with the described first default label similarity value
Or the second default label of the first preset quantity equal to the first preset threshold；And
Generation step: it is determined according to the second default label of first preset quantity matched and recommends book list, generated
Two dimensional code corresponding with the recommendation list simultaneously feeds back to the user terminal.
Further, before the comparison step, this method is further comprising the steps of:
Establish the mapping relations between the described second default label and book categories.
About being discussed in detail for above-mentioned steps, journey of following Fig. 2 about 10 embodiment of intelligent recommendation book list program is please referred to
The explanation of sequence module diagram and Fig. 3 about the method flow schematic diagram of intelligent recommendation book list embodiment of the method.
It is the program module schematic diagram of 10 embodiment of intelligent recommendation book list program in Fig. 1 referring to shown in Fig. 2.Intelligent recommendation
Book list program 10 is divided into multiple modules, and multiple module is stored in memory 11, and is executed by processor 12, to complete
The present invention.The so-called module of the present invention is the series of computation machine program instruction section for referring to complete specific function.
In the present embodiment, the intelligent recommendation book list program 10 includes obtaining module 110, extraction module 120, identification mould
Block 130 establishes module 140, contrast module 150 and generation module 160.
Module 110 is obtained, for obtaining the facial image of user terminal 14 (such as camera) acquisition, and to the face
Image is pre-processed.
The pretreatment of facial image is the influence in order to avoid background to projected image, eliminates ambient noise, to reduce
Below when feature extraction caused by error.Remove the point in input original image in the present embodiment using 3 × 3 median filterings first
Secondly noise obtains edge binary images using adaptive threshold method (LAT), to simplify algorithm and reduce operand.Figure
As the algorithm of edge extracting has very much, there are commonly Laplace operator, Robison operator, Prewitt operator and Sobel calculations
Son, using the Sobel operator in four directions in the present embodiment.It is 3 × 3 templates that Sobel operator uses, wherein matrix below
Middle number represents the gray value of pixel, and range is between 0~255, then the Sobel operator Z0, Z1, Z2 in four directions, Z3 are respectively as follows:
After low-pass filtering, LAT algorithm is as follows:
Here the low-pass filter of g (x, y) representing input images, Zk(x, y) represents 4 outputs of Sobel operator, low pass
Filter is as follows:
At this point it is possible to judge: passing through
If LAT (x, y) > 1, g (x, y) is boundary point, I (x, y)=1；
If LAT (x, y) < 1, g (x, y) is non-boundary point, (x, y)=0；
The binary edge map I (x, y) of image can be obtained in this way.That is the background gray levels of image are 0 (black), and people
Face profile and the gray value for interfering scene border are then 255 (whites).
Extraction module 120, for extracting the feature vector of the facial image, according to predetermined characteristic vector pickup
Algorithm extracts the feature vector of facial image.
In the present embodiment, described eigenvector extraction algorithm includes:
Training the second identification model corresponding with the face-image region of the facial image in advance, the second identification mould
Type is convolutional neural networks model (Convolutional Neural Network, CNN)；
The facial image sample of the 4th preset quantity (such as 100,000) is obtained, is labeled with face in every facial image sample
Image-region；
The facial image sample is divided into the second training set and the second verifying according to the second preset ratio (such as 5:1)
Collect, the facial image sample size in second training set is greater than the facial image sample size that second verifying is concentrated；
Facial image sample in second training set is inputted the convolutional neural networks model to be trained, every
Predetermined period (such as 1000 iteration of every progress) tests the convolutional neural networks model using the second verifying collection
Card concentrates each facial image sample and corresponding face-image region to second identification model using second verifying
Accuracy rate is verified；
When the accuracy rate of verifying is greater than third predetermined threshold value (such as 85%), terminate training, obtains second identification
Model inputs the facial image and inputs the face-image area that second identification model identifies the facial image sample
Domain；And
When the accuracy rate of verifying is less than third predetermined threshold value (such as 85%), then increase the number of the facial image sample
Amount, and above-mentioned steps are re-executed based on increased facial image sample.
