CN106339655A - Video shot marking method and device - Google Patents
Video shot marking method and device Download PDFInfo
- Publication number
- CN106339655A CN106339655A CN201510392380.2A CN201510392380A CN106339655A CN 106339655 A CN106339655 A CN 106339655A CN 201510392380 A CN201510392380 A CN 201510392380A CN 106339655 A CN106339655 A CN 106339655A
- Authority
- CN
- China
- Prior art keywords
- cumulant
- camera lens
- lens
- dimensional motion
- class
- 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
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- 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 invention provides a video shot marking method which is used for achieving a purpose of marking splendid shots in a video. The video shot marking method comprises the steps of extracting a video to be tested from a test library to act as the current video, and acquiring a three-dimensional motion vector cumulant of a current shot of the current video; judging whether the content of the current shot belongs to a program type A, performing the next step if the content of the current short belongs to the program type A, otherwise, not performing the next step; when the content of the current shot belongs to the program type A, matching the three-dimensional motion vector cumulant of the current shot with three-dimensional motion vector cumulants of all splendid shots of the program type A in a shot template library to acquire matching values; if any matching value is greater than a preset shot threshold, regarding the current shot is a splendid shot of the program type A, and marking the current shot; otherwise, regarding the current shot to be not a splendid shot. The video shot marking method saves time and labor, and greatly reduces the labor cost.
Description
Technical field
The present invention relates to the communications field, particularly to a kind of video lens mask method and device.
Background technology
With the raising of people's living standard, sports become an indispensable part in life.
Sports are not only able to improve vitality, promote mental health, also promote improving and development of individual character；
In every sports, basketball movement is well received, and it covers multiple body kinematicses such as race, jump, throwing
Form, and exercise intensity is larger, therefore can comprehensively, effectively, synthetically promote body constitution and function of human body
Development in an all-round way, be that the activities of people lay solid physical basis, thus improving the quality of life；Separately
The activity of outer basketball various informative, there are higher property of participation, interest, contingency, recreational and sports
Property etc., multiple demands of different crowd can be met；Meanwhile, more sports enthusiasts like watching basket
Ball is competed, and athlete may give play to supranormal levels in play at any time, make match be filled with uncertainty,
Each splendid moment can allow basket ball fan extremely excited, and therefore excellent physical culture camera lens is frequently necessary to samsara
Play.
In prior art, in the application of CBIR and video labeling, there is important answering
With being the Highlight in mark video, but it is all artificial mark so far, waste time and energy.
Content of the invention
The present invention provides a kind of video lens mask method and device, in order to reach the excellent mirror in mark video
The purpose of head.
The present invention provides a kind of video lens mask method, comprising:
Extract video to be tested as current video from test library, obtain current lens of current video
Three-dimensional motion vector cumulant；
According to the three-dimensional motion vector cumulant of one current lens, judge the content of described current lens
Whether belonging to a class in program category, if belonging to a class in program category, carrying out next step, otherwise
Do not carry out next step；
When belonging to a class in program category when the content of described current lens, by one current lens
All Highlights of a class program category in three-dimensional motion vector cumulant, with default camera lens template base
Three-dimensional motion vector cumulant mated respectively, draw each matching value；If its of described each matching value
In any one matching value be more than default camera lens threshold value then it is assumed that described current lens are the essences in a class program
Color camera lens is simultaneously labeled；Otherwise it is assumed that described current lens are not Highlights.
The beneficial effect of the embodiment of the present invention includes: extracts video to be tested from test library as working as forward sight
Frequently, obtain the three-dimensional motion vector cumulant of current lens of current video；According to one current
The three-dimensional motion vector cumulant of camera lens, judges whether the content of described current lens belongs in program category
A class, if belonging to a class in program category, carries out next step, does not otherwise carry out next step；When described
When the content of current lens belongs to a class in program category, by the three-dimensional motion of one current lens to
Amount cumulant, with default camera lens template base in a class program category a default camera lens three-dimensional motion to
Amount cumulant is mated, and draws matching value；If described matching value be more than default camera lens threshold value then it is assumed that
Described current lens are Highlights and are labeled；Otherwise it is assumed that described current lens are not Highlights；
Highlight in video, save trouble and labor are marked by methods described, greatly reduces cost of labor；And
Judged twice by program category and Highlight, improve the accuracy of mark.
In one embodiment, with the video of artificial mark, training obtains camera lens template base: pre- for one
If each two field picture of camera lens, select a certain window, extract in described certain window three-dimensional motion to
Amount, then by the three-dimensional motion vector of each two field picture of one default camera lens, in each three-dimensional dimension
Upper carry out accumulation respectively and obtain a default camera lens three-dimensional motion vector cumulant, the window of described certain window
It is inside Highlight, outside window be and Highlight irrelevant contents；
The three-dimensional motion vector cumulant of one default camera lens is added to camera lens template base.
