CN110348529A - A kind of intelligent clothes Trend of fashion prediction technique and system - Google Patents
A kind of intelligent clothes Trend of fashion prediction technique and system Download PDFInfo
- Publication number
- CN110348529A CN110348529A CN201910642697.5A CN201910642697A CN110348529A CN 110348529 A CN110348529 A CN 110348529A CN 201910642697 A CN201910642697 A CN 201910642697A CN 110348529 A CN110348529 A CN 110348529A
- Authority
- CN
- China
- Prior art keywords
- clothes
- star
- style
- video
- dress
- 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.)
- Granted
Links
Classifications
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
Abstract
The present invention provides a kind of intelligent clothes Trend of fashion prediction technique and system, and main includes the TV and film variety show for obtaining current hot broadcast；By the list of preset trend video display star and leading role, the camera lens segment for having star is obtained；The clothes worn according to star；It obtains and merges the most complete frame of clothes in video, clothes is extracted；According to garment length, color, style, quality, relief features, style and features extraction is carried out to the clothes that the star wears, unusual or exotic dress is removed, is clustered according to clothes style and features；According to the multiple classifications formed after cluster, classification sequence is carried out；It is sorted according to star's clothes style popularity, for similar style and design, control and prediction can be carried out to the amount of stocking up of the following dress ornament style and clothes by carrying out dress designing or the prediction of the material amount of the stocking up present invention, and designer is assisted to carry out clothes design.
Description
Technical field
The present invention relates to computer application technology more particularly to a kind of intelligent clothes Trend of fashion prediction technique with
System.
Background technique
In clothes design process, necessary not only for the unique design of designer, also to meet the direction of tide of society.And
The clothes halfpace in addition to comparing minority of trend is led, there are also the styles of wearing the clothes of public star.In contrast, contemporary TV dramas,
The audients such as film are more, therefore are easier to form worship of idols and follow the wind.Current Taobao's dress advertisements etc., also often
It can be appreciated that and be turned on posters such as ' clothes are the same money of Liu Shishi '.It can be seen that the dressing of film star is to clothes trend
Trend has large effect power.But, common performer and it is unable to reach the effect of this idolatry, therefore their dress
Clothes can not be used for the prediction of Trend of fashion.
The wearing the clothes of the telecine leading role of fashion theme style of arranging in pairs or groups will affect trend, pass through how much amount TV and video network
On standing, the clothes of hero or video display star are worn, and predict subsequent clothes Trend of fashion, are done for production and the market demand
Prepare.
Summary of the invention
The present invention provides a kind of intelligent clothes Trend of fashion prediction techniques, mainly comprise the steps that
Obtain the TV and film variety show of current hot broadcast；Pass through the list of preset trend video display star and leading role；It obtains
Star's face, and recognition of face is carried out to the personage in program, obtain star's camera lens；
According to star's camera lens, the clothes of their dresses is obtained；According to clothes, multiple video clips are merged, obtains and merges
The most complete frame of clothes, extracts clothes in video；
According to garment length, color, style, quality, relief features, style and features extraction is carried out to the clothes that star wears, according to
Clothes style and features are clustered.
According to the multiple classifications formed after cluster, classification sequence is carried out；Wherein, the classification sequence refers to that each is poly-
A kind of clothes of style of category stands after class；The quantity of each classification is calculated, number of computations specific gravity carries out classification
Sequence obtains the sequence of star's clothes style popularity.
It is sorted according to star's clothes style popularity, for similar style and design, carries out dress designing.
Still optionally further, in method as described above, the TV and film variety show for obtaining current hot broadcast；It is logical
The list for crossing preset trend video display star and leading role, specifically includes that
The major video display ranking list of current network is obtained, it is for statistical analysis to ranking list, obtain the TV and film of current hot broadcast
Program；
By obtaining Baidu search roll of the hour interface, current actors ranking list is obtained, top n for obtaining foremost are bright
Star；
Still optionally further, in method as described above, acquisition star's face, and face knowledge is carried out to the personage in program
Not, star's camera lens is obtained, according to star's camera lens, the clothes of their dresses is obtained, specifically includes that
Recognition of face is carried out to the personage in video, obtains the video clip with the star,
It is worn clothes according to personage in video, segment is merged.