By each face-image region cutting be the second preset quantity (such as 16) elementary area, using in advance really
The gradient-norm and gradient direction of each pixel in described image unit is calculated in fixed computation rule；
The histogram of gradients of each described image unit is established according to the calculated gradient-norm and gradient direction；And
Each histogram of gradients is connected according to correspondence image unit in the position in the face-image region, with
Form the feature vector of every facial image sample, such as histograms of oriented gradients (Histogram of Oriented
Gradient, HOG) feature vector.
Wherein, the predetermined computation rule includes:
Wherein, Gx(x, y) and Gy(x, y) respectively indicates the ladder horizontally and vertically at current pixel point (x, y)
Angle value,Gradient-norm and the direction at current pixel point (x, y) are respectively indicated with θ (x, y).
Identification module 130, the first default label corresponding with facial image (includes but are not limited to user's for identification
Age, gender and character trait, such as " age: 10-15 years old "；" gender: female "；" personality: export-oriented ").Utilize training in advance
The first identification model corresponding with the described first default label, input described eigenvector may recognize that and the facial image
Corresponding first default label.
In the present embodiment, first identification model be support vector machines (Support Vector Machine, SVM,
It is a kind of common method of discrimination, is the learning model for having supervision in machine learning field, is commonly used to carry out mode knowledge
Not, classification and regression analysis), the training process of corresponding first identification model of the first default label includes:
The facial image sample of the second preset quantity (such as 100,000) is obtained, is that each facial image sample distributes one
Unique first default label；
The facial image sample is divided into the first training set and the first training according to the first preset ratio (such as 4:1)
Collect, the image pattern quantity in first training set is greater than the image pattern quantity in first training set；
According to the predetermined characteristic vector pickup algorithm, extracts the first training set of institute and the first verifying is concentrated often
The feature vector of facial image sample, such as histograms of oriented gradients (Histogram of Oriented Gradient,
HOG) feature vector；
Image pattern in the training set is inputted the support vector machines to be trained, every predetermined period (such as
1000 iteration of every progress) support vector machines is verified using verifying collection, each face is concentrated using verifying
The feature vector of image pattern and corresponding first default label verify the accuracy rate of first identification model；
When the accuracy rate of verifying is greater than the second preset threshold (such as 95%), terminate training, obtains first identification
Model；And
When the accuracy rate of verifying is less than the second preset threshold (such as 95%), then increase the number of the facial image sample
Amount, and above-mentioned steps are re-executed based on increased facial image sample.
Module 140 is established, for obtaining second default label of the corresponding presupposed information of each books as each books.
In the present embodiment, the presupposed information includes that the history for every books being stored in Library Information System is borrowed
It reads the age distribution of user, gender, borrow duration, character trait (such as export-oriented, introversive).In addition, the second default label
It can also include the build-in attribute of books, such as the type of books, length, the publication time limit.The default label information of the second of books
As the running in library is regular or real-time update.
Further, described to establish module 140 further include: to establish reflecting between the described second default label and book categories
Penetrate relationship.
Different books are sorted out, such as science and technology, life kind etc. are different classes of.User oneself is liked in inquiry
The classification that books can be first selected before books, further reduces range of choice, improves the recommendation precision of system.
Contrast module 150 is greater than for matching from the multiple second default labels with the first default label similarity value
Or it is equal to the second default label of first preset quantity (such as 5) of the first preset threshold (such as 85%).
Generation module 160, for generating two dimensional code corresponding with recommendation list and showing user terminal 14.
It is determined according to the second default label matched and recommends book list, and generated and the recommendation list corresponding two
Tie up code.Wherein, two dimensional code can obtain whole relevant informations of books, author, publication time including books, it is simple introduce,
The academic probation of former chapters, other people evaluation, borrow the frequency recently, can also show surplus, the location information of books, and according to user
Current location generate route map guide user go to the position where books.