In this embodiment, by the video with artificial mark, training obtains camera lens template base: pre- for one
If each two field picture of camera lens, select a certain window, extract in described certain window three-dimensional motion to
Amount, then by the three-dimensional motion vector of each two field picture of one default camera lens, in each three-dimensional dimension
Upper carry out accumulation respectively and obtain a default camera lens three-dimensional motion vector cumulant, the window of described certain window
It is inside Highlight, outside window be and Highlight irrelevant contents；Three maintenance and operations by one default camera lens
Moving vector cumulant is added to camera lens template base；By camera lens template base storehouse as a comparison, judge described current mirror
Whether head is Highlight, further increases annotating efficiency, saves cost of labor.
In one embodiment, described three-dimensional motion vector includes: horizontally rotate vector, vertical rotation is vectorial,
Focal length axle motion-vector；
Described accumulated respectively in each three-dimensional dimension, comprising: one default camera lens is each
The vector that horizontally rotates of two field picture carries out summation in the horizontal direction and obtains horizontally rotating vectorial cumulant, by institute
The vertical rotation vector stating each two field picture of a default camera lens carries out summation in vertical direction and obtains vertically
Rotating vector cumulant, by the focal length axle rotating vector of each two field picture of one default camera lens in focal length axle
Summation is carried out on direction obtain horizontally rotating vectorial cumulant.
In this embodiment, described three-dimensional motion vector includes: horizontally rotates vector, vertical rotation vector, Jiao
Away from axle motion-vector；Described accumulated respectively in each three-dimensional dimension, comprising: will be one pre-
If the vector that horizontally rotates of each two field picture of camera lens carries out summation in the horizontal direction and obtains horizontally rotating vector
Cumulant, the vertical rotation vector of each two field picture of one default camera lens is asked in vertical direction
With obtain vertical rotation vector cumulant, by the focal length axle of each two field picture of one default camera lens rotate to
Amount carries out summation on focal length direction of principal axis and obtains horizontally rotating vectorial cumulant；Further to three-dimensional motion vector
It is optimized with three-dimensional motion vector cumulant.
In one embodiment, the three-dimensional motion vector accumulation of described current lens obtaining current video
Amount, comprising: current video is cut into different camera lenses according to color, for current lens, obtains
Three-dimensional motion vector cumulant；
Whether the described content judging described current lens belongs to a class in program category, comprising: will be described
The three-dimensional motion vector cumulant of one current lens, with one of default class template storehouse default camera lens three
Maintenance and operation moving vector cumulant is mated, and draws matching value, and the content of one default camera lens belongs to program
A class in type；
If described matching value is more than default class threshold value then it is assumed that the content of described current lens belongs to program class
A class in type；Otherwise it is assumed that the content of described current lens is not belonging to a class in program category.
In this embodiment, current video is cut into different camera lenses according to color, for current lens,
Obtain three-dimensional motion vector cumulant；First current video is cut into different camera lenses, obtains in current lens
Three-dimensional motion vector cumulant, is optimized further to methods described.
In one embodiment, the image in described Highlight is carried out as the image in new camera lens defeated
Go out；The Highlight of mark in test library is exported, is formed the featured videos collection of choice specimens.
In this embodiment, the image in described Highlight is exported as the image in new camera lens；Will
The Highlight of mark in test library is exported, and forms the featured videos collection of choice specimens；Save trouble and labor, drops significantly
Low cost of labor.
The invention provides a kind of video lens annotation equipment is it is characterised in that include:
Extraction module, for extracting video to be tested from test library as current video, obtains current video
Current lens three-dimensional motion vector cumulant；
Sort module, for the three-dimensional motion vector cumulant according to one current lens, judges described
Whether the content of current lens belongs to a class in program category, if belonging to a class in program category, enters
Row next step, does not otherwise carry out next step；
Matching module, during for belonging to a class in program category when the content of described current lens, will be described
The three-dimensional motion vector cumulant of one current lens, with a class program category in default camera lens template base
All Highlights three-dimensional motion vector cumulant mated respectively, draw each matching value；If institute
One matching value of any of which stating each matching value is more than default camera lens threshold value then it is assumed that described current lens are
Highlight in a class program is simultaneously labeled；Otherwise it is assumed that described current lens are not Highlights.