The segment merging refers to identify in multiple video clips, whether the star wears same dress, if so,
Then multiple video clips are merged.
Still optionally further, in method as described above, in the multiple video clips of identification, it is same whether star wears
Part clothes, specifically includes that
By image recognition technology, the clothes of star's dress is extracted in the first video and the second video, using opencv or
Matlab extract video frame in clothes color characteristic, gray feature, textural characteristics, for three kinds of features feature to
Amount；The similarity of feature vector is calculated, if the similarity of two videos is greater than certain threshold value, assert the first video and second
Video dress is same dress.
Still optionally further, in method as described above, the acquisition merges the most complete frame of clothes in video, to clothes
It extracts, specifically includes that
For the video after merging, the frame that star's whole body can be taken in video is extracted,
If the Full Body Video frame of star can not be extracted, the maximum frame of clothes area that can be observed is extracted；
The clothes worn with star is extracted, according to the clothes for taking multiple angles, obtains clothes in the feature of all angles,
Including length, color, style, quality, flower pattern.
Still optionally further, in method as described above, it is described according to each span access location length of clothes, color, structure, quality,
Flower pattern feature carries out style and features extraction to the clothes that star wears, specifically includes that
Clothes in video is extracted, the ratio between the length and each component at each position of clothes is calculated；
Clothes color is extracted, rgb value is converted into；
Clothes quality is analyzed, including chill mark quantity is extracted by openvc software, judgement is soft type or endures
Refreshing type；By extracting the reflective degree of garment surface, judgement is gloss type or dim type；
Flower picture recognition is carried out according to the flower on clothes, obtains flower pattern size, color, position feature；
Still optionally further, described to be clustered according to clothes style and features in method as described above, it specifically includes that
According to length, color, structure, quality, flower pattern feature, feature extraction is carried out, construction feature matrix is used according to feature
The bitch clustering technique of scikit-learn software carries out style cluster, the clothes with similar style is aggregated to together.
Still optionally further, described according to the multiple classifications formed after cluster in method as described above, carry out classification row
Sequence specifically includes that
According to Baidupedia to the classification method of style, classify to style, it is main include elegant series, stamp series, when
Still series, evening clothing series, leisure series, Sports Series；
Classified according to style, obtain the online clothes sample of each series, trained style disaggregated model, to need to carry out clothes into
Row classification；
It for the clothes classification that each has been classified, then is once clustered, obtains the popularity sequence under the class categories.
Still optionally further, described to be sorted according to star's clothes style popularity in method as described above, for
Similar style and design, carry out dress designing, specifically include that
It is sorted according to star's clothes style popularity, the prediction current year all types of demand specific gravity with style measure ratio according to demand
Weight carries out clothes production into purchase clothes raw material specific gravity；
It is sorted according to star's clothes style popularity, designer targetedly extracts most popular design style, in novel clothes
Fashion elements are added.
The invention discloses it is a kind of based on video display star dress clothes Trend of fashion forecasting system, the system comprises:
Movie and video programs and star obtain module, for obtaining movie and video programs and the video display star of most fashion；
Star's clothes obtain module, for identification star and identify star dress clothes；
Garment features extraction module, for extracting the garment features of star；
Style popularity sorting module, the quantity for being worn according to star's clothes style, judges which kind of style is more worn by star
It obtains more；
Prediction and design module improve the design that material is stocked up with clothes for the Trend of fashion according to prediction；
Technical solution provided in an embodiment of the present invention can include the following benefits:
The present invention is by the style of wearing the clothes of acquisition video display star, and prediction is because of idol effect, the stream for the clothes style that may cause
Row, the present invention can exclude star and wear the influence of inaccuracy caused by unusual or exotic dress, predict following epidemic factor.It can be to factory
Play the role of prediction and assistance to the control of the following dress ornament style to the judgement and designer of material buying of wearing the clothes.