In addition, the present invention also provides a kind of methods of intelligent recommendation book list.It is intelligent recommendation of the present invention referring to shown in Fig. 3
The method flow schematic diagram of the embodiment of the method for book list.The processor 12 of server 1 executes the intelligence stored in memory 11
The following steps of the method for intelligent recommendation book list are realized when recommendation one way sequence 10:
Step S110, obtain user terminal 14 (such as camera) acquisition facial image, and to the facial image into
Row pretreatment.
The pretreatment of facial image is the influence in order to avoid background to projected image, eliminates ambient noise, to reduce
Below when feature extraction caused by error.Remove the point in input original image in the present embodiment using 3 × 3 median filterings first
Secondly noise obtains edge binary images using adaptive threshold method (LAT), to simplify algorithm and reduce operand.Figure
As the algorithm of edge extracting has very much, there are commonly Laplace operator, Robison operator, Prewitt operator and Sobel calculations
Son, using the Sobel operator in four directions in the present embodiment.It is 3 × 3 templates that Sobel operator uses, wherein matrix below
Middle number represents the gray value of pixel, and range is between 0~255, then the Sobel operator Z0, Z1, Z2 in four directions, Z3 are respectively as follows:
After low-pass filtering, LAT algorithm is as follows:
Here the low-pass filter of g (x, y) representing input images, Zk(x, y) represents 4 outputs of Sobel operator, low pass
Filter is as follows:
At this point it is possible to judge: passing through
If LAT (x, y) > 1, g (x, y) is boundary point, I (x, y)=1；
If LAT (x, y) < 1, g (x, y) is non-boundary point, (x, y)=0；
The binary edge map I (x, y) of image can be obtained in this way.That is the background gray levels of image are 0 (black), and people
Face profile and the gray value for interfering scene border are then 255 (whites).
Step S120, training the second identification model corresponding with the face-image region of the facial image, described in advance
Second identification model is convolutional neural networks model (Convolutional Neural Network, CNN)；
The facial image sample of the 4th preset quantity (such as 100,000) is obtained, is labeled with face in every facial image sample
Image-region；
The facial image sample is divided into the second training set and the second verifying according to the second preset ratio (such as 5:1)
Collect, the facial image sample size in second training set is greater than the facial image sample size that second verifying is concentrated；
Facial image sample in second training set is inputted the convolutional neural networks model to be trained, every
Predetermined period (such as 1000 iteration of every progress) tests the convolutional neural networks model using the second verifying collection
Card concentrates each facial image sample and corresponding face-image region to second identification model using second verifying
Accuracy rate is verified；
When the accuracy rate of verifying is greater than third predetermined threshold value (such as 85%), terminate training, obtains second identification
Model inputs the facial image and inputs the face-image area that second identification model identifies the facial image sample
Domain；And
When the accuracy rate of verifying is less than third predetermined threshold value (such as 85%), then increase the number of the facial image sample
Amount, and above-mentioned steps are re-executed based on increased facial image sample.
By each face-image region cutting be the second preset quantity (such as 16) elementary area, using in advance really
The gradient-norm and gradient direction of each pixel in described image unit is calculated in fixed computation rule；
The histogram of gradients of each described image unit is established according to the calculated gradient-norm and gradient direction；And
Each histogram of gradients is connected according to correspondence image unit in the position in the face-image region, with
Form the feature vector of every facial image sample, such as histograms of oriented gradients (Histogram of Oriented
Gradient, HOG) feature vector.
Wherein, the predetermined computation rule includes:
Wherein, Gx(x, y) and Gy(x, y) respectively indicates the ladder horizontally and vertically at current pixel point (x, y)
Angle value,Gradient-norm and the direction at current pixel point (x, y) are respectively indicated with θ (x, y).
