CN110413874A - A kind of clothes recommended method based on dress ornament attributes match - Google Patents
A kind of clothes recommended method based on dress ornament attributes match Download PDFInfo
- Publication number
- CN110413874A CN110413874A CN201910521390.XA CN201910521390A CN110413874A CN 110413874 A CN110413874 A CN 110413874A CN 201910521390 A CN201910521390 A CN 201910521390A CN 110413874 A CN110413874 A CN 110413874A
- Authority
- CN
- China
- Prior art keywords
- picture
- attribute
- recommended
- clothes
- user
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/53—Querying
- G06F16/535—Filtering based on additional data, e.g. user or group profiles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- 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
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
Abstract
The invention discloses a kind of clothes recommended methods based on dress ornament attributes match, it include: (1) building clothes detection of attribute model, clothes attribute data collection is selected, category is trained clothes detection of attribute model, obtains the corresponding model parameter file of each classification;(2) garment image that selection user uploads carries out detection of attribute using the different classes of model of above-mentioned training, counts the corresponding frequency of each attribute in each classification;(3) garment inventory picture to be recommended is selected, detection of attribute is carried out to every picture respectively using the different classes of model of above-mentioned training, obtains the attribute that each classification is possessed in every picture to be recommended;(4) attributes similarity between every picture to be recommended and the picture of user's upload is calculated, and every picture to be recommended is ranked up by similarity size, according to the forward picture to be recommended of recommended amount selected and sorted.Using the present invention, makes to recommend that more there is specific aim, improve the accuracy rate of clothes recommendation.
Description
Technical field
The invention belongs to commercial product recommending technical fields, more particularly, to a kind of clothes attribute by matching user and businessman
Similarity is come the method that carries out clothes recommendation.
Background technique
Recommended technology is a kind of Characteristic of Interest according to user and buying behavior, to the interested information of user recommended user
With the recommended technology of commodity, recommended technology currently on the market is very mature, has been widely used in major electric business website,
Such as Amazon, Taobao, beauty, Suning, Wal-Mart etc. are gathered in Jingdone district.
With the increasingly raising of current living standard, people start increasingly to pay close attention to personal dress up.It usually can be in Taobao
This season fashion style is bought on equal websites, but in face of the clothes of magnanimity on website, buyer can usually be at a loss.Based on this
The personalized recommendation technology of a little demands, existing major website is mainly based upon the purchaser record of user, and user is in which clothes
Webpage residence time, and the recommendation based on user's similarity progress clothes, such as Taobao, they are mainly by the purchase before user
Record information storage is bought, personalized calculating is carried out, to analyze the preference of user;Either think to buy the user of same part clothes
Preference having the same calculates similarity between user, can be other purchases for the consumer for buying same clothes with the user
Commodity are recommended.
The Chinese patent literature of Publication No. CN109859004A discloses a kind of commercial product recommending side based on historical data
Method and system, for realizing the shopping information of combining target user, by the previous purchase article of user or the record of service,
The user for possessing similar interests or purchaser record with target user is found, calculates in neighbours and uses further according to polymerization mimicry function
The potential attachment relationship size of family and target user obtain the Recommendations of target user according to the purchase information of associated user
Information.
The main problem of existing technology has: traditional method is the purchaser record by calculating user mostly, browses net
Page the time carry out personality analysis user preferences, but this be it is more unilateral and unreasonable, user oneself buy not
It must be clothes required for himself, it is also possible to there is the case where helping others to buy on behalf, such as famous " beer and diaper " example
Son, using this as the basis of user preference analysis, there are errors.
Summary of the invention
Of the existing technology to overcome the problems, such as, the present invention provides a kind of clothes recommendation sides based on dress ornament attributes match
Method improves the accuracy rate of clothes recommendation.
Technical scheme is as follows:
A kind of clothes recommended method based on dress ornament attributes match, comprising the following steps:
(1) clothes detection of attribute model is constructed, clothes attribute data collection is selected, by whole Attribute transposition to several classes
In not, category is respectively trained clothes detection of attribute model, obtains the corresponding model parameter file of each classification;
(2) garment image that selection user uploads, carries out detection of attribute using the different classes of model of above-mentioned training respectively,
Count the corresponding frequency of each attribute in each classification;
(3) garment inventory picture to be recommended is selected, using the different classes of model of above-mentioned training respectively to every picture
Detection of attribute is carried out, the attribute that each classification is possessed in every picture to be recommended is obtained;
(4) attributes similarity between every picture to be recommended and the picture of user's upload is calculated, and presses similarity size
Every picture to be recommended is ranked up, according to the forward picture to be recommended of recommended amount selected and sorted.
