CN108921185A - A kind of shelf sales promotion information recognition methods based on image recognition, device and system - Google Patents
A kind of shelf sales promotion information recognition methods based on image recognition, device and system Download PDFInfo
- Publication number
- CN108921185A CN108921185A CN201810421154.6A CN201810421154A CN108921185A CN 108921185 A CN108921185 A CN 108921185A CN 201810421154 A CN201810421154 A CN 201810421154A CN 108921185 A CN108921185 A CN 108921185A
- Authority
- CN
- China
- Prior art keywords
- text
- training
- pictures
- information
- recognition
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The shelf sales promotion information recognition methods that the invention discloses a kind of based on image recognition, device and system, method acquire training pictures using camera;Then the markup information and training text of picture in training pictures are obtained;Training obtains the identification engine for having promotion mark recognition capability, line of text recognition capability and text recognition;Acquisition shelf picture is input to identification engine as test pictures and after being pre-processed and is identified to obtain sales promotion information.It is trained to obtain identification engine by inputting training pictures to training pattern in the present invention, only need to obtain the identification that artificial markup information can be automatically performed sales promotion information, recognition speed is fast, accuracy rate is high, complete sales promotion information data can be obtained, conducive to subsequent data statistic analysis and decision, working efficiency is improved.The present invention can be widely applied to field of image recognition as a kind of shelf sales promotion information recognition methods based on image recognition, device and system.
Description
Technical field
The present invention relates to field of image recognition, especially a kind of shelf sales promotion information recognition methods based on image recognition,
Device and system.
Background technique
Demand of the Kuai Xiao product enterprise to the sales promotion information statistics of each terminal sale channel under line is obtained is very urgent always, needs
Shops's visit periodically carried out to channel each under line, and acquire the sales promotion information of commodity, to understand promotion strategy in channel terminal
Whether execute in place, the decision of promotion strategy and market management to enterprise is significant.
Brand quotient is it is to be understood that its commodity the sales promotion information on major shopping mall supermarket shelf the case where, needs to arrange special pin
Representative or office worker is sold to specified shopping mall supermarket to carry out sales promotion information acquisition to the shelf for putting its commodity, count and will unite
Result feedback is counted to company, current solution has the following two kinds:
1, representative of sales & marketing or related office worker carry out manually counting sales promotion information before shelf, i.e. artificial acquisition promotion at the scene
Listed, text information of promotion paster of information etc., Information Statistics after the completion of work summarized and submitted;
2, representative of sales & marketing or related office worker take pictures before shelf and acquire its relevant sales promotion information photo, after the completion of acquisition
The related promotion etc. counted needed for recording in picture, Information Statistics after the completion of work summarized and submitted.
It usually requiring that a large amount of manpowers is spent to assess work to realize in existing scheme, working efficiency is low, and quality varies with each individual,
It is difficult to unified control, and is confined to human cost, can only often sample collecting part data, be difficult in data integrity full
The requirement of sufficient mass data analysis.And in the prior art, such as a kind of identification of pallets method based on image recognition technology
(CN107045641A), it is provided solely for the recognition methods of shelf Global Information, it is difficult to which the characteristics of being directed to sales promotion information carries out quasi-
Really identify.
Summary of the invention
In order to solve the above-mentioned technical problem, the purpose of the present invention is:A kind of realization based on image recognition technology is provided to mention
Picking frame sales promotion information carries out knowledge method for distinguishing.
In order to solve the above-mentioned technical problem, it is another object of the present invention to:A kind of reality based on image recognition technology is provided
Now extract the device that shelf sales promotion information is identified.
In order to solve the above-mentioned technical problem, it is another object of the present invention to:A kind of reality based on image recognition technology is provided
Now extract the system that shelf sales promotion information is identified.