Step S130, the corresponding with facial image first default label (includes but are not limited to the year of user for identification
Age, gender and character trait, such as " age: 10-15 years old "；" gender: female "；" personality: export-oriented ").Using in advance training with
Corresponding first identification model of the first default label, input described eigenvector may recognize that and the facial image pair
The default label of first answered.
In the present embodiment, first identification model be support vector machines (Support Vector Machine, SVM,
It is a kind of common method of discrimination, is the learning model for having supervision in machine learning field, is commonly used to carry out mode knowledge
Not, classification and regression analysis), the training process of corresponding first identification model of the first default label includes:
The facial image sample of the second preset quantity (such as 100,000) is obtained, is that each facial image sample distributes one
Unique first default label；
The facial image sample is divided into the first training set and the first training according to the first preset ratio (such as 4:1)
Collect, the image pattern quantity in first training set is greater than the image pattern quantity in first training set；
According to the predetermined characteristic vector pickup algorithm, extracts first training set and the first verifying is concentrated
The feature vector of every facial image sample, such as histograms of oriented gradients (Histogram of Oriented Gradient,
HOG) feature vector；
Image pattern in the training set is inputted the support vector machines to be trained, every predetermined period (such as
1000 iteration of every progress) support vector machines is verified using verifying collection, each face is concentrated using verifying
The feature vector of image pattern and corresponding first default label verify the accuracy rate of first identification model；
When the accuracy rate of verifying is greater than the second preset threshold (such as 95%), terminate training, obtains first identification
Model；And
When the accuracy rate of verifying is less than the second preset threshold (such as 95%), then increase the number of the facial image sample
Amount, and above-mentioned steps are re-executed based on increased facial image sample.
Step S140, for obtaining second default label of the corresponding presupposed information of each books as each books.
In the present embodiment, the presupposed information includes that the history for every books being stored in Library Information System is borrowed
It reads the age distribution of user, gender, borrow duration, character trait (such as export-oriented, introversive).In addition, the second default label
It can also include the build-in attribute of books, such as the type of books, length, the publication time limit.The default label information of the second of books
As the running in library is regular or real-time update.
Further, this method further include: establish the mapping relations between the described second default label and book categories.
Different books are sorted out, such as science and technology, life kind etc. are different classes of.User oneself is liked in inquiry
The classification that books can be first selected before books, further reduces range of choice, improves the recommendation precision of system.
Step S150 is matched from the multiple second default labels and is greater than or equal to the with the first default label similarity value
Second default label of the first preset quantity (such as 5) of one preset threshold (such as 85%).
Step S160 generates two dimensional code corresponding with recommendation list and shows user terminal 14.
It is determined according to the second default label matched and recommends book list, and generated and the recommendation list corresponding two
Tie up code.Wherein, two dimensional code can obtain whole relevant informations of books, author, publication time including books, it is simple introduce,
The academic probation of former chapters, other people evaluation, borrow the frequency recently, can also show surplus, the location information of books, and according to user
Current location generate route map guide user go to the position where books.
For example, in practical application, it is (or logical that user carries out recognition of face first before the computer camera in library
Cross mobile phone and log in library automation, mobile phone camera is utilized to carry out recognition of face) obtain user the first default label (such as
" age: 10-15 years old "；" gender: female "；" personality: export-oriented "), the book categories oneself liked then are inputted on the computer page,
System is matched from the multiple second default labels to be formed with the first default higher 5 the second default labels of label similarity value
Recommend book list, and generate two dimensional code corresponding with recommendation list and show user, user obtains recommending book list by barcode scanning.