The beneficial effect of the embodiment of the present invention includes: extracts video to be tested from test library as working as forward sight
Frequently, obtain the three-dimensional motion vector cumulant of current lens of current video；According to one current
The three-dimensional motion vector cumulant of camera lens, judges whether the content of described current lens belongs in program category
A class, if belonging to a class in program category, carries out next step, does not otherwise carry out next step；When described
When the content of current lens belongs to a class in program category, by the three-dimensional motion of one current lens to
Amount cumulant, with default camera lens template base in a class program category a default camera lens three-dimensional motion to
Amount cumulant is mated, and draws matching value；If described matching value be more than default camera lens threshold value then it is assumed that
Described current lens are Highlights and are labeled；Otherwise it is assumed that described current lens are not Highlights；
Highlight in video, save trouble and labor are marked by methods described, greatly reduces cost of labor；And
Judged twice by program category and Highlight, improve the accuracy of mark.
In one embodiment, described device, comprising:
Presetting module, for by the video of artificial mark, training obtains camera lens template base: default for one
Each two field picture of camera lens, selects a certain window, extracts three-dimensional motion vector in described certain window,
Then by the three-dimensional motion vector of each two field picture of one default camera lens, in each three-dimensional dimension
Carry out summation respectively and obtain a default camera lens three-dimensional motion vector cumulant, in the window of described certain window
It is Highlight, outside window be and Highlight irrelevant contents；
The three-dimensional motion vector cumulant of one default camera lens is added to camera lens template base.
In one embodiment, described three-dimensional motion vector, comprising: horizontally rotate vector, vertical rotation to
Amount, focal length axle motion-vector；
Described presetting module, comprising: accumulation submodule, for by each two field picture of one default camera lens
The vector that horizontally rotates carry out summation in the horizontal direction and obtain horizontally rotating vectorial cumulant, will be one
The vertical rotation vector of each two field picture of default camera lens carry out in vertical direction summation obtain vertical rotation to
Amount cumulant, by the focal length axle rotating vector of each two field picture of one default camera lens on focal length direction of principal axis
Carry out summation to obtain horizontally rotating vectorial cumulant.
In one embodiment, described matching module, including matched sub-block, for by current video according to
Color is cut into different camera lenses, for current lens, obtains three-dimensional motion vector cumulant；
Described sort module, including classification submodule, for by the three-dimensional motion of one current lens to
Amount cumulant, is carried out with one of default class template storehouse default camera lens three-dimensional motion vector cumulant
Join, draw matching value, the content of one default camera lens belongs to a class in program category；
If described matching value is more than default class threshold value then it is assumed that the content of described current lens belongs to program class
A class in type；Otherwise it is assumed that the content of described current lens is not belonging to a class in program category.
In one embodiment, described device, comprising:
Output module, for being exported the image in described Highlight as the image in new camera lens；
The Highlight of mark in test library is exported, is formed the featured videos collection of choice specimens.
Other features and advantages of the present invention will illustrate in the following description, and, partly from explanation
Become apparent in book, or understood by implementing the present invention.The purpose of the present invention and other advantages can
Realized by specifically noted structure in the description write, claims and accompanying drawing and obtain
?.
Below by drawings and Examples, technical scheme is described in further detail.
Brief description
Accompanying drawing is used for providing a further understanding of the present invention, and constitutes a part for description, with this
Bright embodiment is used for explaining the present invention together, is not construed as limiting the invention.In the accompanying drawings:
Fig. 1 is a kind of video lens mask method flow chart shown in the present invention one exemplary embodiment；
Fig. 2 is a kind of video lens mask method flow chart shown in the present invention one exemplary embodiment；
Fig. 3 is a kind of video lens mask method flow chart shown in the present invention one exemplary embodiment；
Fig. 4 is a kind of video lens mask method flow chart shown in the present invention one exemplary embodiment；
Fig. 5 is a kind of video lens mask method flow chart shown in the present invention one exemplary embodiment；
Fig. 6 is a kind of video lens annotation equipment block diagram shown in the present invention one exemplary embodiment；
Fig. 7 is a kind of video lens annotation equipment block diagram shown in the present invention one exemplary embodiment；
Fig. 8 is a kind of video lens annotation equipment block diagram shown in the present invention one exemplary embodiment；
Fig. 9 is a kind of video lens annotation equipment block diagram shown in the present invention one exemplary embodiment；
Figure 10 is a kind of video lens annotation equipment block diagram shown in the present invention one exemplary embodiment；
Figure 11 is a kind of video lens annotation equipment block diagram shown in the present invention one exemplary embodiment；
Specific embodiment
Below in conjunction with accompanying drawing, the preferred embodiments of the present invention are illustrated it will be appreciated that described herein
Preferred embodiment is merely to illustrate and explains the present invention, is not intended to limit the present invention.