In step (1), by whole Attribute transpositions at eight classifications, including collar design collar, the design of neck line
The long design skirt of neckline, skirt, sleeve length design sleeve, neck design neck, the long design coat of clothing, lapel design lapel
And trousers length designs pant.
In step (2), the garment image that user uploads is the clothes figure that garment image person, the user in user's wardrobe like
The combination of piece or both.
In order to improve the efficiency of detection of attribute, when carrying out detection of attribute, defines a predict function while calling complete
The other network model of category realizes multitask output.
When carrying out detection of attribute, the determined property to shield portions in picture is " being not present ", and is counting each classification
In deleted before the corresponding frequency of each attribute.
In step (4), the calculation formula of the attributes similarity are as follows:
Si=Si_1+Si_2+……+Si_n
Wherein, SiFor the attributes similarity between i-th picture to be recommended and user's picture, Si_nFor i-th figure to be recommended
Local similarity in piece and user's picture between n-th of attribute, if there is no in user's picture n-th for i-th picture to be recommended
A attribute, then Si_n=0, if it is present Si_nCalculation formula are as follows:
Si_n=| rn+1|2
Wherein, rnIt is the frequency of n-th of attribute in user's picture.
In step (4), every picture to be recommended is ranked up by similarity size using bubble sort.
Compared with prior art, the invention has the following advantages:
1, picture attribute label of the invention has more pinpoint accuracy on description user preference, and more vivid understanding is used
Family hobby.
2, advanced features are added in clothes recommendation in the present invention --- and clothes attribute can be recommended higher with user's compatible degree
Clothes, have better recommendation effect.
Detailed description of the invention
Fig. 1 is a kind of flow diagram of the clothes recommended method based on dress ornament attributes match of the present invention;
Fig. 2 is the clothes attribute of the method for the present invention detection;
Fig. 3 is more detection of attribute result schematic diagrams of single picture.
Specific embodiment
The invention will be described in further detail with reference to the accompanying drawings and examples, it should be pointed out that reality as described below
It applies example to be intended to convenient for the understanding of the present invention, and does not play any restriction effect to it.
As shown in Figure 1, a kind of clothes recommended method based on dress ornament attributes match, comprising the following steps:
S01 establishes clothes detection of attribute model
(1.1) prepare clothes attribute data collection, carry out clothes detection of attribute and mainly Alibaba is used to exist
The data set provided in 2018FashionAI global challenge match, mainly having 8 classifications, (collar designs collar, the design of neck line
The long design skirt of neckline, skirt, sleeve length design sleeve, neck design neck, the long design coat of clothing, lapel design
Lapel, trousers length design pant), 54 attributes.Next, 8 classifications are separately trained as single task.Clothes attribute is such as
Shown in Fig. 2.
(1.2) data preparation is carried out first: a train_valid is created at data, as data after all arrangements
Catalogue, read path and the label of every picture, wherein label is the character string of several n and y composition, and alphabetical y goes out
Existing position is exactly the corresponding type of picture, and it is right under data/train_valid catalogue that this picture is put into according to its label
In the category list answered, with the ratio cut partition training set of 9:1 and verifying collection.
(1.3) transfer learning is borrowed, has been trained on ImageNet data set in advance using the importing of the library MXNet
The model output of resnet50_v2 model, the training on ImageNet is 1000 dimensions, it would be desirable to according to the label of every one kind
Character string number adjusts output dimension accordingly, and to the weight random initializtion of output layer, redefines new resnet50_
v2。
(1.4) some training parameters (learning rate, epoch) and the picture augmentation for training set and verifying collection are defined
Function carries out random cropping and overturning to picture.
(1.5) pass through the data preparation in 1.2, the picture in train data set passes through in the library MXNet
Gluon.data.DataLoader interface is read in.Use cross entropy as loss function.
(1.6) picture is trained using the resnet50_v2 model modified, passes through the save_ in the library MXNet
Parameters preservation model Parameter File under model file directory uses after convenient.
(1.7) all to eight classifications (collar, neckline, skirt, sleeve, neck, coat, lapel, pant)
It repeats the above steps, the model parameter file of 8 detection respective classes attributes can be obtained.