In addition, the embodiment of the present invention also proposes a kind of computer readable storage medium, computer readable storage medium can be with
It is hard disk, multimedia card, SD card, flash card, SMC, read-only memory (ROM), Erasable Programmable Read Only Memory EPROM
(EPROM), any one in portable compact disc read-only memory (CD-ROM), USB storage etc. or several any
Combination.It include intelligent recommendation book list program 10 in computer readable storage medium, the computer readable storage medium of the present invention
Specific embodiment is roughly the same with the specific embodiment of above-mentioned intelligent recommendation book list method and server, no longer superfluous herein
It states.
It should be noted that the serial number of the above embodiments of the invention is only for description, do not represent the advantages or disadvantages of the embodiments.And
The terms "include", "comprise" herein or any other variant thereof is intended to cover non-exclusive inclusion, so that packet
Process, device, article or the method for including a series of elements not only include those elements, but also including being not explicitly listed
Other element, or further include for this process, device, article or the intrinsic element of method.Do not limiting more
In the case where, the element that is limited by sentence "including a ...", it is not excluded that including process, device, the article of the element
Or there is also other identical elements in method.
The serial number of the above embodiments of the invention is only for description, does not represent the advantages or disadvantages of the embodiments.Pass through above embodiment party
The description of formula, it is required general that those skilled in the art can be understood that above-described embodiment method can add by software
The mode of hardware platform is realized, naturally it is also possible to which by hardware, but in many cases, the former is more preferably embodiment.It is based on
Such understanding, substantially the part that contributes to existing technology can be with software product in other words for technical solution of the present invention
Form embody, which is stored in a storage medium (such as ROM/RAM, magnetic disk, light as described above
Disk) in, including some instructions use is so that a terminal device (can be mobile phone, computer, server or the network equipment
Deng) execute method described in each embodiment of the present invention.
The above is only a preferred embodiment of the present invention, is not intended to limit the scope of the invention, all to utilize this hair
Equivalent structure or equivalent flow shift made by bright specification and accompanying drawing content is applied directly or indirectly in other relevant skills
Art field, is included within the scope of the present invention.
Claims (10)
1. a kind of method of intelligent recommendation book list is applied to server, which is characterized in that this method comprises:
Obtaining step: the facial image of user terminal acquisition is obtained, the facial image is pre-processed；
Extraction step: using preset characteristic vector pickup algorithm extracted from the pretreated facial image feature to
Amount；
Identification step: the first identification model corresponding with the first default label of described eigenvector input training in advance is known
Not Chu the facial image the first default label；
Establishment step: second default label of the corresponding presupposed information of each books as each books is obtained；
Comparison step: it is matched from multiple second default labels and is greater than or waits with the described first default label similarity value
In the second default label of the first preset quantity of the first preset threshold；And
Generation step: determining according to the second default label of first preset quantity matched and recommend book list, generation and institute
It states the corresponding two dimensional code of recommendation list and feeds back to the user terminal.
2. the method for intelligent recommendation book list as described in claim 1, which is characterized in that first identification model be support to
The training process of amount machine, corresponding first identification model of the first default label includes:
The facial image sample of the second preset quantity is obtained, is that each facial image sample distribution one unique first is default
Label；
The facial image sample is divided into the first training set according to the first preset ratio and the first verifying collects, first training
The facial image sample size of concentration is greater than the facial image sample size that first verifying is concentrated；
According to the predetermined characteristic vector pickup algorithm, extracts first training set and the first verifying concentrates every
The feature vector of facial image sample；
Facial image sample in first training set is inputted the support vector machines to be trained, is made every predetermined period
The support vector machines is verified with the first verifying collection, concentrates each facial image sample using first verifying
Feature vector and corresponding first default label the accuracy rate of first identification model is verified；And
When the accuracy rate of verifying is greater than the second preset threshold, terminates training, obtain first identification model.