Detailed description of the invention
Fig. 1 is the flow chart of the clothes Trend of fashion prediction technique embodiment of intelligence of the invention；
Fig. 2 is the structure chart of the clothes Trend of fashion forecasting system embodiment of intelligence of the invention.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention clearer, right in the following with reference to the drawings and specific embodiments
The present invention is described in detail.
Fig. 1 is the flow chart of the clothes Trend of fashion prediction technique of the invention based on video leading role identification.Such as Fig. 1 institute
Show, the clothes Trend of fashion prediction technique based on video leading role identification of the present embodiment can specifically include following steps:
Step 101, modern video display film and hot star are obtained.
Obtain the major video display ranking list of current network, it is for statistical analysis to ranking list, obtain current hot broadcast TV and
Movie program；By film types interface, modern piece and city piece are obtained.Because only that the film and television of city sheet type or comprehensive
Skill program, the clothes of the inside can be just that current trend is corresponding, and modern there is no fear of dressing up inside study costume piece, i.e.,
Just learn, also not will form trend.
By obtaining Baidu search roll of the hour interface, current actors ranking list is obtained, the top n of foremost is obtained
A star；Star's temperature is higher, he/her becomes people's idol and the probability being imitated is higher.Moreover, this imitate is with the epoch
And change.Before 30 years, people can more imitate the stars such as Hong Kong, Macao and Taiwan star such as Liu Dehua, Guo Fucheng wear the clothes and
Hair off-centre parting hairstyle.And the modern times more imitate hot star, such as poplar is purple, the fashionable dress apperance of the modern star of tf-boy.No longer
Follow fashionable dress and the hair style etc. of old star.
Step 102, star's face is obtained, and recognition of face is carried out to the personage in program, star's camera lens is obtained, according to bright
Star camera lens obtains the clothes of their dresses.
Recognition of face is carried out to the personage in video, obtains the video clip with the star.Recognition of face is mature
Technical method, belong to well-known technique, as long as getting the human face photo of star on the net, so that it may in video, to star
Face carries out video tracking, obtains all video clips with star's camera lens.
It is worn clothes according to personage in video, segment is merged.
The merging of video clip refers to, identifies in multiple video clips, and whether detection star wears same dress, if
It is then to merge multiple video clips.Multiple video mergings can be in subsequent analysis, on the one hand by same part clothing
Clothes carry out aggregate analysis, and some star's clothes overall picture camera lenses that can show on the other hand has been avoided to be missed.
It identifies in multiple video clips, the method whether star wears same dress includes:
By image recognition technology, the clothes of star's dress is extracted in the first video and the second video, using opencv or
Matlab extracts the color characteristic of clothes in video frame, gray feature, textural characteristics these three feature vectors；Calculate feature vector
Similarity, if the similarity of two videos is greater than certain threshold value, assert that the first video and the second video wear is same
Dress.
Step 103, it obtains and merges the most complete frame of clothes in video, clothes is extracted.
For the video after merging, the frame that star's whole body can be taken in video is extracted,
If the Full Body Video frame of star can not be extracted, the maximum frame of clothes area that can be observed is extracted；This was searched for
Journey, by analyzing in each frame, the maximum area that clothes can be taken by camera lens screens camera lens.Because only that clapping
The angle taken the photograph is good, can completely take clothes, could preferably analyze the clothes with star.
It extracts video underpants and extracts the clothes worn with star, according to the clothes for taking multiple angles, obtain clothes
In the feature of all angles, including length, color, style, quality, flower pattern.
