CN108664845A - The method and device screened image in fashion show video, identify model's types of garments - Google Patents
The method and device screened image in fashion show video, identify model's types of garments Download PDFInfo
- Publication number
- CN108664845A CN108664845A CN201710192736.7A CN201710192736A CN108664845A CN 108664845 A CN108664845 A CN 108664845A CN 201710192736 A CN201710192736 A CN 201710192736A CN 108664845 A CN108664845 A CN 108664845A
- Authority
- CN
- China
- Prior art keywords
- image
- model
- show video
- fashion show
- frame image
- 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
- 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
- G06V20/47—Detecting features for summarising video content
Abstract
The present invention provides a kind of method and devices screened image in fashion show video, identify model's types of garments, wherein the method for image includes in screening fashion show video：It obtains in fashion show video per frame image；Image characteristic analysis is carried out to every frame image, characteristics of image in every frame image is met to the image-erasing of the first pre-set image feature；In the remaining image per frame, the image for meeting the second pre-set image feature is filtered out.Above-mentioned technical proposal improves the accuracy for the image for meeting pre-set image feature in screening fashion show video.
Description
Technical field
The present invention relates to video identification technology field, more particularly to image, identification model in a kind of screening fashion show video
The method and device of types of garments.
Background technology
Currently, meeting the image of pre-set image feature in screening fashion show video, need for the figure in fashion show video
As being analyzed, which is a dynamic process, this needs to face more complicated scene image, in complicated scene graph
The accuracy rate that the image for meeting pre-set image feature in fashion show video is filtered out as in is low.
Further, it in fashion show video, is analyzed by the video and image intelligent of computer, it can be with automatic identification mould
Special habited style and type.The identification process is generally：First, model couple is extracted in complicated scene image
As then, then the style and type of the worn clothes of model being identified.Identification the worn clothes of model are all multiple at these at present
Model's object is extracted in miscellaneous scene image, the habited style of model institute and type that extract is identified, identification is accurate
Exactness is low.
Invention content
An embodiment of the present invention provides a kind of methods of image in screening fashion show video, are regarded to improve screening fashion show
Meet the accuracy of the image of pre-set image feature in frequency, this method includes：
It obtains in fashion show video per frame image；
Image characteristic analysis is carried out to every frame image, characteristics of image in every frame image is met into the first pre-set image feature
Image-erasing；
In the remaining image per frame, the image for meeting the second pre-set image feature is filtered out.
In one embodiment, the first pre-set image feature includes：Preset model's posture；
Image characteristic analysis is carried out to every frame image, characteristics of image in every frame image is met into the first pre-set image feature
Image-erasing, including：Image characteristic analysis is carried out to every frame image, it will be per model's posture in frame image and preset model's posture
Consistent image-erasing.
In one embodiment, the first pre-set image feature includes：
Preset model's clarity；
Image characteristic analysis is carried out per frame image to described, it is special that the first pre-set image will be met per characteristics of image in frame image
The image-erasing of sign, including：
Image characteristic analysis is carried out per frame image to described, it is clear that preset model will be less than per model's clarity in frame image
The image-erasing of clear degree；
And/or preset model's quantity；
Image characteristic analysis is carried out per frame image to described, it is special that the first pre-set image will be met per characteristics of image in frame image
The image-erasing of sign, including：
Image characteristic analysis is carried out per frame image to described, preset model's quantity will be more than per model's quantity in frame image
Image-erasing.
In one embodiment, the first pre-set image feature includes：
Preset camera angle range；
Image characteristic analysis is carried out per frame image to described, it is special that the first pre-set image will be met per characteristics of image in frame image
The image-erasing of sign, including：
Image characteristic analysis is carried out per frame image to described, it will be per camera angle in frame image at preset video camera angle
Spend the image-erasing in range；
And/or preset bright and dark light degree；
Image characteristic analysis is carried out per frame image to described, it is special that the first pre-set image will be met per characteristics of image in frame image
The image-erasing of sign, including：
Image characteristic analysis is carried out per frame image to described, it is bright that preset light will be less than per bright and dark light degree in frame image
The image-erasing of darkness.