3. the method for intelligent recommendation book list as claimed in claim 2, which is characterized in that described eigenvector extraction algorithm packet
It includes:
By the second identification mould corresponding with the face-image region of the facial image of facial image input training in advance
Type identifies the face-image region of the facial image；
It is the elementary area of the second preset quantity by each face-image region cutting, utilizes predetermined computation rule
The gradient-norm and gradient direction of each pixel in described image unit is calculated；
The histogram of gradients of each described image unit is established according to the calculated gradient-norm and gradient direction；And
Each histogram of gradients is formed into every face figure in the position in the face-image region according to correspondence image unit
The feature vector of picture.
4. the method for intelligent recommendation book list as claimed in any one of claims 1-3, which is characterized in that the second default class
Type identification model is convolutional neural networks model, and the training process of the convolutional neural networks model is as follows:
The facial image sample of third preset quantity is obtained, is labeled with face-image region in every facial image sample；
The facial image sample is divided into the second training set according to the second preset ratio and the second verifying collects, second training
The image pattern quantity of concentration is greater than the image pattern quantity that second verifying is concentrated；
Facial image sample in second training set is inputted the convolutional neural networks model to be trained, every default
Period verifies the convolutional neural networks model using the second verifying collection, concentrates each using second verifying
The accuracy rate of second identification model is verified in facial image and corresponding face-image region；And
When the accuracy rate of verifying is greater than third predetermined threshold value, terminates training, obtain second identification model.
5. the method for intelligent recommendation book list as claimed in claim 3, which is characterized in that the predetermined computation rule packet
It includes:
Wherein, Gx(x, y) and Gy(x, y) respectively indicates the gradient horizontally and vertically at current pixel point (x, y)
Value,Gradient-norm and the direction at current pixel point (x, y) are respectively indicated with θ (x, y).
6. the method for intelligent recommendation book list as described in claim 1, which is characterized in that before the comparison step, the party
Method is further comprising the steps of:
Establish the mapping relations between the described second default label and book categories.
7. a kind of server, which is characterized in that the server includes memory and processor, is stored with intelligence on the memory
Energy recommendation one way sequence, the intelligent recommendation book list program realize following steps when being executed by the processor:
Obtaining step: the facial image of user terminal acquisition is obtained, the facial image is pre-processed；
Extraction step: using preset characteristic vector pickup algorithm extracted from the pretreated facial image feature to
Amount；
Identification step: the first identification model corresponding with the first default label of described eigenvector input training in advance is known
Not Chu the facial image the first default label；
Establishment step: second default label of the corresponding presupposed information of each books as each books is obtained；
Comparison step: it is matched from multiple second default labels and is greater than or waits with the described first default label similarity value
In the second default label of the first preset quantity of the first preset threshold；And
Generation step: determining according to the second default label of first preset quantity matched and recommend book list, generation and institute
It states the corresponding two dimensional code of recommendation list and feeds back to the user terminal.
8. server as claimed in claim 7, which is characterized in that first identification model is support vector machines, described the
The training process of corresponding first identification model of one default label includes:
The facial image sample of the second preset quantity is obtained, is that each facial image sample distribution one unique first is default
Label；
The facial image sample is divided into the first training set according to the first preset ratio and the first verifying collects, first training
The facial image sample size of concentration is greater than the facial image sample size that first verifying is concentrated；
According to the predetermined characteristic vector pickup algorithm, extracts first training set and the first verifying concentrates every
The feature vector of facial image sample；
Facial image sample in first training set is inputted the support vector machines to be trained, is made every predetermined period
The support vector machines is verified with the first verifying collection, concentrates each facial image sample using first verifying
Feature vector and corresponding first default label the accuracy rate of first identification model is verified；And
When the accuracy rate of verifying is greater than the second preset threshold, terminates training, obtain first identification model.
9. server as claimed in claim 8, which is characterized in that described eigenvector extraction algorithm includes:
By the second identification mould corresponding with the face-image region of the facial image of facial image input training in advance
Type identifies the face-image region of the facial image；
It is the elementary area of the second preset quantity by each face-image region cutting, utilizes predetermined computation rule
The gradient-norm and gradient direction of each pixel in described image unit is calculated；
The histogram of gradients of each described image unit is established according to the calculated gradient-norm and gradient direction；And
Each histogram of gradients is formed into every face figure in the position in the face-image region according to correspondence image unit
The feature vector of picture.