The technical scheme adopted by the invention is that:A kind of shelf sales promotion information recognition methods based on image recognition, including
There are following steps:
360 degree of camera shooting acquisition shelf pictures are carried out as training pictures using camera;
The markup information of picture in training pictures is obtained, the markup information includes promotion home position markup information
With line of text markup information;
Obtain the training text for being used for text identification model training;
Training pictures and training text are input to training airplane, carry out the training of promotion label detection, line of text detection instruction
Experienced and text identification model training obtains having promotion mark recognition capability, line of text recognition capability and text recognition
Identify engine;
Acquisition shelf picture is as test pictures and test pictures image data pretreatment is carried out to it;
The pretreated test pictures of image data are input to identification engine to be identified to obtain sales promotion information.
Further, it is described promotion label detection training the specific steps are:
It detects identification network to RFCN according to the picture and markup information trained in pictures to be trained, to obtain
The promotion mark identification model for having promotion mark recognition capability.
Further, line of text detection training the specific steps are:
It detects identification network to RFCN according to the picture and markup information trained in pictures to be trained, to obtain
The line of text identification model for having line of text recognition capability.
Further, the text identification model training the specific steps are:
The training text collection with label character information and digital markup information is generated based on training text, then using instruction
Practice text set to be trained the CNN-LSTM network based on Attension mechanism, to obtain having text recognition
Text identification model.
Another technical solution of the present invention is:A kind of shelf sales promotion information identification device based on image recognition,
Include:
Image capture module, for using camera carry out 360 degree camera shootings acquisition shelf pictures be used as train pictures or
Test pictures;
Labeling module, for obtaining the markup information of picture in trained pictures, the markup information includes promotion mark
Know position markup information and line of text markup information;
Text collection module, for obtaining the training text for being used for text identification model training;
Training module, for pictures and training text will to be trained to be input to training airplane, carry out the training of promotion label detection,
Line of text detection training and text identification model training obtain having promotion mark recognition capability, line of text recognition capability and text
The identification engine of this recognition capability;
Preprocessing module, for carrying out image data pretreatment to test pictures;
Engine is identified, for being identified to obtain sales promotion information to the pretreated test pictures of image data.
Further, the identification engine includes promotion mark identification model, line of text identification model and text identification mould
Type.
Further, the training module includes promotion label detection training submodule, for according in training pictures
Picture and markup information identification network is detected to RFCN and is trained, to obtain having the rush of promotion mark recognition capability
Pin mark identification model.
Further, the training module includes that line of text detects training submodule, for according in training pictures
Picture and markup information detect identification network to RFCN and are trained, to obtain the line of text for having line of text recognition capability
Identification model.
Further, the training module includes text identification model training submodule, for being generated based on training text
Training text collection with label character information and digital markup information, then using training text collection to based on Attension
The CNN-LSTM network of mechanism is trained, to obtain the text identification model for having text recognition.
Another technical solution of the present invention is:A kind of shelf sales promotion information identifying system based on image recognition,
Include:
At least one processor;
At least one processor, for storing at least one program;
When at least one described program is executed by least one described processor, so that at least one described processor is realized
Such as a kind of above-mentioned shelf sales promotion information recognition methods based on image recognition.
Beneficial effects of the present invention are:It is trained, is obtained by inputting training pictures to training pattern in the present invention
The identification engine for having identification sales promotion information ability, it is only necessary to obtain the knowledge that artificial markup information can be automatically performed sales promotion information
Not, recognition speed is fast, accuracy rate is high, can obtain complete sales promotion information data, is conducive to subsequent data statistic analysis and decision,
Improve working efficiency.
Detailed description of the invention
Fig. 1 is the step flow chart of the method for the present invention.
Specific embodiment
Specific embodiments of the present invention will be further explained with reference to the accompanying drawing:
Referring to Fig.1, a kind of shelf sales promotion information recognition methods based on image recognition, includes following steps:
360 degree of camera shooting acquisition shelf pictures are carried out as training pictures using camera;Utilize high-definition camera and rotation
Turn base device, round-the-clock automatic collection shelf image;
The markup information of picture in training pictures is obtained, the markup information includes promotion home position markup information
With line of text markup information;
The markup information can be realized by marking software system:The shelf pictures that will be identified comprising promotion by administrator
Task is uploaded and assigned, personnel is marked and enters mark interface in personal task list, the position that mark is promoted in picture is used
Box mark;Promotion mark pictures under being cut by administrator upload and assign task, and mark personnel arrange in personal task
Each line of text in picture is marked with box in table.Mark completes system evaluation accuracy after result is submitted, administrator's root
It determines whether mark personnel need to mark again according to accuracy, ensure that obtain in this method above-mentioned steps and scheme in training pictures
The accuracy of the markup information of piece;
Employee's account can be used to log in for the software systems, using pc client or the form of mobile client, system according to
Mobile phone, which positions, determines staff position (GPS is anti-fake), and staff takes pictures to shelf and uploads and now take a picture that (picture is anti-
It is pseudo-), system rapidly carries out shelf beyond the clouds and verifies work, generates report, and report is synchronized to the App of enterprise administrator
In end, enterprise administrator can carry out information check to picture;
Obtain the training text for being used for text identification model training;
The source of training text can there are many, such as using crawler by microblogging or other social networks writing text believe
Breath is crawled, and is then required according to the text information and required dimension of picture, font etc. that crawl, is generated using class identifying code
Mode generates corresponding textual image, composing training text;
Training pictures and training text are input to training airplane, carry out the training of promotion label detection, line of text detection instruction
Experienced and text identification model training obtains having promotion mark recognition capability, line of text recognition capability and text recognition
Identify engine;
Acquisition shelf picture is as test pictures and test pictures image data pretreatment is carried out to it;
The pretreated test pictures of image data are input to identification engine to be identified to obtain sales promotion information;It is described
Shelf picture is put into identification engine in the function of having had detection identification favor information mark after training by identification engine
Afterwards, engine one by one outlines promotion mark all on shelf by promoting label detection network first, further according to the figure in frame
Piece information detects network by line of text and detects the position of each style of writing originally, finally passes through text to each row textual image and knows
Other model is identified, all favor informations are extracted, and generates corresponding data or report;
It is trained in the present invention by inputting training pictures to training pattern, obtains having identification sales promotion information ability
Identification engine, it is only necessary to obtain the identification that artificial markup information can be automatically performed sales promotion information, recognition speed is fast, accuracy rate
Height can obtain complete sales promotion information data, be conducive to subsequent data statistic analysis and decision, improve working efficiency.
Be further used as preferred embodiment, the promotion label detection training the specific steps are:
It detects identification network to RFCN according to the picture and markup information trained in pictures to be trained, to obtain
The promotion mark identification model for having promotion mark recognition capability;
The method that identification of pallets mainly uses target detection obtains the data with mark, the instruction as an only class object
Practice collection and identification network is detected to RFCN and be trained, trains the mould that can be detected promotion mark from shelf picture
Type promotes basis of the label detection as subsequent step.
Be further used as preferred embodiment, the line of text detection training the specific steps are:
It detects identification network to RFCN according to the picture and markup information trained in pictures to be trained, to obtain
The line of text identification model for having line of text recognition capability.
Be further used as preferred embodiment, the text identification model training the specific steps are:
The training text collection with label character information and digital markup information is generated based on training text, then using instruction
Practice text set to be trained the CNN-LSTM network based on Attension mechanism, to obtain having text recognition
Text identification model;
It is trained for sales promotion information using the CNN-LSTM network based on Attension mechanism, effectively promotion network
Robustness, have identification random length text information textual image ability;Avoid in existing text identification network
When being identified for sales promotion information, the shortcomings that the textual image of fixed length can only be identified.
A kind of corresponding above-mentioned shelf sales promotion information recognition methods based on image recognition, the present invention also provides one kind based on figure
As the shelf sales promotion information identification device of identification, include:
Image capture module, for using camera carry out 360 degree camera shootings acquisition shelf pictures be used as train pictures or
Test pictures;
Labeling module, for obtaining the markup information of picture in trained pictures, the markup information includes promotion mark
Know position markup information and line of text markup information;
Text collection module, for obtaining the training text for being used for text identification model training;
Training module, for pictures and training text will to be trained to be input to training airplane, carry out the training of promotion label detection,
Line of text detection training and text identification model training obtain having promotion mark recognition capability, line of text recognition capability and text
The identification engine of this recognition capability;
Preprocessing module, for carrying out image data pretreatment to test pictures;
Engine is identified, for being identified to obtain sales promotion information to the pretreated test pictures of image data.
It is further used as preferred embodiment, the identification engine includes promotion mark identification model, line of text knowledge
Other model and text identification model.
It is further used as preferred embodiment, the training module includes promotion label detection training submodule, is used
It is trained in detecting identification network to RFCN according to the picture and markup information trained in pictures, to obtain having rush
The promotion of pin mark recognition capability identifies identification model.
It is further used as preferred embodiment, the training module includes that line of text detects training submodule, is used for
It detects identification network to RFCN according to the picture and markup information trained in pictures to be trained, to obtain having text
The line of text identification model of row recognition capability.
It is further used as preferred embodiment, the training module includes text identification model training submodule, is used
In generating the training text collection with label character information and digital markup information based on training text, training text is then used
Collection is trained the CNN-LSTM network based on Attension mechanism, so that the text for obtaining having text recognition is known
Other model.
A kind of corresponding above-mentioned shelf sales promotion information recognition methods based on image recognition, the present invention also provides one kind based on figure
As the shelf sales promotion information identifying system of identification, include:
At least one processor;
At least one processor, for storing at least one program;
When at least one described program is executed by least one described processor, so that at least one described processor is realized
Such as a kind of above-mentioned shelf sales promotion information recognition methods based on image recognition.
It is to be illustrated to preferable implementation of the invention, but the invention is not limited to the implementation above
Example, those skilled in the art can also make various equivalents on the premise of without prejudice to spirit of the invention or replace
It changes, these equivalent deformations or replacement are all included in the scope defined by the claims of the present application.
Claims (10)
1. a kind of shelf sales promotion information recognition methods based on image recognition, which is characterized in that include following steps:
360 degree of camera shooting acquisition shelf pictures are carried out as training pictures using camera;
The markup information of picture in training pictures is obtained, the markup information includes promotion home position markup information and text
Current row markup information;
Obtain the training text for being used for text identification model training;
Will training pictures and training text be input to training airplane, carry out promotion label detection training, line of text detection training and
Text identification model training obtains the identification for having promotion mark recognition capability, line of text recognition capability and text recognition
Engine;
Acquisition shelf picture is as test pictures and test pictures image data pretreatment is carried out to it;
The pretreated test pictures of image data are input to identification engine to be identified to obtain sales promotion information.
2. a kind of shelf sales promotion information recognition methods based on image recognition according to claim 1, which is characterized in that institute
State promotion label detection training the specific steps are:
It detects identification network to RFCN according to the picture and markup information trained in pictures to be trained, to be had
The promotion of promotion mark recognition capability identifies identification model.
3. a kind of shelf sales promotion information recognition methods based on image recognition according to claim 1, which is characterized in that institute
State line of text detection training the specific steps are:
It detects identification network to RFCN according to the picture and markup information trained in pictures to be trained, to be had
The line of text identification model of line of text recognition capability.
4. a kind of shelf sales promotion information recognition methods based on image recognition according to claim 1, which is characterized in that institute
State text identification model training the specific steps are:
The training text collection with label character information and digital markup information is generated based on training text, then using training text
This collection is trained the CNN-LSTM network based on Attension mechanism, to obtain the text for having text recognition
Identification model.
5. a kind of shelf sales promotion information identification device based on image recognition, it is characterised in that:Include
Image capture module, for carrying out 360 degree of camera shooting acquisition shelf pictures as training pictures or test using camera
Pictures;
Labeling module, for obtaining the markup information of picture in trained pictures, the markup information includes promotion marker
Set markup information and line of text markup information;
Text collection module, for obtaining the training text for being used for text identification model training;
Training module carries out the training of promotion label detection, text for pictures and training text will to be trained to be input to training airplane
Row detects trained and text identification model training, obtains having promotion mark recognition capability, line of text recognition capability and text knowledge
The identification engine of other ability;
Preprocessing module, for carrying out image data pretreatment to test pictures;
Engine is identified, for being identified to obtain sales promotion information to the pretreated test pictures of image data.
6. a kind of shelf sales promotion information identification device based on image recognition according to claim 5, it is characterised in that:Institute
Stating identification engine includes promotion mark identification model, line of text identification model and text identification model.
7. a kind of shelf sales promotion information identification device based on image recognition according to claim 6, it is characterised in that:Institute
Stating training module includes promotion label detection training submodule, for according to the picture and markup information in training pictures
It detects identification network to RFCN to be trained, to obtain the promotion mark identification model for having promotion mark recognition capability.
8. a kind of shelf sales promotion information identification device based on image recognition according to claim 6, it is characterised in that:Institute
Stating training module includes that line of text detects training submodule, for according to the picture and markup information pair in training pictures
RFCN detection identification network is trained, to obtain the line of text identification model for having line of text recognition capability.
9. a kind of shelf sales promotion information identification device based on image recognition according to claim 6, it is characterised in that:Institute
Stating training module includes text identification model training submodule, for being generated based on training text with label character information and
The training text collection of digital markup information, then using training text collection to the CNN-LSTM network based on Attension mechanism
It is trained, to obtain the text identification model for having text recognition.
10. a kind of shelf sales promotion information identifying system based on image recognition, which is characterized in that include:
At least one processor;
At least one processor, for storing at least one program;
When at least one described program is executed by least one described processor, so that at least one described processor is realized as weighed
Benefit requires a kind of shelf sales promotion information recognition methods based on image recognition described in any claim in 1-4.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810421154.6A CN108921185A (en) | 2018-05-04 | 2018-05-04 | A kind of shelf sales promotion information recognition methods based on image recognition, device and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810421154.6A CN108921185A (en) | 2018-05-04 | 2018-05-04 | A kind of shelf sales promotion information recognition methods based on image recognition, device and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108921185A true CN108921185A (en) | 2018-11-30 |
Family
ID=64402322
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810421154.6A Pending CN108921185A (en) | 2018-05-04 | 2018-05-04 | A kind of shelf sales promotion information recognition methods based on image recognition, device and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108921185A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109858475A (en) * | 2019-01-08 | 2019-06-07 | 平安科技(深圳)有限公司 | Picture character localization method, device, medium and computer equipment |
CN110222695A (en) * | 2019-06-19 | 2019-09-10 | 拉扎斯网络科技(上海)有限公司 | Certificate picture processing method and device, medium and electronic equipment |
CN114061784A (en) * | 2021-11-18 | 2022-02-18 | 贵州电网有限责任公司 | Transformer conducting rod lead head overheating predictive maintenance method based on machine vision |
Citations (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101571875A (en) * | 2009-05-05 | 2009-11-04 | 程治永 | Realization method of image searching system based on image recognition |
JP2010146064A (en) * | 2008-12-16 | 2010-07-01 | Sharp Corp | Purchase system, electronic shelf tag and order processing method |
CN101789111A (en) * | 2009-12-30 | 2010-07-28 | 中国科学院自动化研究所 | Site advertisement and system for real-time interaction of electronic advertisement data |
CN101809601A (en) * | 2007-08-31 | 2010-08-18 | 埃森哲环球服务有限公司 | Planogram extraction based on image processing |
CN102376061A (en) * | 2011-08-26 | 2012-03-14 | 浙江工业大学 | Omni-directional vision-based consumer purchase behavior analysis device |
US9589433B1 (en) * | 2013-07-31 | 2017-03-07 | Jeff Thramann | Self-checkout anti-theft device |
CN106777125A (en) * | 2016-12-16 | 2017-05-31 | 广东顺德中山大学卡内基梅隆大学国际联合研究院 | A kind of iamge description generation method based on neutral net and image attention point |
CN106952151A (en) * | 2017-04-19 | 2017-07-14 | 广州海舶服装有限公司 | A kind of method of commerce based on new retail intelligent steward and newly it is sold intelligent steward |
CN106960206A (en) * | 2017-02-08 | 2017-07-18 | 北京捷通华声科技股份有限公司 | Character identifying method and character recognition system |
CN107045641A (en) * | 2017-04-26 | 2017-08-15 | 广州图匠数据科技有限公司 | A kind of identification of pallets method based on image recognition technology |
CN107145845A (en) * | 2017-04-26 | 2017-09-08 | 中山大学 | The pedestrian detection method merged based on deep learning and multi-characteristic points |
CN107247950A (en) * | 2017-06-06 | 2017-10-13 | 电子科技大学 | A kind of ID Card Image text recognition method based on machine learning |
CN107527053A (en) * | 2017-08-31 | 2017-12-29 | 北京小米移动软件有限公司 | Object detection method and device |
CN107545460A (en) * | 2017-07-25 | 2018-01-05 | 广州智选网络科技有限公司 | One kind digitlization color page promotion management and analysis method, storage device and mobile terminal |
CN107562792A (en) * | 2017-07-31 | 2018-01-09 | 同济大学 | A kind of question and answer matching process based on deep learning |
CN107609460A (en) * | 2017-05-24 | 2018-01-19 | 南京邮电大学 | A kind of Human bodys' response method for merging space-time dual-network stream and attention mechanism |
CN107609945A (en) * | 2017-09-15 | 2018-01-19 | 泾县麦蓝网络技术服务有限公司 | A kind of Method of Commodity Recommendation and system applied to physical stores |
CN107767564A (en) * | 2017-09-28 | 2018-03-06 | 中南大学 | Automatic vending machine automatic defrosting system and method based on image recognition |
-
2018
- 2018-05-04 CN CN201810421154.6A patent/CN108921185A/en active Pending
Patent Citations (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101809601A (en) * | 2007-08-31 | 2010-08-18 | 埃森哲环球服务有限公司 | Planogram extraction based on image processing |
JP2010146064A (en) * | 2008-12-16 | 2010-07-01 | Sharp Corp | Purchase system, electronic shelf tag and order processing method |
CN101571875A (en) * | 2009-05-05 | 2009-11-04 | 程治永 | Realization method of image searching system based on image recognition |
CN101789111A (en) * | 2009-12-30 | 2010-07-28 | 中国科学院自动化研究所 | Site advertisement and system for real-time interaction of electronic advertisement data |
CN102376061A (en) * | 2011-08-26 | 2012-03-14 | 浙江工业大学 | Omni-directional vision-based consumer purchase behavior analysis device |
US9589433B1 (en) * | 2013-07-31 | 2017-03-07 | Jeff Thramann | Self-checkout anti-theft device |
CN106777125A (en) * | 2016-12-16 | 2017-05-31 | 广东顺德中山大学卡内基梅隆大学国际联合研究院 | A kind of iamge description generation method based on neutral net and image attention point |
CN106960206A (en) * | 2017-02-08 | 2017-07-18 | 北京捷通华声科技股份有限公司 | Character identifying method and character recognition system |
CN106952151A (en) * | 2017-04-19 | 2017-07-14 | 广州海舶服装有限公司 | A kind of method of commerce based on new retail intelligent steward and newly it is sold intelligent steward |
CN107045641A (en) * | 2017-04-26 | 2017-08-15 | 广州图匠数据科技有限公司 | A kind of identification of pallets method based on image recognition technology |
CN107145845A (en) * | 2017-04-26 | 2017-09-08 | 中山大学 | The pedestrian detection method merged based on deep learning and multi-characteristic points |
CN107609460A (en) * | 2017-05-24 | 2018-01-19 | 南京邮电大学 | A kind of Human bodys' response method for merging space-time dual-network stream and attention mechanism |
CN107247950A (en) * | 2017-06-06 | 2017-10-13 | 电子科技大学 | A kind of ID Card Image text recognition method based on machine learning |
CN107545460A (en) * | 2017-07-25 | 2018-01-05 | 广州智选网络科技有限公司 | One kind digitlization color page promotion management and analysis method, storage device and mobile terminal |
CN107562792A (en) * | 2017-07-31 | 2018-01-09 | 同济大学 | A kind of question and answer matching process based on deep learning |
CN107527053A (en) * | 2017-08-31 | 2017-12-29 | 北京小米移动软件有限公司 | Object detection method and device |
CN107609945A (en) * | 2017-09-15 | 2018-01-19 | 泾县麦蓝网络技术服务有限公司 | A kind of Method of Commodity Recommendation and system applied to physical stores |
CN107767564A (en) * | 2017-09-28 | 2018-03-06 | 中南大学 | Automatic vending machine automatic defrosting system and method based on image recognition |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109858475A (en) * | 2019-01-08 | 2019-06-07 | 平安科技(深圳)有限公司 | Picture character localization method, device, medium and computer equipment |
CN110222695A (en) * | 2019-06-19 | 2019-09-10 | 拉扎斯网络科技(上海)有限公司 | Certificate picture processing method and device, medium and electronic equipment |
CN114061784A (en) * | 2021-11-18 | 2022-02-18 | 贵州电网有限责任公司 | Transformer conducting rod lead head overheating predictive maintenance method based on machine vision |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2018126742A1 (en) | Electronic batch processing method and system for stored archives, and storage medium | |
CN111507283A (en) | Student behavior identification method and system based on classroom scene | |
CN108921185A (en) | A kind of shelf sales promotion information recognition methods based on image recognition, device and system | |
CN111401314A (en) | Dressing information detection method, device, equipment and storage medium | |
CN105373978A (en) | Artificial test paper judgment processing device and artificial test paper judgment processing method based on OCR | |
WO2016169499A1 (en) | Random anti-counterfeiting marker-based anti-counterfeiting system, and anti-counterfeiting method therefor | |
CN104408785A (en) | Distribution line tour-inspection system based on GIS and GPS | |
CN102982332A (en) | Retail terminal goods shelf image intelligent analyzing system based on cloud processing method | |
CN110865654A (en) | Power grid unmanned aerial vehicle inspection defect processing method | |
CN106056112A (en) | IMAGE PROCESSING APPARATUS and IMAGE PROCESSING METHOD | |
CN105205673B (en) | A kind of false proof trace-back technique using in bar cigarette stamp feature | |
CN109710705A (en) | Map point of interest treating method and apparatus | |
CN109213858B (en) | Automatic identification method and system for network water army | |
CN111178845A (en) | Data annotation system and method based on network service platform | |
WO2023103329A1 (en) | Data labeling method, apparatus, and system, device, and storage medium | |
CN107992527B (en) | Data mark checking method, server and storage medium | |
JP2012003632A (en) | Manufacturing process management system | |
CN110517123A (en) | A kind of unmanned bid evaluation system of AI based on technology of Internet of things and method | |
CN113076730A (en) | Test report intelligent management method, device, system and storage medium | |
CN110543839A (en) | commodity goods laying rate acquisition method based on computer vision | |
CN111666735A (en) | Image text label display method and device and storage medium | |
CN111047731A (en) | AR technology-based telecommunication room inspection method and system | |
CN108875830A (en) | A kind of commodity on shelf identified off-line method, system and device based on image recognition | |
CN113989268B (en) | Method for identifying GIS sleeve of specified equipment in infrared image and storage medium | |
CN111339939B (en) | Attendance checking method and device based on image recognition |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181130 |