Step 104, according to each span access location length of clothes, color, structure, quality, flower pattern feature, the clothes that star wears is carried out
Style and features extract.
Video frame by the clothes taken can be calculated according to the height and weight that the star can obtain on the net
The length and ratio at each position of clothes of his/her dress are obtained out.Clothes color can be further analyzed according to video frame
And colour match, clothes quality, flower pattern etc..Specific method includes:
Calculate the ratio between the length and each component at each position of clothes；
Clothes color is extracted, rgb value is converted into；
Clothes quality is analyzed, including chill mark quantity is extracted by openvc software, judgement is soft type or endures
Refreshing type；By extracting the reflective degree of garment surface, judgement is gloss type or dim type；
Flower picture recognition is carried out according to the flower on clothes, obtains flower pattern size, color, position feature；
These features all construct the minutia of clothes from all angles, can be very good to distinguish the style between different clothes
Feature.
Step 105, unusual or exotic dress is removed.
According to length, color, structure, quality, flower pattern feature, feature is extracted, there is the odd of notable difference to fill feature
Different clothes, are removed；Some variety shows, star singer etc., dress is peculiar, and this clothes is also only suitble to few groups of people,
It is hardly formed fashion, therefore the unusual or exotic dress of hot star dress, belongs to the noise in all clothes Trend of fashion predictions
Data.It should be removed.
The recognition methods of unusual or exotic dress is mainly the feature difference for judging current garment and conventional garments, such as conventional garments
99% sleeve is all the length scale for being suitble to arm width.And current clothes width is 30 back of normal arm width, then
Have exceeded normal range (NR).Or traditional clothes, collar are all there was only three or four centimetres in neck director's degree, and the collar of the star
Head is extended to, length there are 10 centimetres, then illustrates that clothes is too special, and such clothes should be excluded away.In short, being on duty
It is different be greater than certain threshold value when, be judged as unusual or exotic dress, it should be removed.
Step 106, classified to clothes according to style, then again for each sorted subdivision classification, then into
Row cluster.To multiple classifications that cluster is formed, classification sequence is carried out, this process specifically includes that
According to Baidupedia to the classification method of style, classify to style, it is main include elegant series, stamp series, when
Still series, evening clothing series, leisure series, Sports Series；
To be classified according to style, obtains the online clothes sample of each series, these samples contain the garment type of various systems,
By manually labelling, various types of other training data is obtained, these data, training style disaggregated model are then passed through；This point
Class model can be constructed using deep learning algorithm, by preparatory trained clothes genre classification model, to clothes into
Row classification, obtains the style for each garment type that star wears.
Then, classify first to the clothes of star；The reason of being classified, the clothes being between different styles,
Although being able to carry out the comparison of quantity, this comparison is not accurate enough, compares and is difficult to objectively show clothes money
Difference between formula one.And if first classified to style, such as the different evening clothings worn just for star carry out style comparison, just
It can be directed to the type of evening clothing series, the prevailing style is more likely to form and is predicted.If do not classified first, evening clothing is with fortune
Dynamic clothes are compared, and it is not accurate enough to will lead to comparison result.
Subsequently, it for the clothes classification that each has been classified, then is once clustered, obtains the stream under the class categories
The sequence of row degree.The feature of cluster may be used also in addition to the length, color, structure, quality, the flower pattern that have obtained above, same reason
It include collar, sleeve, trouser legs, pocket, button, waistband, skirt angle etc., more careful features to be added.It can make to cluster in this way
The difference of sample is bigger, more has method to carry out accurately clustering.
Feature extraction, construction feature matrix, using scikit-learn software are carried out to the above-mentioned various features mentioned
Bitch clustering technique carries out style cluster, the clothes with similar style is aggregated to together.And according to each after polymerization
The quantity of a classification, is ranked up.It obtains star and wears most styles, they would be possible to become most popular clothes wind
Lattice.