As Fig. 1, the present invention provides a kind of video lens mask method, including step 101-103:
Step 101, extracts video to be tested from test library as current video, obtains the one of current video
The three-dimensional motion vector cumulant of individual current lens；
Step 102, according to the three-dimensional motion vector cumulant of one current lens, judges described current
Whether the content of camera lens belongs to a class in program category, if belonging to a class in program category, under carrying out
One step, does not otherwise carry out next step；
Step 103, when belonging to a class in program category when the content of described current lens, will be one
The institute of a class program category in the three-dimensional motion vector cumulant of current lens, with default camera lens template base
The three-dimensional motion vector cumulant having Highlight is mated respectively, draws each matching value；If described each
One matching value of any of which of matching value is more than default camera lens threshold value then it is assumed that described current lens are a classes
Highlight in program is simultaneously labeled；Otherwise it is assumed that described current lens are not Highlights.
The beneficial effect of the embodiment of the present invention includes: extracts video to be tested from test library as working as forward sight
Frequently, obtain the three-dimensional motion vector cumulant of current lens of current video；According to one current
The three-dimensional motion vector cumulant of camera lens, judges whether the content of described current lens belongs in program category
A class, if belonging to a class in program category, carries out next step, does not otherwise carry out next step；When described
When the content of current lens belongs to a class in program category, by the three-dimensional motion of one current lens to
Amount cumulant, with default camera lens template base in a class program category a default camera lens three-dimensional motion to
Amount cumulant is mated, and draws matching value；If described matching value be more than default camera lens threshold value then it is assumed that
Described current lens are Highlights and are labeled；Otherwise it is assumed that described current lens are not Highlights；
Highlight in video, save trouble and labor are marked by methods described, greatly reduces cost of labor；And
Judged twice by program category and Highlight, improve the accuracy of mark.
In one embodiment, as Fig. 2, step 100, with the video of artificial mark, train and obtain camera lens mould
Plate storehouse: for each two field picture of a default camera lens, select a certain window, extract described specific
Three-dimensional motion vector in window, then by the three-dimensional motion vector of each two field picture of one default camera lens,
Accumulation is carried out respectively on each three-dimensional dimension and obtains a default camera lens three-dimensional motion vector cumulant, institute
Stating in the window of certain window is Highlight, outside window is and Highlight irrelevant contents；
The three-dimensional motion vector cumulant of one default camera lens is added to camera lens template base.
In this embodiment, by the video with artificial mark, training obtains camera lens template base: pre- for one
If each two field picture of camera lens, select a certain window, extract in described certain window three-dimensional motion to
Amount, then by the three-dimensional motion vector of each two field picture of one default camera lens, in each three-dimensional dimension
Upper carry out accumulation respectively and obtain a default camera lens three-dimensional motion vector cumulant, the window of described certain window
It is inside Highlight, outside window be and Highlight irrelevant contents；Three maintenance and operations by one default camera lens
Moving vector cumulant is added to camera lens template base；By camera lens template base storehouse as a comparison, judge described current mirror
Whether head is Highlight, further increases annotating efficiency, saves cost of labor.
In one embodiment, described three-dimensional motion vector includes: horizontally rotate vector, vertical rotation is vectorial,
Focal length axle motion-vector；
Described accumulated respectively in each three-dimensional dimension, comprising: one default camera lens is each
The vector that horizontally rotates of two field picture carries out summation in the horizontal direction and obtains horizontally rotating vectorial cumulant, by institute
The vertical rotation vector stating each two field picture of a default camera lens carries out summation in vertical direction and obtains vertically
Rotating vector cumulant, by the focal length axle rotating vector of each two field picture of one default camera lens in focal length axle
Summation is carried out on direction obtain horizontally rotating vectorial cumulant.
In this embodiment, described three-dimensional motion vector includes: horizontally rotates vector, vertical rotation vector, Jiao
Away from axle motion-vector；Described accumulated respectively in each three-dimensional dimension, comprising: will be one pre-
If the vector that horizontally rotates of each two field picture of camera lens carries out summation in the horizontal direction and obtains horizontally rotating vector
Cumulant, the vertical rotation vector of each two field picture of one default camera lens is asked in vertical direction
With obtain vertical rotation vector cumulant, by the focal length axle of each two field picture of one default camera lens rotate to
Amount carries out summation on focal length direction of principal axis and obtains horizontally rotating vectorial cumulant；Further to three-dimensional motion vector
It is optimized with three-dimensional motion vector cumulant.
In one embodiment, as Fig. 3, step 101, described current lens obtaining current video
Three-dimensional motion vector cumulant, comprising: step 301, current video is cut into different according to color
Camera lens, for current lens, obtains three-dimensional motion vector cumulant；
As Fig. 4, step 102, whether the described content judging described current lens belongs to a in program category
Class, including step 401-403:
Step 401, by the three-dimensional motion vector cumulant of one current lens, with default class template
One of storehouse default camera lens three-dimensional motion vector cumulant is mated, and draws matching value, one pre-
If the content of camera lens belongs to a class in program category；Judge described matching value whether more than default class threshold value:
If described matching value is more than default class threshold value, carry out step 402；Otherwise carry out step 403；
Step 402 is it is believed that the content of described current lens belongs to a class in program category；
Step 403 is it is believed that the content of described current lens is not belonging to a class in program category.