10. a kind of computer readable storage medium, which is characterized in that be stored with intelligence on the computer readable storage medium and push away
Book list program is recommended, the intelligent recommendation book list program can be executed by one or more processor, to realize such as claim 1-6
Any one of described in intelligent recommendation book list method the step of.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910298660.5A CN110135889A (en) | 2019-04-15 | 2019-04-15 | Method, server and the storage medium of intelligent recommendation book list |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910298660.5A CN110135889A (en) | 2019-04-15 | 2019-04-15 | Method, server and the storage medium of intelligent recommendation book list |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110135889A true CN110135889A (en) | 2019-08-16 |
Family
ID=67569934
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910298660.5A Pending CN110135889A (en) | 2019-04-15 | 2019-04-15 | Method, server and the storage medium of intelligent recommendation book list |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110135889A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110990670A (en) * | 2019-10-30 | 2020-04-10 | 华东师范大学 | Growth incentive book recommendation method and system |
| CN112035738A (en) * | 2020-08-14 | 2020-12-04 | 北京奇艺世纪科技有限公司 | Electronic book list recommendation method and device and electronic equipment |
| CN112150194A (en) * | 2020-09-14 | 2020-12-29 | 卖点国际展示(深圳)有限公司 | Marketing analysis method, system and storage medium |
| CN112307530A (en) * | 2020-09-22 | 2021-02-02 | 郑州轻工业大学 | Big data based binding design method and device for reading books of children |
| CN112464093A (en) * | 2020-11-30 | 2021-03-09 | 江苏商贸职业学院 | Reader-oriented intelligent book searching robot and electronic equipment |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105760547A (en) * | 2016-03-16 | 2016-07-13 | 中山大学 | Book recommendation method and system based on user clustering |
| CN106407418A (en) * | 2016-09-23 | 2017-02-15 | Tcl集团股份有限公司 | A face identification-based personalized video recommendation method and recommendation system |
| CN107977633A (en) * | 2017-12-06 | 2018-05-01 | 平安科技（深圳）有限公司 | Age recognition methods, device and the storage medium of facial image |
| CN108595628A (en) * | 2018-04-24 | 2018-09-28 | 百度在线网络技术（北京）有限公司 | Method and apparatus for pushed information |
| CN108763897A (en) * | 2018-05-22 | 2018-11-06 | 平安科技（深圳）有限公司 | Method of calibration, terminal device and the medium of identity legitimacy |
| CN109299709A (en) * | 2018-12-04 | 2019-02-01 | 中山大学 | Data recommendation method, device, server end and client based on recognition of face |
| CN109558535A (en) * | 2018-11-05 | 2019-04-02 | 重庆中科云丛科技有限公司 | The method and system of personalized push article based on recognition of face |
-
2019
- 2019-04-15 CN CN201910298660.5A patent/CN110135889A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105760547A (en) * | 2016-03-16 | 2016-07-13 | 中山大学 | Book recommendation method and system based on user clustering |
| CN106407418A (en) * | 2016-09-23 | 2017-02-15 | Tcl集团股份有限公司 | A face identification-based personalized video recommendation method and recommendation system |
| CN107977633A (en) * | 2017-12-06 | 2018-05-01 | 平安科技（深圳）有限公司 | Age recognition methods, device and the storage medium of facial image |
| CN108595628A (en) * | 2018-04-24 | 2018-09-28 | 百度在线网络技术（北京）有限公司 | Method and apparatus for pushed information |
| CN108763897A (en) * | 2018-05-22 | 2018-11-06 | 平安科技（深圳）有限公司 | Method of calibration, terminal device and the medium of identity legitimacy |
| CN109558535A (en) * | 2018-11-05 | 2019-04-02 | 重庆中科云丛科技有限公司 | The method and system of personalized push article based on recognition of face |
| CN109299709A (en) * | 2018-12-04 | 2019-02-01 | 中山大学 | Data recommendation method, device, server end and client based on recognition of face |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110990670A (en) * | 2019-10-30 | 2020-04-10 | 华东师范大学 | Growth incentive book recommendation method and system |