S02, user property forming label
(2.1) model parameter file is imported by load_parameters, obtains the network of eight category attributes of detection
Model.
(2.2) since purpose is desirable to that all properties testing result of a picture can be obtained simultaneously, so defining one
A predict function calls 8 network models simultaneously, realizes multitask output.
(2.3) picture for reading user file, be in user file user store clothes picture in oneself wardrobe or
It is oneself satisfied garment image.Detection of attribute, one figure of every detection are carried out to user's picture using predict function
Piece, the attribute results that will test are saved in list Attribute_list.More detection of attribute result such as Fig. 3 of single picture
It is shown.
(2.4) since certain pictures presence of data set is blocked or invisible problem, cause model less can accurate judgement
Picture classification will be judged as " Invisible (being not present) " that this attribute is for formulating attribute mark by Picture section attribute
There are errors for label, so needing to delete " Invisible " in list Attribute_list.
(2.5) attribute number sorts: after having traversed the picture of all user's storages in user file, Attribute_
The attribute information of stored all pictures in list, using counter come to the attribute in list Attribute_list into
Row counts, and then arranges it from high to low by frequency of occurrence and saves in list Attribute_count.
(2.6) Attributes Frequency calculates: the single attribute frequency of occurrence in Attribute_count is gone out divided by all properties
Occurrence number obtains single attribute aiFrequency ri, by attribute aiFrequency r corresponding with itiSuccessively it is saved in list Attribute_
User is embodied in prob, in Attribute_prob to the preference degree of each attribute.It indicates are as follows:
[[attribute a1, frequency r1], [attribute a2, frequency r2], [attribute a3, frequency r3] ... ...]
(2.7) using dataframe.to_csv () method in the library pandas by the content of list Attribute_prob
Write-in csv file is saved under current directory, is called in attributes similarity matcher after convenient.
S03, shop clothes detection of attribute
(3.1) garment image in the file of shop is read, what is saved in the file of shop is that clothes shop will recommend
The garment inventory picture of user carries out detection of attribute to each picture using predict function, by recycling each attribute
It is saved in list 1s with Attributes Frequency (being defaulted as 1), such as [attribute, 1].
(3.2) 1s is successively then nested into the list pred_ containing picture name after the completion of picture detection
In result, the attribute tags list of single picture, the form of expression are formed are as follows:
[picture name, [attribute 1,1], [attribute 2,1], [attribute 3,1] ... ...]
(3.3) pred_result is nested into list Shop_Attribute again, it is each in Shop_Attribute
A element is all the attribute tags list of a picture, contains the picture pathname of every picture, picture attribute and attribute
Frequency.The specific format of list Shop_Attribute can indicate are as follows:
[[picture p1, [attribute p1_1, 1], [attribute p1_2, 1], [attribute p1_3, 1] ... ...],
[picture p2, [attribute p2_1, 1], [attribute p2_2, 1], [attribute p2_3, 1] ... ...],
[picture pn, [attribute pn_1, 1], [attribute pn_2, 1], [attribute pn_3, 1] ... ...]]
(3.4) using dataframe.to_csv () method in the library pandas by the content of list Shop_Attribute
Write-in csv file is saved under current directory, is called in attributes similarity matching after convenient.
S04, user-shop clothes attributes similarity matching
(4.1) csv file related with user and shop is read, the content of the inside is stored in list user and shop
In, list user is indicated are as follows:
[[attribute a1, frequency r1], [attribute a2, frequency r2], [attribute a3, frequency r3] ... ...]
List shop (explanation after for convenience, with the citing of the first picture):
[picture p1, [attribute p1_1, 1], [attribute p1_2, 1], [attribute p1_3, 1] ... ...]
(4.2) by looping through their two identical or different attributes of two list lookups, user-shop figure is calculated
Similarity (Similarity degree) between piece, is denoted as Si。
(4.3) if list user and picture p1Between attribute it is completely dissimilar, i.e. user and p1Between do not have be overlapped category
Property, then user user and picture p1Between similarity be zero, S1=0.Then in picture p1List last bit addition element S1:
[picture p1, [attribute p1_1, 1], [attribute p1_2, 1], [attribute p1_3, 1] ... ..., S1]
(4.4) if the attribute of user includes picture p1All properties, i.e. picture p1Property set be user property set
Subset, then the attribute (shared attribute) existed simultaneously in two property sets is calculated.
Such as: if there are attribute a in user1: V neckline, in picture p1In there are same alike result p1_1: V
Neckline, then a1(p1_1) it is exactly user and picture p1Shared attribute, similarity (and the user and picture p of the attribute1's
Local similarity) it calculates are as follows: S1_1=| r1+1|2,
(4.5) it then sums to the similarity of all shared attributes, to ask user and picture p1The overall situation it is similar
Degree: S1=S1_1+S1_3+ ..., then in picture p1List last bit addition element S1。
(4.6) if the attribute of user and picture p1Attribute exist intersection, i.e. picture p1Part attribute be institute in user
No, then ignore this part attribute, for picture p1The step of remaining attribute repeats above-mentioned (4.4), (4.5).
(4.7) bubble sort is used, sequence from big to small is carried out by comparing the similarity S of every picture, exports Si
Ranking is the picture of first five.
Technical solution of the present invention and beneficial effect is described in detail in embodiment described above, it should be understood that
Above is only a specific embodiment of the present invention, it is not intended to restrict the invention, it is all to be done in spirit of the invention
Any modification, supplementary, and equivalent replacement, should all be included in the protection scope of the present invention.
Claims (7)
1. a kind of clothes recommended method based on dress ornament attributes match, which comprises the following steps:
(1) clothes detection of attribute model is constructed, clothes attribute data collection is selected, by whole Attribute transposition to several classifications
In, category is respectively trained clothes detection of attribute model, obtains the corresponding model parameter file of each classification;
(2) garment image that selection user uploads, carries out detection of attribute using the different classes of model of above-mentioned training respectively, counts
The corresponding frequency of each attribute in each classification;
(3) garment inventory picture to be recommended is selected, every picture is carried out respectively using the different classes of model of above-mentioned training
Detection of attribute obtains the attribute that each classification is possessed in every picture to be recommended;
(4) attributes similarity between every picture to be recommended and the picture of user's upload is calculated, and by similarity size to every
It opens picture to be recommended to be ranked up, according to the forward picture to be recommended of recommended amount selected and sorted.
2. the clothes recommended method according to claim 1 based on dress ornament attributes match, which is characterized in that in step (1),
By whole Attribute transpositions at eight classifications, including collar design, the design of neck line, the long design of skirt, sleeve length design, neck design, clothing
Long design, lapel design and trousers length design.
3. the clothes recommended method according to claim 1 based on dress ornament attributes match, which is characterized in that in step (2),
The garment image that user uploads is the combination of clothes picture that garment image person, the user in user's wardrobe like or both.
4. the clothes recommended method according to claim 1 based on dress ornament attributes match, which is characterized in that in step (2),
It when carrying out detection of attribute, defines a predict function while calling the network model of whole classifications, realize multitask output.
5. the clothes recommended method according to claim 1 based on dress ornament attributes match, which is characterized in that in step (2),
When carrying out detection of attribute, the determined property to shield portions in picture is " being not present ", and each category in each classification of statistics
Property corresponding frequency before deleted.
6. the clothes recommended method according to claim 1 based on dress ornament attributes match, which is characterized in that in step (4),
The calculation formula of the attributes similarity are as follows:
Si=Si_1+Si_2+……+Si_n
Wherein, SiFor the attributes similarity between i-th picture to be recommended and user's picture, Si_nFor i-th picture to be recommended with
Local similarity in user's picture between n-th of attribute, if there is no n-th of categories in user's picture for i-th picture to be recommended
Property, then Si_n=0, if it is present Si_nCalculation formula are as follows:
Si_n=| rn+1|2
Wherein, rnIt is the frequency of n-th of attribute in user's picture.