In this embodiment, current video is cut into different camera lenses according to color, for current lens,
Obtain three-dimensional motion vector cumulant；First current video is cut into different camera lenses, obtains in current lens
Three-dimensional motion vector cumulant, is optimized further to methods described.
In one embodiment, as Fig. 5, further include step 104, by the figure in described Highlight
As being exported as the image in new camera lens；The Highlight of mark in test library is exported, group
Become the featured videos collection of choice specimens.
In this embodiment, the image in described Highlight is exported as the image in new camera lens；Will
The Highlight of mark in test library is exported, and forms the featured videos collection of choice specimens；Save trouble and labor, drops significantly
Low cost of labor.
As Fig. 6, the invention provides a kind of video lens annotation equipment is it is characterised in that include:
Extraction module 601, for extracting video to be tested from test library as current video, obtains current
The three-dimensional motion vector cumulant of one current lens of video；
Sort module 602, for the three-dimensional motion vector cumulant according to one current lens, judges
Whether the content of described current lens belongs to a class in program category, if belonging to a class in program category
Then carry out next step, otherwise do not carry out next step；
Matching module 603, during for belonging to a class in program category when the content of described current lens, will
The three-dimensional motion vector cumulant of one current lens, with a class program in default camera lens template base
The three-dimensional motion vector cumulant of all Highlights of type is mated respectively, draws each matching value；As
One matching value of any of which of really described each matching value is more than default camera lens threshold value then it is assumed that described current mirror
Head is Highlight in a class program and is labeled；Otherwise it is assumed that described current lens are not Highlights.
The beneficial effect of the embodiment of the present invention includes: extracts video to be tested from test library as working as forward sight
Frequently, obtain the three-dimensional motion vector cumulant of current lens of current video；According to one current
The three-dimensional motion vector cumulant of camera lens, judges whether the content of described current lens belongs in program category
A class, if belonging to a class in program category, carries out next step, does not otherwise carry out next step；When described
When the content of current lens belongs to a class in program category, by the three-dimensional motion of one current lens to
Amount cumulant, with default camera lens template base in a class program category a default camera lens three-dimensional motion to
Amount cumulant is mated, and draws matching value；If described matching value be more than default camera lens threshold value then it is assumed that
Described current lens are Highlights and are labeled；Otherwise it is assumed that described current lens are not Highlights；
Highlight in video, save trouble and labor are marked by methods described, greatly reduces cost of labor；And
Judged twice by program category and Highlight, improve the accuracy of mark.
In one embodiment, as Fig. 7, described device, comprising:
Presetting module 604, for by the video of artificial mark, training obtains camera lens template base: pre- for one
If each two field picture of camera lens, select a certain window, extract in described certain window three-dimensional motion to
Amount, then by the three-dimensional motion vector of each two field picture of one default camera lens, in each three-dimensional dimension
Summation is carried out respectively on degree and obtains a default camera lens three-dimensional motion vector cumulant, the window of described certain window
It is Highlight in mouthful, outside window be and Highlight irrelevant contents；
The three-dimensional motion vector cumulant of one default camera lens is added to camera lens template base.
In one embodiment, described three-dimensional motion vector, comprising: horizontally rotate vector, vertical rotation to
Amount, focal length axle motion-vector；
As Fig. 8, described presetting module 604, comprising: accumulation submodule 801, for will one preset
The vector that horizontally rotates of each two field picture of camera lens carries out summation in the horizontal direction to obtain horizontally rotating vector tired
Accumulated amount, the vertical rotation vector of each two field picture of one default camera lens is sued for peace in vertical direction
Obtain vertical rotation vector cumulant, by the focal length axle rotating vector of each two field picture of one default camera lens
Summation is carried out on focal length direction of principal axis obtain horizontally rotating vectorial cumulant.
In one embodiment, as Fig. 9, described matching module 603, including matched sub-block 901, it is used for
Current video is cut into different camera lenses according to color, for current lens, obtain three-dimensional motion to
Amount cumulant；
As Figure 10, described sort module 602, including classification submodule 1001, for will one currently
The three-dimensional motion vector cumulant of camera lens, with one of default class template storehouse default camera lens three-dimensional motion to
Amount cumulant is mated, and draws matching value, the content of one default camera lens belongs in program category
A class；
If described matching value is more than default class threshold value then it is assumed that the content of described current lens belongs to program class
A class in type；Otherwise it is assumed that the content of described current lens is not belonging to a class in program category.
In one embodiment, as Figure 11, described device, comprising:
Output module 605, defeated for carrying out the image in described Highlight as the image in new camera lens
Go out；
The Highlight of mark in test library is exported, is formed the featured videos collection of choice specimens.
Those skilled in the art are it should be appreciated that embodiments of the invention can be provided as method, system or meter
Calculation machine program product.Therefore, the present invention can be using complete hardware embodiment, complete software embodiment or knot
Close the form of the embodiment of software and hardware aspect.And, the present invention can adopt and wherein wrap one or more
Computer-usable storage medium containing computer usable program code (including but not limited to disk memory and
Optical memory etc.) the upper computer program implemented form.
The present invention is to produce with reference to method according to embodiments of the present invention, equipment (system) and computer program
The flow chart of product and/or block diagram are describing.It should be understood that can by computer program instructions flowchart and
/ or block diagram in each flow process and/or the flow process in square frame and flow chart and/or block diagram and/
Or the combination of square frame.These computer program instructions can be provided to general purpose computer, special-purpose computer, embed
The processor of formula datatron or other programmable data processing device is to produce a machine so that passing through to calculate
The instruction of the computing device of machine or other programmable data processing device produces for realizing in flow chart one
The device of the function of specifying in individual flow process or multiple flow process and/or one square frame of block diagram or multiple square frame.
These computer program instructions may be alternatively stored in and computer or other programmable datas can be guided to process and set
So that being stored in this computer-readable memory in the standby computer-readable memory working in a specific way
Instruction produce and include the manufacture of command device, the realization of this command device is in one flow process or multiple of flow chart
The function of specifying in flow process and/or one square frame of block diagram or multiple square frame.
These computer program instructions also can be loaded in computer or other programmable data processing device, makes
Obtain and series of operation steps is executed on computer or other programmable devices to produce computer implemented place
Reason, thus the instruction of execution is provided for realizing in flow chart one on computer or other programmable devices
The step of the function of specifying in flow process or multiple flow process and/or one square frame of block diagram or multiple square frame.
Obviously, those skilled in the art can carry out various changes and modification without deviating from this to the present invention
The spirit and scope of invention.So, if these modifications of the present invention and modification belong to the claims in the present invention
And its within the scope of equivalent technologies, then the present invention is also intended to comprise these changes and modification.
Claims (10)
1. a kind of video lens mask method is it is characterised in that include:
Extract video to be tested as current video from test library, obtain current lens of current video
Three-dimensional motion vector cumulant；
According to the three-dimensional motion vector cumulant of one current lens, judge the content of described current lens
Whether belonging to a class in program category, if belonging to a class in program category, carrying out next step, otherwise
Do not carry out next step；
When belonging to a class in program category when the content of described current lens, by one current lens
All Highlights of a class program category in three-dimensional motion vector cumulant, with default camera lens template base
Three-dimensional motion vector cumulant mated respectively, draw each matching value；If its of described each matching value
In any one matching value be more than default camera lens threshold value then it is assumed that described current lens are the essences in a class program
Color camera lens is simultaneously labeled；Otherwise it is assumed that described current lens are not Highlights.
2. the method for claim 1 is it is characterised in that include:
With the video of artificial mark, train and obtain camera lens template base: for each frame figure of a default camera lens
Picture, selects a certain window, extracts three-dimensional motion vector in described certain window, then by described one
The three-dimensional motion vector of each two field picture of individual default camera lens, carries out accumulating in each three-dimensional dimension respectively
To a default camera lens three-dimensional motion vector cumulant, it is Highlight in the window of described certain window, window
Outside mouthful it is and Highlight irrelevant contents；
The three-dimensional motion vector cumulant of one default camera lens is added to camera lens template base.
3. method as claimed in claim 2 is it is characterised in that described three-dimensional motion vector includes: level
Rotating vector, vertical rotation vector, focal length axle motion-vector；
Described accumulated respectively in each three-dimensional dimension, comprising: one default camera lens is each
The vector that horizontally rotates of two field picture carries out summation in the horizontal direction and obtains horizontally rotating vectorial cumulant, by institute
The vertical rotation vector stating each two field picture of a default camera lens carries out summation in vertical direction and obtains vertically
Rotating vector cumulant, by the focal length axle rotating vector of each two field picture of one default camera lens in focal length axle
Summation is carried out on direction obtain horizontally rotating vectorial cumulant.
4. the method for claim 1 is it is characterised in that one of described acquisition current video is worked as
The three-dimensional motion vector cumulant of front camera lens, comprising: current video is cut into different camera lenses according to color,
For current lens, obtain three-dimensional motion vector cumulant；
Whether the described content judging described current lens belongs to a class in program category, comprising: will be described
The three-dimensional motion vector cumulant of one current lens, with one of default class template storehouse default camera lens three
Maintenance and operation moving vector cumulant is mated, and draws matching value, and the content of one default camera lens belongs to program
A class in type；
If described matching value is more than default class threshold value then it is assumed that the content of described current lens belongs to program class
A class in type；Otherwise it is assumed that the content of described current lens is not belonging to a class in program category.
5. the method for claim 1 is it is characterised in that include:
Image in described Highlight is exported as the image in new camera lens；
The Highlight of mark in test library is exported, is formed the featured videos collection of choice specimens.
6. a kind of video lens annotation equipment is it is characterised in that include:
Extraction module, for extracting video to be tested from test library as current video, obtains current video
Current lens three-dimensional motion vector cumulant；
Sort module, for the three-dimensional motion vector cumulant according to one current lens, judges described
Whether the content of current lens belongs to a class in program category, if belonging to a class in program category, enters
Row next step, does not otherwise carry out next step；
Matching module, during for belonging to a class in program category when the content of described current lens, will be described
The three-dimensional motion vector cumulant of one current lens, with a class program category in default camera lens template base
All Highlights three-dimensional motion vector cumulant mated respectively, draw each matching value；If institute
One matching value of any of which stating each matching value is more than default camera lens threshold value then it is assumed that described current lens are
Highlight in a class program is simultaneously labeled；Otherwise it is assumed that described current lens are not Highlights.
7. device as claimed in claim 6 is it is characterised in that described device, comprising:
Presetting module, for by the video of artificial mark, training obtains camera lens template base: default for one
Each two field picture of camera lens, selects a certain window, extracts three-dimensional motion vector in described certain window,
Then by the three-dimensional motion vector of each two field picture of one default camera lens, in each three-dimensional dimension
Carry out summation respectively and obtain a default camera lens three-dimensional motion vector cumulant, in the window of described certain window
It is Highlight, outside window be and Highlight irrelevant contents；
The three-dimensional motion vector cumulant of one default camera lens is added to camera lens template base.
8. device as claimed in claim 7 is it is characterised in that described three-dimensional motion is vectorial, comprising: water
Flat turn moving vector, vertical rotation vector, focal length axle motion-vector；
Described presetting module, comprising: accumulation submodule, for by each two field picture of one default camera lens
The vector that horizontally rotates carry out summation in the horizontal direction and obtain horizontally rotating vectorial cumulant, will be one
The vertical rotation vector of each two field picture of default camera lens carry out in vertical direction summation obtain vertical rotation to
Amount cumulant, by the focal length axle rotating vector of each two field picture of one default camera lens on focal length direction of principal axis
Carry out summation to obtain horizontally rotating vectorial cumulant.
9. device as claimed in claim 6 is it is characterised in that described matching module, sub including coupling
Module, for current video is cut into different camera lenses according to color, for current lens, obtains
Three-dimensional motion vector cumulant；
Described sort module, including classification submodule, for by the three-dimensional motion of one current lens to
Amount cumulant, is carried out with one of default class template storehouse default camera lens three-dimensional motion vector cumulant
Join, draw matching value, the content of one default camera lens belongs to a class in program category；
If described matching value is more than default class threshold value then it is assumed that the content of described current lens belongs to program class
A class in type；Otherwise it is assumed that the content of described current lens is not belonging to a class in program category.
10. device as claimed in claim 6 is it is characterised in that described device, comprising:
Output module, for being exported the image in described Highlight as the image in new camera lens；
The Highlight of mark in test library is exported, is formed the featured videos collection of choice specimens.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510392380.2A CN106339655A (en) | 2015-07-06 | 2015-07-06 | Video shot marking method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510392380.2A CN106339655A (en) | 2015-07-06 | 2015-07-06 | Video shot marking method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106339655A true CN106339655A (en) | 2017-01-18 |
Family
ID=57826301
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510392380.2A Pending CN106339655A (en) | 2015-07-06 | 2015-07-06 | Video shot marking method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106339655A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108924576A (en) * | 2018-07-10 | 2018-11-30 | 武汉斗鱼网络科技有限公司 | A kind of video labeling method, device, equipment and medium |
| CN109040773A (en) * | 2018-07-10 | 2018-12-18 | 武汉斗鱼网络科技有限公司 | A kind of video improvement method, apparatus, equipment and medium |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101013444A (en) * | 2007-02-13 | 2007-08-08 | 华为技术有限公司 | Method and apparatus for adaptively generating abstract of football video |
| CN101018347A (en) * | 2006-02-09 | 2007-08-15 | 智辉研发股份有限公司 | Apparatus for detecting highlights of media stream and related method |
| CN101127866A (en) * | 2007-08-10 | 2008-02-20 | 西安交通大学 | A method for detecting wonderful section of football match video |
| CN101201822A (en) * | 2006-12-11 | 2008-06-18 | 南京理工大学 | Method for searching visual lens based on contents |
| CN101420579A (en) * | 2007-10-22 | 2009-04-29 | 皇家飞利浦电子股份有限公司 | Method, apparatus and system for detecting exciting part |
| CN101479767A (en) * | 2006-06-30 | 2009-07-08 | Nxp股份有限公司 | A method and device for video stitching |
| CN101599179A (en) * | 2009-07-17 | 2009-12-09 | 北京邮电大学 | Method for automatically generating field motion wonderful scene highlights |
| CN102930553A (en) * | 2011-08-10 | 2013-02-13 | 中国移动通信集团上海有限公司 | Method and device for identifying objectionable video content |
| CN103065300A (en) * | 2012-12-24 | 2013-04-24 | 安科智慧城市技术（中国）有限公司 | Method for video labeling and device for video labeling |
-
2015
- 2015-07-06 CN CN201510392380.2A patent/CN106339655A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101018347A (en) * | 2006-02-09 | 2007-08-15 | 智辉研发股份有限公司 | Apparatus for detecting highlights of media stream and related method |
| CN101479767A (en) * | 2006-06-30 | 2009-07-08 | Nxp股份有限公司 | A method and device for video stitching |
| CN101201822A (en) * | 2006-12-11 | 2008-06-18 | 南京理工大学 | Method for searching visual lens based on contents |
| CN101013444A (en) * | 2007-02-13 | 2007-08-08 | 华为技术有限公司 | Method and apparatus for adaptively generating abstract of football video |
| CN101127866A (en) * | 2007-08-10 | 2008-02-20 | 西安交通大学 | A method for detecting wonderful section of football match video |
| CN101420579A (en) * | 2007-10-22 | 2009-04-29 | 皇家飞利浦电子股份有限公司 | Method, apparatus and system for detecting exciting part |
| CN101599179A (en) * | 2009-07-17 | 2009-12-09 | 北京邮电大学 | Method for automatically generating field motion wonderful scene highlights |
| CN102930553A (en) * | 2011-08-10 | 2013-02-13 | 中国移动通信集团上海有限公司 | Method and device for identifying objectionable video content |
| CN103065300A (en) * | 2012-12-24 | 2013-04-24 | 安科智慧城市技术（中国）有限公司 | Method for video labeling and device for video labeling |
Non-Patent Citations (1)
| Title |
|---|
| 李秀强: "视频镜头边界检测与体育视频分类算法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108924576A (en) * | 2018-07-10 | 2018-11-30 | 武汉斗鱼网络科技有限公司 | A kind of video labeling method, device, equipment and medium |
| CN109040773A (en) * | 2018-07-10 | 2018-12-18 | 武汉斗鱼网络科技有限公司 | A kind of video improvement method, apparatus, equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Huang et al. | Tracknet: A deep learning network for tracking high-speed and tiny objects in sports applications | |
| Parmar et al. | What and how well you performed? a multitask learning approach to action quality assessment | |
| US11544928B2 (en) | Athlete style recognition system and method | |
| Ma et al. | Action recognition and localization by hierarchical space-time segments | |
| Dai et al. | Temporal context network for activity localization in videos | |
| CN105022982B (en) | Hand motion recognition method and apparatus | |
| Theagarajan et al. | Soccer: Who has the ball? Generating visual analytics and player statistics | |
| KR102106135B1 (en) | Apparatus and method for providing application service by using action recognition | |
| CN107730458A (en) | A kind of fuzzy facial reconstruction method and system based on production confrontation network | |
| Gourgari et al. | Thetis: Three dimensional tennis shots a human action dataset | |
| Rangasamy et al. | Deep learning in sport video analysis: a review | |
| CN109886165A (en) | A kind of action video extraction and classification method based on moving object detection | |
| CN108875730A (en) | A kind of deep learning sample collection method, apparatus, equipment and storage medium | |
| Karaman et al. | Human daily activities indexing in videos from wearable cameras for monitoring of patients with dementia diseases | |
| Baby et al. | Dynamic vision sensors for human activity recognition | |
| CN111160134A (en) | Human-subject video scene analysis method and device | |
| CN109961039A (en) | A kind of individual's goal video method for catching and system | |
| Dubois et al. | Using image processing for biomechanics measures in swimming | |
| Fani et al. | Swim stroke analytic: Front crawl pulling pose classification | |
| CN106326910A (en) | Video shot classified labeling method and video shot classified labeling device | |
| CN107169503B (en) | Indoor scene classification method and device | |
| CN106339655A (en) | Video shot marking method and device | |
| CN114618142A (en) | Auxiliary training system and method for table tennis sports | |
| CN109460724B (en) | Object detection-based separation method and system for ball-stopping event | |
| CN106599779A (en) | Human ear recognition method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170118 |
|
| RJ01 | Rejection of invention patent application after publication |