The specific embodiment of this process, for example, after system is classified automatically, all stars wear in video evening clothing
Classification, total has 1000 evening clothing clothes, then needing once to cluster the clothes of this 1000 evening clothings again, makes late
Dress style also obtains the polymerization of further stylistic differences.Such as 8 classifications, respectively 533 have been aggregated into, and 234,109
Part, 65,24,23,9,3, then just can recognize that, first maximum classification has 533 styles, is
The most probable popular classification of current year evening clothing style.
Step 107, it is sorted according to star's clothes style popularity, for similar style and design, carries out dress designing, or
Person's material is stocked up planning.
It is sorted according to star's clothes style popularity, the prediction current year all types of demand specific gravity with style, according to demand
Specific gravity is measured, into purchase clothes raw material specific gravity, carries out clothes production；
It is sorted according to star's clothes style popularity, designer targetedly extracts most popular design style, in novel clothes
Fashion elements are added.
Through the above description of the embodiments, those skilled in the art can be understood that above-described embodiment can
The mode of necessary general hardware platform can also be added to realize by software by software realization.Based on this understanding,
The technical solution of above-described embodiment can be embodied in the form of software products, which can store non-easy at one
In the property lost storage medium (can be CD-ROM, USB flash disk, mobile hard disk etc.), including some instructions are with so that a computer is set
Standby (can be personal computer, server or the network equipment etc.) executes method described in each embodiment of the present invention.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Within mind and principle, any modification, equivalent substitution, improvement and etc. done be should be included within the scope of the present invention.
Claims (10)
1. a kind of intelligent clothes Trend of fashion prediction technique, which is characterized in that the described method includes:
Obtain the TV and film variety show of current hot broadcast；By the list of preset trend video display star and leading role, obtain
Star's face；And recognition of face is carried out to the personage in program, obtain the camera lens segment for having star；
According to star's camera lens segment, the clothes of star's dress is obtained；According to identical clothes, multiple video clips are carried out
Merge, obtains and merge the most complete frame of clothes in video, clothes is extracted；
According to garment length, color, style, quality, relief features, style and features extraction is carried out to the clothes that the star wears,
Unusual or exotic dress is removed, is clustered according to clothes style and features；
According to the multiple classifications formed after cluster, classification sequence is carried out；Wherein, the classification sequence refers to, after each cluster
A kind of clothes of style of category stands；The quantity of each classification is calculated, number of computations specific gravity carries out classification row
Sequence obtains the sequence of star's clothes style popularity；
Sorted according to star's clothes style popularity, for similar style and design, carry out dress designing or material into
The prediction of goods amount.
2. according to the method described in claim 1, wherein, the TV and film variety show for obtaining current hot broadcast；Pass through
The list of preset trend video display star and leading role obtain star's face, specifically include that
The major video display ranking list of current network is obtained, it is for statistical analysis to ranking list, obtain the TV and film of current hot broadcast
Program；By video display style interface, the video display type of modern piece and city piece is obtained；
By obtaining Baidu search roll of the hour interface, current actors ranking list is obtained, top n for obtaining foremost are bright
Star；And obtain the face picture of star.
3. obtaining star's mirror according to the method described in claim 1, wherein, the personage in program carries out recognition of face
Head obtains the clothes of star's dress, specifically includes that according to star's camera lens
Recognition of face is carried out to the personage in video, obtains the video clip with the star,
It is worn clothes according to personage in video, if clothes is identical, segment is merged；
The segment merging refers to identify in multiple video clips, whether the star wears same dress, if so, will
Multiple video clips merge.
4. according to the method described in claim 3, wherein, in the multiple video clips of identification, whether star wears same part
Clothes specifically includes that
By image recognition technology, the clothes of star's dress is extracted in the first video and the second video, using opencv or
Matlab extract video frame in clothes color characteristic, gray feature, textural characteristics, for three kinds of features feature to
Amount；The similarity of feature vector is calculated, if the similarity of two videos is greater than certain threshold value, assert the first video and second
Video dress is same dress.