7. the clothes recommended method according to claim 5 based on dress ornament attributes match, which is characterized in that in step (4),
Every picture to be recommended is ranked up by similarity size using bubble sort.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910521390.XA CN110413874A (en) | 2019-06-17 | 2019-06-17 | A kind of clothes recommended method based on dress ornament attributes match |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910521390.XA CN110413874A (en) | 2019-06-17 | 2019-06-17 | A kind of clothes recommended method based on dress ornament attributes match |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110413874A true CN110413874A (en) | 2019-11-05 |
Family
ID=68359184
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910521390.XA Pending CN110413874A (en) | 2019-06-17 | 2019-06-17 | A kind of clothes recommended method based on dress ornament attributes match |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110413874A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111414949A (en) * | 2020-03-13 | 2020-07-14 | 杭州海康威视系统技术有限公司 | Picture clustering method and device, electronic equipment and storage medium |
CN112819533A (en) * | 2021-01-29 | 2021-05-18 | 深圳脉腾科技有限公司 | Information pushing method and device, electronic equipment and storage medium |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140025533A1 (en) * | 2012-07-20 | 2014-01-23 | Alibaba Group Holding Limited | Method and Apparatus of Recommending Clothing Products |
CN106156297A (en) * | 2016-06-29 | 2016-11-23 | 北京小米移动软件有限公司 | Method and device recommended by dress ornament |
CN108052952A (en) * | 2017-12-19 | 2018-05-18 | 中山大学 | A kind of the clothes similarity determination method and its system of feature based extraction |
CN109685121A (en) * | 2018-12-11 | 2019-04-26 | 中国科学院苏州纳米技术与纳米仿生研究所 | Training method, image search method, the computer equipment of image encrypting algorithm |
-
2019
- 2019-06-17 CN CN201910521390.XA patent/CN110413874A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140025533A1 (en) * | 2012-07-20 | 2014-01-23 | Alibaba Group Holding Limited | Method and Apparatus of Recommending Clothing Products |
CN106156297A (en) * | 2016-06-29 | 2016-11-23 | 北京小米移动软件有限公司 | Method and device recommended by dress ornament |
CN108052952A (en) * | 2017-12-19 | 2018-05-18 | 中山大学 | A kind of the clothes similarity determination method and its system of feature based extraction |
CN109685121A (en) * | 2018-12-11 | 2019-04-26 | 中国科学院苏州纳米技术与纳米仿生研究所 | Training method, image search method, the computer equipment of image encrypting algorithm |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111414949A (en) * | 2020-03-13 | 2020-07-14 | 杭州海康威视系统技术有限公司 | Picture clustering method and device, electronic equipment and storage medium |
CN111414949B (en) * | 2020-03-13 | 2023-06-27 | 杭州海康威视系统技术有限公司 | Picture clustering method and device, electronic equipment and storage medium |
CN112819533A (en) * | 2021-01-29 | 2021-05-18 | 深圳脉腾科技有限公司 | Information pushing method and device, electronic equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102760163B (en) | Personalized recommendation method and device of characteristic information | |
US20220138831A1 (en) | Method of Providing Fashion Item Recommendation Service Using User's Body Type and Purchase History | |
US20150095184A1 (en) | Recommending a personalized ensemble | |
US8533225B2 (en) | Representing and processing inter-slot constraints on component selection for dynamic ads | |
US20130041781A1 (en) | Digital personal market and shopping network | |
CN106055710A (en) | Video-based commodity recommendation method and device | |
CN109542916A (en) | Platform commodity enter method, apparatus, computer equipment and storage medium | |
CN106294420B (en) | The method and device of business object collocation information is provided | |
KR101677128B1 (en) | Cosmetics information mobile application using cosmetic products component information | |
CN103336796A (en) | Method and system for displaying door buster directly | |
CN101454771A (en) | System and method of segmenting and tagging entities based on profile matching using a multi-media survey | |
CN106910108A (en) | A kind of items list methods of exhibiting and system | |
KR20200045668A (en) | Method, apparatus and computer program for style recommendation | |
CN109190030A (en) | Merge the implicit feedback recommended method of node2vec and deep neural network | |
CN107194758A (en) | Evaluation recording method based on credible friend | |
CN107832338A (en) | A kind of method and system for identifying core product word | |
CN110413874A (en) | A kind of clothes recommended method based on dress ornament attributes match | |
Furukawa et al. | Visualisation for analysing evolutionary dynamics of fashion trends | |
KR20220039697A (en) | Method, apparatus and computer program for style recommendation | |
Gong et al. | Aesthetics, personalization and recommendation: A survey on deep learning in fashion | |
KR20220019737A (en) | Method, apparatus and computer program for fashion item recommendation | |
CN110807691A (en) | Cross-commodity-class commodity recommendation method and device | |
KR20210131198A (en) | Method, apparatus and computer program for advertising recommended product | |
CN110035111A (en) | A kind of method for pushing and device dispensing entity | |
KR102378072B1 (en) | Method, apparatus and computer program for style recommendation |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191105 |
|
RJ01 | Rejection of invention patent application after publication |