In one embodiment, the first pre-set image feature includes：The quantity of personage and article in preset background；
Image characteristic analysis is carried out to every frame image, characteristics of image in every frame image is met into the first pre-set image feature
Image-erasing, including：
Image characteristic analysis is carried out to every frame image, the quantity of personage and article in background in every frame image is more than default
Background in the quantity of personage and article picture delete.
In one embodiment, the second pre-set image feature includes：Preset model's position range；
In the remaining image per frame, the image for meeting the second pre-set image feature is filtered out, including：
In the remaining image per frame, model's figure of the position in preset model's position range in the picture is filtered out
Picture；
And/or preset model's magnitude range；
In the remaining image per frame, the image for meeting the second pre-set image feature is filtered out, including：
In the remaining image per frame, model's figure of the size in preset model's magnitude range in the picture is filtered out
Picture.
In one embodiment, preset model's position range is that model is placed in the middle in every frame image；Preset model is big
Small range is：The size of model is more than 1/100 with the ratio per frame image size, and is less than 2/3.
The embodiment of the present invention additionally provides a kind of device screening image in fashion show video, to improve screening fashion show
Meet the accuracy of the image of pre-set image feature in video, which includes：
Acquisition module, for obtaining in fashion show video per frame image；
Image analysis module will meet for carrying out image characteristic analysis to every frame image per characteristics of image in frame image
The image-erasing of first pre-set image feature；
Picture recognition module, in the remaining image per frame, filtering out the image for meeting the second pre-set image feature.
With the image for meeting pre-set image feature in fashion show video is filtered out in complicated image in the prior art
Scheme compares, and the scheme of image is by obtaining in video per frame figure in screening fashion show video provided in an embodiment of the present invention
Picture；Image characteristic analysis is carried out to every frame image, characteristics of image in every frame image is met to the image of the first pre-set image feature
It deletes；It is remaining per frame image in, filter out the image for meeting the second pre-set image feature, to image in fashion show video into
It has gone simplification, has improved the accuracy for the image for meeting pre-set image feature in screening fashion show video.
The embodiment of the present invention additionally provides a kind of method identifying the worn types of garments of model in fashion show video, to carry
The accuracy of model worn types of garments and style, this method include in height identification fashion show video：
Using the method for image in above-mentioned screening fashion show video, filters out and meet the second pre-set image in fashion show video
The image of feature；
In the image for meeting the second pre-set image feature, the style and type of the worn clothes of model are identified.
The embodiment of the present invention additionally provides a kind of device identifying the worn types of garments of model in fashion show video, to carry
The accuracy of the worn types of garments of model, the device include in height identification fashion show video：
Optical sieving module, for using the device such as image in above-mentioned screening fashion show video, filtering out fashion show and regarding
Meet the image of the second pre-set image feature in frequency；
Types of garments identification module, in the image for meeting the second pre-set image feature, identifying the worn clothes of model
Style and type.
With in the prior art in complex scene identify the worn clothes of model style and type scheme compared with, this hair
The scheme of the worn types of garments of model, utilizes image in above-mentioned screening fashion show video in the identification video that bright embodiment provides
Scheme filters out the image for meeting the second pre-set image feature in fashion show video, and letter has been carried out to image in fashion show video
Change, has obtained the image of suitable clothes fashion type analysis；In the image for meeting the second pre-set image feature, model institute is identified
The style and type for wearing clothes improve the accuracy of types of garments and style identification.
Description of the drawings
Attached drawing described herein is used to provide further understanding of the present invention, and is constituted part of this application, not
Constitute limitation of the invention.In the accompanying drawings：
Fig. 1 is the flow diagram that the method for image in fashion show video is screened in the embodiment of the present invention；
Fig. 2 is the structural schematic diagram that the device of image in fashion show video is screened in the embodiment of the present invention；
Fig. 3 is the flow signal that the method for the worn types of garments of model in fashion show video is identified in the embodiment of the present invention
Figure；
Fig. 4 is the structural representation that the device of the worn types of garments of model in fashion show video is identified in the embodiment of the present invention
Figure.
Specific implementation mode
To make the objectives, technical solutions, and advantages of the present invention clearer, right with reference to embodiment and attached drawing
The present invention is described in further details.Here, the exemplary embodiment and its explanation of the present invention be for explaining the present invention, but simultaneously
It is not as a limitation of the invention.