5. the acquisition merges the most complete frame of clothes in video according to the method described in claim 1, wherein, to clothes into
Row extracts, and specifically includes that
For the video after merging, the frame that star's whole body can be taken in video is extracted,
If the Full Body Video frame of star can not be extracted, the maximum frame of clothes area that can be observed is extracted；
The clothes worn with star is extracted, according to the clothes for taking multiple angles, obtains clothes in the feature of all angles,
Including length, color, style, quality, flower pattern.
6. described according to each span access location length of clothes, color, structure, quality, flower pattern according to the method described in claim 1, wherein
Feature carries out style and features extraction to the clothes that star wears, specifically includes that
Clothes in video is extracted, the ratio between the length and each component at each position of clothes is calculated；
Clothes color is extracted, rgb value is converted into；
Clothes quality is analyzed, including chill mark quantity is extracted by openvc software, judgement is soft type or endures
Refreshing type；By extracting the reflective degree of garment surface, judgement is gloss type or dim type；
Flower picture recognition is carried out according to the flower on clothes, obtains flower pattern size, color, position feature.
7. described to be removed to unusual or exotic dress according to the method described in claim 1, wherein, according to clothes style and features into
Row cluster, specifically includes that
According to length, color, structure, quality, flower pattern feature, feature is extracted, there is the unusual or exotic dress of notable difference for feature,
It is removed；
The recognition methods of the unusual or exotic dress includes the feature difference for judging current garment and conventional garments, when difference is greater than one
When determining threshold value, it is judged as unusual or exotic dress, it should be removed；
After removing unusual or exotic dress,
According to the extraction feature, construction feature matrix carries out wind using the bitch clustering technique of scikit-learn software
Lattice cluster, the clothes with similar style is aggregated to together.
8. it is described to carry out classification sequence according to the multiple classifications formed after cluster according to the method described in claim 1, wherein,
It specifically includes that
According to Baidupedia to the classification method of style, classify to style, it is main include elegant series, stamp series, when
Still series, evening clothing series, leisure series, Sports Series；
Classified according to style, obtain the online clothes sample of each series, trained style disaggregated model, to need to carry out clothes into
Row classification；
It for the clothes classification that each has been classified, then is once clustered, obtains the popularity sequence under the class categories.
9. it is described to be sorted according to star's clothes style popularity according to the method described in claim 1, wherein, for class
Like style and design, dress designing or the prediction of the material amount of stocking up are carried out, is specifically included that
It is sorted according to star's clothes style popularity, the prediction current year all types of demand specific gravity with style measure ratio according to demand
Weight carries out clothes production into purchase clothes raw material specific gravity；
It is sorted according to star's clothes style popularity, designer targetedly extracts most popular design style, in novel clothes
Fashion elements are added.