| CN110990670B (en) * | 2019-10-30 | 2023-11-10 | 华东师范大学 | Growth incentive book recommendation method and recommendation system |
| CN112035738A (en) * | 2020-08-14 | 2020-12-04 | 北京奇艺世纪科技有限公司 | Electronic book list recommendation method and device and electronic equipment |
| CN112035738B (en) * | 2020-08-14 | 2023-09-26 | 北京奇艺世纪科技有限公司 | Electronic bill recommendation method and device and electronic equipment |
| CN112150194A (en) * | 2020-09-14 | 2020-12-29 | 卖点国际展示(深圳)有限公司 | Marketing analysis method, system and storage medium |
| CN112307530A (en) * | 2020-09-22 | 2021-02-02 | 郑州轻工业大学 | Big data based binding design method and device for reading books of children |
| CN112307530B (en) * | 2020-09-22 | 2024-02-02 | 郑州轻工业大学 | Method and device for framing book of child books based on big data |
| CN112464093A (en) * | 2020-11-30 | 2021-03-09 | 江苏商贸职业学院 | Reader-oriented intelligent book searching robot and electronic equipment |
| CN112464093B (en) * | 2020-11-30 | 2023-04-18 | 江苏商贸职业学院 | Reader-oriented intelligent book searching robot and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110135889A (en) | Method, server and the storage medium of intelligent recommendation book list | |
| CN112232293B (en) | Image processing model training method, image processing method and related equipment | |
| CN110033018B (en) | Graph similarity judging method and device and computer readable storage medium | |
| US9098888B1 (en) | Collaborative text detection and recognition | |
| CN112396613B (en) | Image segmentation method, device, computer equipment and storage medium | |
| CN110009556A (en) | Image background weakening method, device, storage medium and electronic equipment | |
| CN110506276A (en) | The efficient image analysis of use environment sensing data | |
| CN108629043A (en) | Extracting method, device and the storage medium of webpage target information | |
| CN112016438A (en) | Method and system for identifying certificate based on graph neural network | |
| CN113705554A (en) | Training method, device and equipment of image recognition model and storage medium | |
| CN112396049A (en) | Text error correction method and device, computer equipment and storage medium | |
| CN113704531A (en) | Image processing method, image processing device, electronic equipment and computer readable storage medium | |
| US20090228510A1 (en) | Generating congruous metadata for multimedia | |
| CN110135421A (en) | Licence plate recognition method, device, computer equipment and computer readable storage medium | |
| CN110008997A (en) | Image texture similarity recognition method, device and computer readable storage medium | |
| CN113939791A (en) | Image labeling method, device, equipment and medium | |
| CN109300041A (en) | Typical karst ecosystem recommended method, electronic device and readable storage medium storing program for executing | |
| CN108648189A (en) | Image fuzzy detection method, apparatus, computing device and readable storage medium storing program for executing | |
| CN108268641A (en) | Invoice information recognition methods and invoice information identification device, equipment and storage medium | |
| CN110781925A (en) | Software page classification method and device, electronic equipment and storage medium | |
| CN108932533A (en) | Identification model construction method and device, character identifying method and device | |
| He et al. | Aggregating local context for accurate scene text detection | |
| CN114581710A (en) | Image recognition method, device, equipment, readable storage medium and program product | |
| CN112613508A (en) | Object identification method, device and equipment | |
| Rehman et al. | Efficient coarser‐to‐fine holistic traffic sign detection for occlusion handling |
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 |