In fashion show video, is analyzed, can be worn the clothes with automatic identification model by the video and image intelligent of computer
The style and type of clothes.This is a dynamic process, needs to be analyzed for the picture in video, extracts model first
Object, the style and type of the clothes then worn again to model are identified.This needs to face more complicated scene.At one section
In fashion show video, the picture of some segments is not appropriate for carrying out clothes automatic identification, for example, photographer is shooting model's
In the process, video camera can convert multiple angles and focal length, cause the complexity of identification.
In order to solve the above-mentioned technical problem, image, identification fashion show regard in a kind of screening fashion show video of present invention proposition
The scheme of the worn types of garments of model in frequency, the program are under clothing show environment, and to avoid wrong identification, it is suitable only to retain
Image removes unfavorable image, promotes the automatic distinguishing method for image of types of garments recognition accuracy.Below to the program into
Row is discussed in detail.
The scheme of image in screening fashion show video is introduced first.
Fig. 1 is the flow diagram that the method for image in fashion show video is screened in the embodiment of the present invention, as shown in Figure 1,
This method comprises the following steps：
Step 101：It obtains in fashion show video per frame image；
Step 102：Image characteristic analysis is carried out to every frame image, the first default figure will be met per characteristics of image in frame image
As the image-erasing of feature；
Step 103：In the remaining image per frame, the image for meeting the second pre-set image feature is filtered out.
With the image for meeting pre-set image feature in fashion show video is filtered out in complicated image in the prior art
Scheme compares, and the scheme of image is every in fashion show video by obtaining in screening fashion show video provided in an embodiment of the present invention
Frame image；Image characteristic analysis is carried out to every frame image, characteristics of image in every frame image is met into the first pre-set image feature
Image-erasing；In the remaining image per frame, the image for meeting the second pre-set image feature is filtered out, to scheming in fashion show video
As being simplified, the accuracy for the image for meeting pre-set image feature in screening fashion show video is improved.
When it is implemented, in above-mentioned steps 101, obtains in fashion show video per frame image, need in fashion show video
In, picture image is analyzed frame by frame, finds every frame image.
When it is implemented, above-mentioned steps 102 are that characteristics of image in every frame image is met to the figure of the first pre-set image feature
As deletion, that is, unfavorable scene image being removed, there are many unsuitable characteristics of image, such as：The angle of video camera, Mo Teqing
Clear degree and bright and dark light degree etc..The image for deleting step 102 the first pre-set image feature of satisfaction below is situated between in detail
It continues as follows.
First, inventor has found, if the posture of identified model is improper, then it is assumed that the inappropriate scene graph of the posture
It seem a scene image for being not suitable for analysis.
In one embodiment, the first pre-set image feature includes：Preset model's posture；
Image characteristic analysis is carried out to every frame image, characteristics of image in every frame image is met into the first pre-set image feature
Image-erasing, including：Image characteristic analysis is carried out to every frame image, it will be per model's posture in frame image and preset model's posture
Consistent image-erasing.
When it is implemented, suitable posture refers to that subject goal (model) should be based on the walking postures of front, because of mistake
It is unfavorable for identifying the worn types of garments of model and style in fashion show video in complicated posture.For example, the posture of standing akimbo of model,
Front folded arm posture etc., these are all unfavorable for identifying that the worn types of garments of model in fashion show video, the purpose of the present embodiment are
Identify that these image-erasings are simplified fashion show by these images for being unfavorable for identifying the worn clothes of model in fashion show video
Image in video.
Second, inventor has found, if identified main object (can be model) is fuzzy, is not considered as one and is suitble to
The scene image of analysis.
In one embodiment, the first pre-set image feature includes：Preset model's clarity；
Image characteristic analysis is carried out to every frame image, characteristics of image in every frame image is met into the first pre-set image feature
Image-erasing, including：
Image characteristic analysis is carried out to every frame image, preset model's clarity will be less than per model's clarity in frame image
Image-erasing.
When it is implemented, main body it is smudgy i.e. model's clarity, refer to due to light or focusing etc., cause by
Model's picture of shooting is unintelligible, is unfavorable for identifying the worn types of garments of model in fashion show video.
Third, inventor have found, if identified discovery has multiple main bodys (model), then it is assumed that this has the field of multiple main bodys
Scape image is an image for being not suitable for analysis.
In one embodiment, the first pre-set image feature includes：Preset model's quantity；
Image characteristic analysis is carried out to every frame image, characteristics of image in every frame image is met into the first pre-set image feature
Image-erasing, including：
Image characteristic analysis is carried out to every frame image, model's quantity in every frame image is more than to the figure of preset model's quantity
As deleting.
When it is implemented, preset model's quantity is to have the case where multiple models, which refers to same in one picture
When there are multiple models to occur, at this time it is possible that the problems such as covering and cover mutually between model, is unfavorable for identifying fashion show
The worn types of garments of model in video.
4th, inventor has found, if identified discovery camera angle is improper, then it is assumed that the camera angle does not conform to
Suitable scene image is a scene image for being not suitable for analysis.Camera angle is improper to be referred to：Photographer is in shooting mould
In special process, video camera can convert multiple angles and focal length, cause the angle of identification complexity.
In one embodiment, the first pre-set image feature includes：Preset camera angle range；
Image characteristic analysis is carried out to every frame image, characteristics of image in every frame image is met into the first pre-set image feature
Image-erasing, including：
Image characteristic analysis is carried out to every frame image, it will be per camera angle in frame image in preset camera angle model
Enclose interior image-erasing.
When it is implemented, most suitable camera angle refers to：Video camera is preferred with normotopia in face of model.It is unsuitable to take the photograph
The angle of camera, i.e., preset camera angle refer to：Due in shooting process, photographer in order to pursue shooting effect,
Model may be shot with multiple angles, for example, from tilt angle, lower depression angle, left side angle, right side angle
Deng.In the step of carrying out final clothes fashion and type identification, need in the clothes and database for the model that will be extracted in advance
The clothes of storage are matched, and the sample of clothes is mainly positive in database, are not suitable in this way if existed in image
Video camera angle shot model, just to being identified that the types of garments of model increases complexity.
5th, inventor has found, if identified picture dim light, causes main body unintelligible, is not easy to identification main body,
Then it is not considered as an image for being suitble to analysis.
In one embodiment, the first pre-set image feature includes：Preset bright and dark light degree；
Image characteristic analysis is carried out to every frame image, characteristics of image in every frame image is met into the first pre-set image feature
Image-erasing, including：
Image characteristic analysis is carried out to every frame image, preset bright and dark light degree will be less than per bright and dark light degree in frame image
Image-erasing.
When it is implemented, dim light refers to individual picture insufficient lights, model's main body can not be protruded, is unfavorable for identifying
Model.Judge that the method for picture insufficient light (bright and dark light degree) is that judgement picture entirety brightness is inadequate, the model master extracted
Also brightness is inadequate for body.Picture entirety brightness can be detected using the statistical value average value of the Y-component of YUV color spaces, it is typical
Ground, if this value is less than 30, it is believed that be that picture entirety brightness is inadequate, can actually used using this thresholding as parameter
Middle adjustment uses.
6th, inventor has found, if identified picture background is chaotic complicated, is not easy to identification main body, is then not considered as
One scene picture for being suitble to analysis.
In one embodiment, the first pre-set image feature includes：The quantity of personage and article in preset background；
Image characteristic analysis is carried out to every frame image, characteristics of image in every frame image is met into the first pre-set image feature
Image-erasing, including：
Image characteristic analysis is carried out to every frame image, the quantity of personage and article in background in every frame image is more than default
Background in the quantity of personage and article picture delete.
When it is implemented, the quantity of personage and article is more than preset quantity representative picture background clutter, including example in background
Such as：Background is auditorium, and spectators' shadow is shaken, and may have complicated light-illuminating, and model is caused to be blended in a picture with spectators
Face occurs, it is not easy to extract main body model, then it is assumed that is an image for being not suitable for analysis.
Above-mentioned steps 103 are that after deleting the image for being not suitable for analysis, suitable analysis is found in remaining scene picture
Scene picture.The step 103 is introduced below as follows.
In clothing show fashion show video, screen shot, automatic identification main target therein (model) are analyzed frame by frame.When
Main target is placed in the middle in entire Pictures location, and the area occupied is in certain suitable range, then it is assumed that this picture generation
The scene of table is a scene for being suitable for analysis.
In one embodiment, the second pre-set image feature includes：Preset model's position range and preset model are big
Small range；
In the remaining image per frame, the image for meeting the second pre-set image feature is filtered out, including：
In the remaining image per frame, model's figure of the position in preset model's position range in the picture is filtered out
Picture；
In the remaining image per frame, model's figure of the size in preset model's magnitude range in the picture is filtered out
Picture.
In one embodiment, preset model's position range is that model is placed in the middle in every frame image；Preset model is big
Small range is：The size of model is more than 1/100 with the ratio per frame image size, and is less than 2/3.
When it is implemented, main target (model) is placed in the middle in entire Pictures location, and the area occupied is in certain conjunction
In suitable range.When cameraman's shooting picture, it will usually subject goal is placed on the center of picture, it is therein " to occupy
In ", subject goal is indicated near picture diagonal line crosspoint, and center ofthe is also placed in the middle up and down." suitable range " indicates：
The area of subject goal in the picture confined is not too small nor excessive.Typically, if the subject goal confined occupies
Area be less than overall picture area 1/100, then be too small；If the area that the subject goal confined occupies is more than total painting
The 2/3 of face area is then excessive.The two threshold values can be used as parameter, and dynamic adjusts in actual use.
Based on same inventive concept, a kind of dress screening image in fashion show video is additionally provided in the embodiment of the present invention
It sets, as described in the following examples.The principle and screening fashionable dress solved the problems, such as due to the device of image in screening fashion show video
The method of image is similar in elegant video, therefore the implementation for screening the device of image in fashion show video may refer to screening fashion show
The implementation of the method for image in video, overlaps will not be repeated.Used below, term " unit " or " module " can be with
Realize the combination of the software and/or hardware of predetermined function.Although device described in following embodiment is preferably come with software real
It is existing, but the realization of the combination of hardware or software and hardware is also that may and be contemplated.
Fig. 2 is the structural schematic diagram that the device of image in fashion show video is screened in the embodiment of the present invention, as shown in Fig. 2,
The device includes：
Acquisition module 02, for obtaining in fashion show video per frame image；
Image analysis module 04 will expire for carrying out image characteristic analysis to every frame image per characteristics of image in frame image
The image-erasing of the first pre-set image feature of foot；
Picture recognition module 06, in the remaining image per frame, filtering out the figure for meeting the second pre-set image feature
Picture.
In one embodiment, the first pre-set image feature includes：Preset model's posture；
Image analysis module 04 is specifically used for carrying out image characteristic analysis to every frame image, will be per model's posture in frame image
The image-erasing consistent with preset model's posture.
In one embodiment, the first pre-set image feature includes：Preset model's clarity and preset model's quantity；
Image analysis module 04 is specifically used for：
Image characteristic analysis is carried out to every frame image, preset model's clarity will be less than per model's clarity in frame image
Image-erasing；
Image characteristic analysis is carried out to every frame image, model's quantity in every frame image is more than to the figure of preset model's quantity
As deleting.
In one embodiment, the first pre-set image feature includes：Preset camera angle range and preset light
Shading value；
Image analysis module 04 is specifically used for：
Image characteristic analysis is carried out to every frame image, it will be per camera angle in frame image in preset camera angle model
Enclose interior image-erasing；
Image characteristic analysis is carried out to every frame image, preset bright and dark light degree will be less than per bright and dark light degree in frame image
Image-erasing.
In one embodiment, the first pre-set image feature includes：The quantity of personage and article in preset background；
Image analysis module 04 is specifically used for：
Image characteristic analysis is carried out to every frame image, the quantity of personage and article in background in every frame image is more than default
Background in the quantity of personage and article picture delete.
In one embodiment, the second pre-set image feature includes：Preset model's position range and preset model are big
Small range；
Picture recognition module 06 is specifically used for：
In the remaining image per frame, model's figure of the position in preset model's position range in the picture is filtered out
Picture；
In the remaining image per frame, model's figure of the size in preset model's magnitude range in the picture is filtered out
Picture.
In one embodiment, preset model's position range is that model is placed in the middle in every frame image；Preset model is big
Small range is：The size of model is more than 1/100 with the ratio per frame image size, and is less than 2/3.
Based on same inventive concept, model in a kind of identification fashion show video is additionally provided in the embodiment of the present invention and wears clothes
The method for filling type, as described in the following examples.By the method solution of the worn types of garments of model in identification fashion show video
Certainly the principle of problem is similar to the method for image in screening fashion show video, therefore identifies the worn clothes of model in fashion show video
The implementation of the method for type may refer to the implementation of the method for image in screening fashion show video, and overlaps will not be repeated.With
Used in lower, the combination of the software and/or hardware of predetermined function may be implemented in term " unit " or " module ".Although following
Device described in embodiment is preferably realized with software, but the realization of the combination of hardware or software and hardware is also
It may and be contemplated.
Fig. 3 is the flow signal that the method for the worn types of garments of model in fashion show video is identified in the embodiment of the present invention
Figure, as shown in figure 3, this method comprises the following steps：
Step 1011：Using the method for image in above-mentioned screening fashion show video, filters out and meet in fashion show video
The image of two pre-set image features；
Step 1012：In the image for meeting the second pre-set image feature, the style and type of the worn clothes of model are identified.
With in the prior art in complex scene identify the worn clothes of model style and type scheme compared with, this hair
The scheme of the worn types of garments of model in the identification fashion show video that bright embodiment provides, using in above-mentioned screening fashion show video
The scheme of image filters out the image for meeting the second pre-set image feature in fashion show video, to image in fashion show video into
It has gone simplification, has obtained the image of suitable clothes fashion type analysis；In the image for meeting the second pre-set image feature, identification
The style and type of the worn clothes of model improve the accuracy of types of garments identification.
When it is implemented, can also identify the side using image in above-mentioned screening fashion show video in the embodiment of the present invention
Method filters out the image for meeting the second pre-set image feature in fashion show video, the style and type of identification model institute wear shoes cap.
Based on same inventive concept, model in a kind of identification fashion show video is additionally provided in the embodiment of the present invention and wears clothes
The device for filling type, as described in the following examples.By the device solution of the worn types of garments of model in identification fashion show video
Certainly the principle of problem is similar to the method for the worn types of garments of model in identification fashion show video, therefore identifies in fashion show video
The implementation of the device of the worn types of garments of model may refer to the method for the worn types of garments of model in identification fashion show video
Implement, overlaps will not be repeated.Used below, the software of predetermined function may be implemented in term " unit " or " module "
And/or the combination of hardware.Although device described in following embodiment is preferably realized with software, hardware or soft
The realization of the combination of part and hardware is also that may and be contemplated.
Fig. 4 is the structural representation that the device of the worn types of garments of model in fashion show video is identified in the embodiment of the present invention
Figure, as shown in figure 4, the device includes：
Optical sieving module 022, for using the device such as image in above-mentioned screening fashion show video, filtering out fashion show
Meet the image of the second pre-set image feature in video；
Types of garments identification module 044, in the image for meeting the second pre-set image feature, identification model to wear clothes
The style and type of dress.
The embodiment of the present invention realizes following technique effect：Technical solution provided in an embodiment of the present invention is in clothing show environment
Under, to avoid wrong identification, it can only retain the picture of suitable scene, remove the picture of unfavorable scene, improve clothes
Fill the accuracy of type identification.
Obviously, those skilled in the art should be understood that each module of the above-mentioned embodiment of the present invention or each step can be with
It is realized with general computing device, they can be concentrated on a single computing device, or be distributed in multiple computing devices
On the network formed, optionally, they can be realized with the program code that computing device can perform, it is thus possible to by it
Store and be performed by computing device in the storage device, and in some cases, can be to be held different from sequence herein
The shown or described step of row, either they are fabricated to each integrated circuit modules or will be multiple in them
Module or step are fabricated to single integrated circuit module to realize.In this way, the embodiment of the present invention be not limited to it is any specific hard
Part and software combine.
The foregoing is only a preferred embodiment of the present invention, is not intended to restrict the invention, for the skill of this field
For art personnel, the embodiment of the present invention can have various modifications and variations.All within the spirits and principles of the present invention, made by
Any modification, equivalent substitution, improvement and etc. should all be included in the protection scope of the present invention.
Claims (10)
1. a kind of method of image in screening fashion show video, which is characterized in that including：
It obtains in fashion show video per frame image；
Image characteristic analysis is carried out per frame image to described, characteristics of image in every frame image is met into the first pre-set image feature
Image-erasing；
In the remaining image per frame, the image for meeting the second pre-set image feature is filtered out.
2. the method for screening image in fashion show video as described in claim 1, which is characterized in that first pre-set image
Feature includes：Preset model's posture；
Image characteristic analysis is carried out per frame image to described, characteristics of image in every frame image is met into the first pre-set image feature
Image-erasing, including：Image characteristic analysis is carried out per frame image to described, it will be per model's posture in frame image and preset model
The consistent image-erasing of posture.
3. the method for screening image in fashion show video as described in claim 1, which is characterized in that first pre-set image
Feature includes：
Preset model's clarity；
Image characteristic analysis is carried out per frame image to described, characteristics of image in every frame image is met into the first pre-set image feature
Image-erasing, including：
Image characteristic analysis is carried out per frame image to described, preset model's clarity will be less than per model's clarity in frame image
Image-erasing；
And/or preset model's quantity；
Image characteristic analysis is carried out per frame image to described, characteristics of image in every frame image is met into the first pre-set image feature
Image-erasing, including：
Image characteristic analysis is carried out per frame image to described, model's quantity in every frame image is more than to the figure of preset model's quantity
As deleting.
4. the method for screening image in fashion show video as described in claim 1, which is characterized in that first pre-set image
Feature includes：
Preset camera angle range；
Image characteristic analysis is carried out per frame image to described, characteristics of image in every frame image is met into the first pre-set image feature
Image-erasing, including：
Image characteristic analysis is carried out per frame image to described, it will be per camera angle in frame image in preset camera angle model
Enclose interior image-erasing；
And/or preset bright and dark light degree；
Image characteristic analysis is carried out per frame image to described, characteristics of image in every frame image is met into the first pre-set image feature
Image-erasing, including：
Image characteristic analysis is carried out per frame image to described, preset bright and dark light degree will be less than per bright and dark light degree in frame image
Image-erasing.
5. the method for screening image in fashion show video as described in claim 1, which is characterized in that first pre-set image
Feature includes：The quantity of personage and article in preset background；
Image characteristic analysis is carried out per frame image to described, characteristics of image in every frame image is met into the first pre-set image feature
Image-erasing, including：
Image characteristic analysis is carried out per frame image to described, the quantity of personage and article in background in every frame image is more than default
Background in the quantity of personage and article picture delete.
6. the method for screening image in fashion show video as described in claim 1, which is characterized in that second pre-set image
Feature includes：
Preset model's position range；
In the remaining image per frame, the image for meeting the second pre-set image feature is filtered out, including：
In the remaining image per frame, model's image of the position in preset model's position range in the picture is filtered out；
And/or preset model's magnitude range；
In the remaining image per frame, the image for meeting the second pre-set image feature is filtered out, including：
In the remaining image per frame, model's image of the size in preset model's magnitude range in the picture is filtered out.
7. the method for screening image in fashion show video as claimed in claim 6, which is characterized in that the preset model position
It sets ranging from：Model is placed in the middle in every frame image；Preset model's magnitude range is：The size of model and every frame image are big
Small ratio is more than 1/100, and is less than 2/3.
8. the device of image in a kind of screening fashion show video, which is characterized in that including：
Acquisition module, for obtaining in fashion show video per frame image；
Image analysis module will meet for carrying out image characteristic analysis per frame image to described per characteristics of image in frame image
The image-erasing of first pre-set image feature；
Picture recognition module, in the remaining image per frame, filtering out the image for meeting the second pre-set image feature.
9. a kind of method of the worn types of garments of model in identification fashion show video, which is characterized in that including：
Using the method for screening image in fashion show video as described in claim 1 to 7 is any, filter out full in fashion show video
The image of the second pre-set image feature of foot；
In the image for meeting the second pre-set image feature, the style and type of the worn clothes of model are identified.
10. the device of the worn types of garments of model in a kind of identification fashion show video, which is characterized in that including：
Optical sieving module filters out fashionable dress for the device using image in screening fashion show video as claimed in claim 8
Meet the image of the second pre-set image feature in elegant video；
Types of garments identification module, in the image for meeting the second pre-set image feature, identifying the money of the worn clothes of model
Formula and type.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710192736.7A CN108664845A (en) | 2017-03-28 | 2017-03-28 | The method and device screened image in fashion show video, identify model's types of garments |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710192736.7A CN108664845A (en) | 2017-03-28 | 2017-03-28 | The method and device screened image in fashion show video, identify model's types of garments |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108664845A true CN108664845A (en) | 2018-10-16 |
Family
ID=63786731
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710192736.7A Pending CN108664845A (en) | 2017-03-28 | 2017-03-28 | The method and device screened image in fashion show video, identify model's types of garments |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108664845A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110414331A (en) * | 2019-06-20 | 2019-11-05 | 深圳壹账通智能科技有限公司 | Model's screening technique, device, equipment and readable storage medium storing program for executing |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104954659A (en) * | 2014-03-28 | 2015-09-30 | 宇龙计算机通信科技(深圳)有限公司 | Method and device for processing image information |
| CN105488526A (en) * | 2015-11-26 | 2016-04-13 | 熊桂荣 | Automatic screening method of group sharing photos |
| CN105631043A (en) * | 2016-01-26 | 2016-06-01 | 公安部第一研究所 | Video processing method and device |
-
2017
- 2017-03-28 CN CN201710192736.7A patent/CN108664845A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104954659A (en) * | 2014-03-28 | 2015-09-30 | 宇龙计算机通信科技(深圳)有限公司 | Method and device for processing image information |
| CN105488526A (en) * | 2015-11-26 | 2016-04-13 | 熊桂荣 | Automatic screening method of group sharing photos |
| CN105631043A (en) * | 2016-01-26 | 2016-06-01 | 公安部第一研究所 | Video processing method and device |
Non-Patent Citations (1)
| Title |
|---|
| 安立新: "服装款式图提取及其模式识别的研究", 《中国博士学位论文全文数据库 信息科技辑》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110414331A (en) * | 2019-06-20 | 2019-11-05 | 深圳壹账通智能科技有限公司 | Model's screening technique, device, equipment and readable storage medium storing program for executing |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105323456B (en) | For the image preview method of filming apparatus, image capturing device | |
| CN105654471B (en) | Augmented reality AR system and method applied to internet video live streaming | |
| CN109068056B (en) | Electronic equipment, filter processing method of image shot by electronic equipment and storage medium | |
| US9129381B2 (en) | Modification of post-viewing parameters for digital images using image region or feature information | |
| US20180013950A1 (en) | Modification of post-viewing parameters for digital images using image region or feature information | |
| CN109671137B (en) | Picture text matching method, electronic equipment and storage medium | |
| US11308335B2 (en) | Intelligent video surveillance system and method | |
| CN111402135A (en) | Image processing method, image processing device, electronic equipment and computer readable storage medium | |
| GB2373124A (en) | A method of selectively storing images | |
| JP2009031939A (en) | Image processing apparatus, method and program | |
| CN106650671A (en) | Human face identification method, apparatus and system | |
| CN109191371A (en) | A method of it judging automatically scenery type and carries out image filters processing | |
| JP2004520735A (en) | Automatic cropping method and apparatus for electronic images | |
| CN109191558A (en) | Image method for polishing and device | |
| CN108600649A (en) | A kind of Intelligent House Light brightness acquisition and control method | |
| CN111654643B (en) | Exposure parameter determination method and device, unmanned aerial vehicle and computer readable storage medium | |
| CN109299299B (en) | Method and device for adjusting color and mobile terminal | |
| CN108093170B (en) | User photographing method, device and equipment | |
| CN109089045A (en) | A kind of image capture method and equipment and its terminal based on multiple photographic devices | |
| CN104978734B (en) | The extracting method and device of foreground image | |
| US20140078321A1 (en) | Motion blur estimation and restoration using light trails | |
| CN108664845A (en) | The method and device screened image in fashion show video, identify model's types of garments | |
| CN114257730A (en) | Image data processing method and device, storage medium and computer equipment | |
| TWI584228B (en) | Method of capturing and reconstructing court lines | |
| KR100350789B1 (en) | Method of raw color adjustment and atmosphere color auto extract in a image reference system |
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: 20181016 |
|
| RJ01 | Rejection of invention patent application after publication |