10. a kind of clothes Trend of fashion forecasting system based on video leading role identification, which is characterized in that the system comprises:
Movie and video programs and star obtain module, for obtaining movie and video programs and the video display star of most fashion；
Star's clothes obtain module, for identification star and identify star dress clothes；
Garment features extraction module, for extracting the garment features of star；
Style popularity sorting module, the quantity for being worn according to star's clothes style, judges which kind of style is more worn by star
It obtains more；
Prediction and design module improve the design that material is stocked up with clothes for the Trend of fashion according to prediction.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910642697.5A CN110348529B (en) | 2019-07-16 | 2019-07-16 | Intelligent clothes fashion style prediction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910642697.5A CN110348529B (en) | 2019-07-16 | 2019-07-16 | Intelligent clothes fashion style prediction method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110348529A true CN110348529A (en) | 2019-10-18 |
| CN110348529B CN110348529B (en) | 2021-10-22 |
Family
ID=68174864
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910642697.5A Active CN110348529B (en) | 2019-07-16 | 2019-07-16 | Intelligent clothes fashion style prediction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110348529B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112465567A (en) * | 2020-12-14 | 2021-03-09 | 武汉纺织大学 | Clothing style fashion prediction system and method |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011152844A1 (en) * | 2010-06-01 | 2011-12-08 | Hewlett-Packard Development Company, L.P. | Image clustering using a personal clothing model |
| CN102381074A (en) * | 2011-06-18 | 2012-03-21 | 江南大学 | Method used for knitted dress design, structural design and popular color forecasting |
| CN103530652A (en) * | 2013-10-23 | 2014-01-22 | 北京中视广信科技有限公司 | Face clustering based video categorization method and retrieval method as well as systems thereof |
| CN104967885A (en) * | 2015-03-27 | 2015-10-07 | 哈尔滨工业大学深圳研究生院 | Advertisement recommending method and system based on video content |
| CN105913275A (en) * | 2016-03-25 | 2016-08-31 | 哈尔滨工业大学深圳研究生院 | Clothes advertisement putting method and system based on video leading role identification |
| US20160314567A1 (en) * | 2015-03-26 | 2016-10-27 | The Board Of Regents Of The University Of Texas System | Systems and methods for image/video recoloring, color standardization, and multimedia analytics |
| CN106383877A (en) * | 2016-09-12 | 2017-02-08 | 电子科技大学 | On-line short text clustering and topic detection method of social media |
| CN107688985A (en) * | 2017-08-23 | 2018-02-13 | 武汉恩祁瑞网络有限公司 | A kind of garment coordination commending system based on style element |
| CN107862241A (en) * | 2017-06-06 | 2018-03-30 | 哈尔滨工业大学深圳研究生院 | A kind of clothes fashion method for digging and visually-perceptible system based on star's identification |
| CN108960499A (en) * | 2018-06-27 | 2018-12-07 | 东华大学 | A kind of Fashion trend predicting system merging vision and non-vision feature |
-
2019
- 2019-07-16 CN CN201910642697.5A patent/CN110348529B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011152844A1 (en) * | 2010-06-01 | 2011-12-08 | Hewlett-Packard Development Company, L.P. | Image clustering using a personal clothing model |
| CN102381074A (en) * | 2011-06-18 | 2012-03-21 | 江南大学 | Method used for knitted dress design, structural design and popular color forecasting |
| CN103530652A (en) * | 2013-10-23 | 2014-01-22 | 北京中视广信科技有限公司 | Face clustering based video categorization method and retrieval method as well as systems thereof |
| US20160314567A1 (en) * | 2015-03-26 | 2016-10-27 | The Board Of Regents Of The University Of Texas System | Systems and methods for image/video recoloring, color standardization, and multimedia analytics |
| CN104967885A (en) * | 2015-03-27 | 2015-10-07 | 哈尔滨工业大学深圳研究生院 | Advertisement recommending method and system based on video content |
| CN105913275A (en) * | 2016-03-25 | 2016-08-31 | 哈尔滨工业大学深圳研究生院 | Clothes advertisement putting method and system based on video leading role identification |
| CN106383877A (en) * | 2016-09-12 | 2017-02-08 | 电子科技大学 | On-line short text clustering and topic detection method of social media |
| CN107862241A (en) * | 2017-06-06 | 2018-03-30 | 哈尔滨工业大学深圳研究生院 | A kind of clothes fashion method for digging and visually-perceptible system based on star's identification |
| CN107688985A (en) * | 2017-08-23 | 2018-02-13 | 武汉恩祁瑞网络有限公司 | A kind of garment coordination commending system based on style element |
| CN108960499A (en) * | 2018-06-27 | 2018-12-07 | 东华大学 | A kind of Fashion trend predicting system merging vision and non-vision feature |
Non-Patent Citations (6)
| Title |
|---|
| HAIJUN ZHANG 等: "Sitcom-star-based clothing retrieval for video advertising: a deep learning framework", 《NEURAL COMPUTING AND APPLICATIONS》 * |
| HAIJUN ZHANG 等: "Sitcom-Stars Oriented Video Advertising via Clothing Retrieval", 《DASFAA 2018：DATABASE SYSTEM FOR ADVANCED APPLICATIONS》 * |
| ZIAD AL-HALAH 等: "Fashion Forward: Forecasting Visual Style in Fashion", 《ICCV 2017》 * |
| 孟志青 等: "时态文本数据流特征流行趋势模型及算法", 《计算机科学》 * |
| 安玉松: "基于视频主角识别的服装广告投放方法研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 项杨雪 等: "服装流行趋势预测网络技术演进路径: 基于科学计量学的分析", 《图书情报工作》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112465567A (en) * | 2020-12-14 | 2021-03-09 | 武汉纺织大学 | Clothing style fashion prediction system and method |
| CN112465567B (en) * | 2020-12-14 | 2022-10-04 | 武汉纺织大学 | Clothing style fashion prediction system and method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110348529B (en) | 2021-10-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liu et al. | Hi, magic closet, tell me what to wear! | |
| CN107273876B (en) | A kind of micro- expression automatic identifying method of ' the macro micro- transformation model of to ' based on deep learning | |
| CN106294489B (en) | Content recommendation method, device and system | |
| Han et al. | Heterogeneous face attribute estimation: A deep multi-task learning approach | |
| Hidayati et al. | Learning and recognition of clothing genres from full-body images | |
| Bialkowski et al. | A database for person re-identification in multi-camera surveillance networks | |
| CN105913275A (en) | Clothes advertisement putting method and system based on video leading role identification | |
| CN106055893B (en) | Garment coordination scheme generation method based on fashion template library and Auto-matching | |
| CN108510000A (en) | The detection and recognition methods of pedestrian's fine granularity attribute under complex scene | |
| EP2149106A1 (en) | Composite person model from image collection | |
| CN109635680A (en) | Multitask attribute recognition approach, device, electronic equipment and storage medium | |
| CN106203313B (en) | A kind of the clothes classification and recommended method of image content-based | |
| US20090091798A1 (en) | Apparel as event marker | |
| CN105260747B (en) | Clothing recognition methods based on clothing co-occurrence information and multi-task learning | |
| CN105210110A (en) | Makeup assistance device, makeup assistance system, makeup assistance method, and makeup assistance program | |
| Hall et al. | Fine-grained classification of pedestrians in video: Benchmark and state of the art | |
| CN109558810A (en) | Divided based on position and merges target person recognition methods | |
| CN108109055B (en) | Cross-scene clothing retrieval method based on image rendering | |
| Zhang et al. | An intelligent fitting room using multi-camera perception | |
| CN109934047A (en) | Face identification system and its face identification method based on deep learning | |
| CN106951448A (en) | A kind of personalization, which is worn, takes recommendation method and system | |
| CN108985873A (en) | Cosmetics recommended method, the recording medium for being stored with program, the computer program to realize it and cosmetics recommender system | |
| Rubio et al. | Multi-modal embedding for main product detection in fashion | |
| CN110348529A (en) | A kind of intelligent clothes Trend of fashion prediction technique and system | |
| Huang et al. | From street photos to fashion trends: leveraging user-provided Noisy labels for fashion understanding |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20210928 Address after: Building C, No.888, Huanhu West 2nd Road, Lingang New District, China (Shanghai) pilot Free Trade Zone, Pudong New Area, Shanghai Applicant after: Shanghai weiyexin Culture Technology Co.,Ltd. Address before: Room f101-12, No.1 incubation and production building, guanshao shuangchuang (equipment) center, Huake City, 42 Baiwang Avenue, Wujiang District, Shaoguan City, Guangdong Province, 512026 Applicant before: Shaoguan Qizhi Information Technology Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |