CN106156700A - The extracting method of characteristics of human body and device in a kind of image procossing - Google Patents
The extracting method of characteristics of human body and device in a kind of image procossing Download PDFInfo
- Publication number
- CN106156700A CN106156700A CN201510149262.9A CN201510149262A CN106156700A CN 106156700 A CN106156700 A CN 106156700A CN 201510149262 A CN201510149262 A CN 201510149262A CN 106156700 A CN106156700 A CN 106156700A
- Authority
- CN
- China
- Prior art keywords
- point
- probability
- center
- gravity
- human body
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Abstract
The present invention relates to extracting method and the device of characteristics of human body in a kind of image procossing, the method includes: in body image, for the human body of central point to be extracted, concentrates from the basic point of this human body, chooses growth point set;Any point concentrated for growing point, when this point meets pre-conditioned, with this point as seed points, determines the seed region of this seed points;For each seed region, calculate the center of gravity of this seed region;For each center of gravity, calculate the prediction probability of the central point that this center of gravity is human body;According to the prediction probability of each center of gravity calculating acquisition, determine the central point of human body.The method provided by the present invention can improve the accuracy extracting central point, and meets the speed requirement at real time processed images extraction center.
Description
Technical field
The present invention relates to image processing field, particularly relate to the extracting method of characteristics of human body in a kind of image procossing
And device.
Background technology
Body-sensing technology, is not only applicable to field of play, it may also be used for security, health medical treatment, amusement are purchased
The fields such as thing.In a word, the development of body-sensing technology, brings revolutionary change to man-machine interaction mode.
Body-sensing technology manipulates somatosensory device by the motion characteristic of identification human body, and in identification maneuver feature
During, need to extract the central point at each position of human body.Such as, in optics senses, it is necessary first to know
Do not go out human body, subdivided go out human body each position (such as head, trunk, ancon, hand etc.), so
After extract from each position central point as identify motion feature parameter.As can be seen here, human body
It is a key technology in body-sensing technology that central point extracts, and the accuracy that central point extracts directly affects dynamic
Make the accuracy of feature identification.And body-sensing technology be somatosensory device manipulation personnel by limb action manipulation set
Standby, then the extraction rate for position central point just requires.
At present in body-sensing technology, it is possible to meet the mainly type heart method of real-time processing speed demand.Type should
The process of method is to liking body image, and the method will belong to each pixel at same position in body image
The coordinate of point is added, then averages, then using the meansigma methods tried to achieve as the central point at this position.
But, type heart method extracts the accuracy of central point, is severely limited by the accuracy of human body classification.
This is that in the human body sorting technique based on probability owing to using at present, the classification at position is inaccurate.
Generally, a position is made up of the pixel that the probit belonging to this position is not 0, and in practice, one
Individual pixel tends to belong to several positions, therefore, boundary the clearest and the most definite between position and position.And
Type heart method does not accounts for position and classifies this factor inaccurate, and simple again is the most only single with the coordinate of pixel
Parameter, extract central point, cause the accuracy of the extraction result of central point to have much room for improvement.
Summary of the invention
It is an object of the invention to provide extracting method and the device of characteristics of human body in a kind of image procossing, to overcome
In correlation technique, human body central point extracts the problem that accuracy is low.
On the one hand, the present invention provides the extracting method of characteristics of human body in a kind of image procossing, including:
For the body image obtained in advance, each pixel calculated in described body image belongs to each individual
The probability of body region;
For the human body of arbitrary central point to be extracted, by belonging to the probability of described human body more than first
The point of predetermined threshold value forms the basic point set of described human body;
Concentrating from described basic point, acquisition probability, more than the point of the second predetermined threshold value, forms growth point set, its
Described in the second predetermined threshold value more than described first predetermined threshold value;
Any point concentrated for described growing point, when this point meets pre-conditioned, the following operation of execution:
With this point as seed points, obtain the seed region corresponding with this seed points by region-growing method;Described default
Condition includes: be not selected as seed points and not in the seed region of any one sub-point;
For each seed region of described human body, by the probability of each point in this seed region
It is considered as the density of this point and according to the density of each point in this seed region, and each point is described
Position coordinates in body image, calculates the center of gravity of this seed region;
For the center of gravity of each seed region, concentrate little with the pixel distance of this center of gravity according to described basic point
In the probability of the point equal to the first predeterminable range, calculate the prediction of the central point that this center of gravity is described human body
Probability;
Basic point is concentrated the point that the center of gravity of range prediction maximum probability is nearest, is defined as described human body
Central point.
On the other hand, the present invention also provides for the extraction element of characteristics of human body in a kind of image procossing, described device
Including:
Probability evaluation entity, for for the body image obtained in advance, calculate in described body image is every
Individual pixel belongs to the probability of each human body;
Basis point set acquisition module, for the human body for arbitrary central point to be extracted, described by belonging to
The probability of human body forms the basic point set of described human body more than the point of the first predetermined threshold value;
Growth point set determines module, and for concentrating from described basic point, acquisition probability is more than the second predetermined threshold value
Point, formed growth point set, wherein said second predetermined threshold value be more than described first predetermined threshold value;
Seed region forms module, for any point concentrated for described growing point, presets when this point meets
During condition, perform following operation: with this point as seed points, obtained and this seed points pair by region-growing method
The seed region answered;Described pre-conditioned include: be not selected as seed points and not at any one sub-point
Seed region in;
Center of gravity calculation module, for each seed region for described human body, by this seed region
In each point probability be considered as this point density and according in this seed region each point density,
And each some position coordinates in described body image, calculate the center of gravity of this seed region;
Prediction probability computing module, for the center of gravity for each seed region, according to described basis point set
In with the pixel distance of this center of gravity less than or equal to the probability of the point of the first predeterminable range, it is described for calculating this center of gravity
The prediction probability of the central point of human body;
Central point determines module, the point that center of gravity for basic point is concentrated range prediction maximum probability is nearest,
It is defined as the central point of described human body.
The present invention at least has the advantages that the embodiment of the present invention, introduces and belongs to central point to be extracted
The probability of human body is as the parameter of extraction central point, and will belong to the dot-dash of human body by seed points
Divide to different seed regions, then according to seed region, draw the center of gravity of each seed region, further root
The probability put according to the surrounding of center of gravity, determines most possibly close to the center of gravity of central point so that final acquisition
Central point is closely connected with probability, thus the human body that reduction position based on probability sorting technique is got is not
Accurately on the impact extracting central point, improve the accuracy extracting central point, and the extraction center of the present invention
The method of point is easy, it is possible to meet the processing speed demand processed in real time.
It should be appreciated that it is only exemplary and explanatory that above general description and details hereinafter describe
, the present invention can not be limited.
Accompanying drawing explanation
Fig. 1 be in the embodiment of the present invention in image procossing the exemplary process diagram of the extracting method of characteristics of human body it
One;
Fig. 2 be in the embodiment of the present invention divide preset range show schematic diagram;
Fig. 3 be in the embodiment of the present invention in image procossing the exemplary process diagram of the extracting method of characteristics of human body it
Two;
Fig. 4 is one of schematic diagram of extraction element of characteristics of human body in image procossing in the embodiment of the present invention;
Fig. 5 is in the embodiment of the present invention in image procossing the two of the schematic diagram of the extraction element of characteristics of human body.
Detailed description of the invention
Below in conjunction with accompanying drawing, embodiments of the invention are illustrated, it will be appreciated that enforcement described herein
Example is merely to illustrate and explains the present invention, is not intended to limit the present invention.
Here will illustrate exemplary embodiment in detail, its example represents in the accompanying drawings.Following retouches
Stating when relating to accompanying drawing, unless otherwise indicated, the same numbers in different accompanying drawings represents same or analogous key element.
Embodiment described in following exemplary embodiment does not represent all enforcements consistent with the present invention
Mode.On the contrary, they only with describe in detail in appended claims, the present invention some in terms of phase
The example of consistent apparatus and method.
Although it is fast for prior art medium-sized heart method processing speed, it is possible to meet the demand of real-time processing speed,
But the problem that the central point accuracy of the human body that the method is extracted has much room for improvement, the embodiment of the present invention provides
The extracting method of characteristics of human body in a kind of image procossing, for extracting the central point of human body, the method phase
Type heart method can not only be improved the accuracy extracting central point, additionally it is possible to meet real time handling requirement and process speed
Spend fast demand.
First, the extracting method of characteristics of human body in the image procossing that the embodiment of the present invention provides, by body image
In the probability of the point human body that belongs to central point to be extracted take into account, be used for determining human body center
The position of point such that it is able to reduce the inaccurate impact on extracting central point of position classification such that it is able to improve
Extract the accuracy of the central point of human body.
Additionally, in the embodiment of the present invention, before extract the central point of human body according to probability, first to genus
Point in this human body carries out pretreatment, such as, filter the little point of probability and isolated point, joined by minimizing
With the quantity of the point extracting central point, reduce amount of calculation, it is possible to increase extract the speed of central point, by filter
Except the point that the probability belonging to human body is little can also improve the accuracy of extraction central point.
Below by simple embodiment, to the extraction side of characteristics of human body in image procossing in the embodiment of the present invention
Method is described in detail.
Embodiment one
As it is shown in figure 1, the showing of extracting method of characteristics of human body in the image procossing provided for the embodiment of the present invention
Example flow chart, the method comprises the following steps:
Step 101: for the body image obtained in advance, calculate each pixel in described body image
Belong to the probability of each human body.
Step 102: for the human body of arbitrary central point to be extracted, by belonging to the general of described human body
Rate forms the basic point set of described human body more than the point of the first predetermined threshold value.
Step 103: concentrate from described basic point, acquisition probability, more than the point of the second predetermined threshold value, is formed raw
Long point set, wherein said second predetermined threshold value is more than described first predetermined threshold value.
Step 104: any point concentrated for described growing point, when this point meets pre-conditioned, performs
Below operation: with this point as seed points, obtains the seed region corresponding with this seed points by region-growing method;
Described pre-conditioned include: be not selected as seed points and not in the seed region of any one sub-point.
Step 105: for each seed region of described human body, each by this seed region
The probability of individual point is considered as the density of this point and according to density of each point in this seed region, and each
Individual some position coordinates in described body image, calculates the center of gravity of this seed region.
Wherein, in one embodiment can be according to the center of gravity of equation below (1) calculating seed region:
Wherein, in formula (1), XiRepresent the abscissa of seed region i-th point;YiRepresent seed zone
The vertical coordinate of territory i-th point;GiRepresent the probability of seed region i-th point;Xc represents the weight of seed region
The abscissa of the heart;Yc represents the vertical coordinate of the center of gravity of seed region;G represent seed region a little general
Rate sum;* represent and seek product.
Step 106: for the center of gravity of each seed region, concentrate and this center of gravity according to described basic point
Pixel distance is less than or equal to the probability of the point of the first predeterminable range, and calculating this center of gravity is in described human body
The prediction probability of heart point.
Step 107: basic point is concentrated the point that the center of gravity of range prediction maximum probability is nearest, is defined as described
The central point of human body.
Below, above steps is described in detail:
1), wherein, step 104 obtains the seed region corresponding with this seed points by region-growing method,
Tool can specifically perform: concentrates from described basic point, and acquisition probability is more than the 3rd predetermined threshold value and with described
The pixel distance of seed points, less than the point of the second predeterminable range, forms seed region;Wherein, described 3rd pre-
If threshold value is less than described second predetermined threshold value.
Wherein, for ease of understanding the forming process of seed region, below by step A1 to step A2 pair
This process is illustrated:
Step A1: after obtaining growth point set, concentrate the most a bit from growing point, as seed points, subordinate
Basic point in described human body is concentrated, and acquisition probability is more than the 3rd predetermined threshold value and with at described human body
On the basis of position in image and the pixel distance of seed points is less than the point of the second predeterminable range, form seed
The seed region of point.
Step A2: then for growing point concentrate any point, when this point be not selected as seed points,
And not when selecting in the seed region doing seed points, using this point as seed points, it is subordinated to described people's body
The basic point of position is concentrated, and acquisition probability is more than the 3rd predetermined threshold value and with the position in described body image
On the basis of and the pixel distance of seed points less than the point of the second predeterminable range, form the seed region of seed points.
Wherein, it be not elected to be into seed points and not at the seed zone of any one sub-point when seed points concentration does not exist
During point in territory, the division of seed region just completes.
Wherein, on the basis of the position in described body image and the pixel distance of seed points is less than second
The point of predeterminable range is e.g.: according to seed points position in body image, concentrates from basic point, searches
To the point that the pixel distance with seed points is presetted pixel point.
2), wherein step 106 can specifically perform as following steps:
Step B1: for the center of gravity of each seed region, concentrates from described basic point and obtains and this center of gravity
Pixel distance less than or equal to the probability of point of the first predeterminable range.
Step B2: the pixel distance of calculating and this center of gravity is less than or equal to the probability of the point of the first predeterminable range
With, using that calculate and as this center of gravity prediction probability.
Wherein, using probability and value as the prediction probability of center of gravity, it is possible to calculating this center of gravity simply and easily is
The probability size of the central point of human body, improves and calculates speed.
Wherein, naturally it is also possible to use other method to calculate the prediction probability of each center of gravity, such as, when default model
Enclose for time centered by center of gravity, with border circular areas as radius of the distance of n (n is as positive integer) individual pixel,
This border circular areas is divided into according to the distance of distance center of gravity the circular annular region of predetermined number.Such as, such as figure
Shown in 2, when n is 3, by the first predeterminable range according to centered by center of gravity O, respectively with 1 pixel
The circle that distance is radius of point, 2 pixels and 3 pixels, divides the scope of the first predeterminable range
Going out 3 circular annular region, these three circular annular region is respectively region 1, region 2 (logos in Fig. 2
Circular annular region) and region 3.Then arranging weight for each region, wherein, region 1 represents distance center of gravity
Recently, then the maximum weight that this region is corresponding, region 3 is farthest apart from center of gravity, then the weights in region 3 are minimum.
To any point in preset range, the probability that this point belongs to human body is multiplied by the weights of this region
As the new probability of this point, then using the new probability of each point in preset range and prediction as center of gravity general
Rate.Owing to the distance the nearlyest weight of center of gravity is the biggest, the effect of the probability of pericentral point so can be increased.
It is of course also possible to the prediction probability of the central point using other method calculating center of gravity to be human body, the present invention
This is not limited.
3), wherein, in step 107, when putting the point of concentration based on the center of gravity that prediction probability is maximum, base
The point that in plinth point set, this center of gravity of distance is nearest is i.e. center of gravity itself.I.e. step 107 includes by prediction probability
The situation of the center of gravity of point set based on big center of gravity, thus, by step 107, the final center obtained
Point can be on human body.Such that make the position of position based on probability sorting technique segmentation be forbidden
Really (even if such as due to environment or the impact of clothing, a position is divided at least two parts, or
Owing to being affected by action, when cavity occurring in the middle of a position), the central point determined according to step 107
Can fall on human body, thus improve the accuracy determining central point.
4), in order to simplify calculating, and the basic point set for extracting central point is optimized, in the embodiment of the present invention,
Before step 103 forms growth point set, it is also possible to filter described basic point and concentrate, with described human body
Other put non-conterminous point.So, it is possible not only to reduce for the data volume extracting central point, additionally it is possible to
Optimize the basic data for extracting central point, thus improve the accuracy extracting central point.
To sum up, the embodiment of the present invention, introduce the probability of the human body belonging to central point to be extracted as extraction
The parameter of central point, and by seed points, the point belonging to human body is divided to different seed region, then
According to seed region, draw the center of gravity of each seed region, the probability put according further to the surrounding of center of gravity,
Determine most possibly close to the center of gravity of central point so that the final central point obtained is closely connected with probability,
Thus reduce the inaccurate shadow to extracting central point of human body that position based on probability sorting technique is got
Ring, improve the accuracy extracting central point, and the method extracting central point of the present invention is easy, it is possible to be full
The processing speed demand processed time full.
Embodiment two
As it is shown on figure 3, as a example by the central point extracting head, at the image provided in the embodiment of the present invention
In reason, the extracting method of characteristics of human body illustrates, and the method comprises the following steps:
Step 301: in the body image obtained in advance, it is not the point of 0 that acquisition belongs to the probability of head,
Form the initial point set of head.
Step 302: filter initial point and concentrate, probability, and filters equal to the point of the first predetermined threshold value less than the
With other non-conterminous point of point of head, basis of formation point set.
Step 303: concentrate from described basic point, acquisition probability, more than the point of the second predetermined threshold value, is formed raw
Long point set.
Step 304: any point concentrated for described growing point, when this point meets pre-conditioned, performs
Below operation: with this point as seed points, obtains the seed region corresponding with this seed points by region-growing method;
Described pre-conditioned include: be not selected as seed points and not in the seed region of any one sub-point.
Step 305: for each seed region of described head, by each point in this seed region
Probability be considered as the density of this point and according to density of each point, and each point in this seed region
Position coordinates in described body image, calculates the center of gravity of this seed region.
Step 306: for the center of gravity of each seed region, concentrates from described basic point and obtains and this center of gravity
Pixel distance less than or equal to the probability of point of the first predeterminable range, and calculate the sum of the probability of the point of acquisition,
Using that calculate and as the central point that this center of gravity is head prediction probability.
Step 307: basic point is concentrated the point that the center of gravity of range prediction maximum probability is nearest, is defined as head
Central point.
The embodiment of the present invention, belongs to the probability of head by basis, and the position coordinates belonging to the point of head carries
Take out the central point of head such that it is able to reduce the inaccurate impact on extracting central point of position classification, improve
Extract the accuracy of central point.
Embodiment three
Based on identical inventive concept, the embodiment of the present invention also provides for carrying of characteristics of human body in a kind of image procossing
Fetching is put, and as shown in Figure 4, described device includes:
Probability evaluation entity 401, for for the body image obtained in advance, calculates in described body image
Each pixel belong to the probability of each human body;
Basis point set acquisition module 402, for the human body for arbitrary central point to be extracted, by belonging to
The probability of described human body forms the basic point set of described human body more than the point of the first predetermined threshold value;
Growth point set determines module 403, and for concentrating from described basic point, acquisition probability is preset more than second
The point of threshold value, forms growth point set, and wherein said second predetermined threshold value is more than described first predetermined threshold value;
Seed region forms module 404, for any point concentrated for described growing point, when this point meets
Time pre-conditioned, perform following operation: with this point as seed points, obtained and this seed by region-growing method
The seed region that point is corresponding;Described pre-conditioned include: be not selected as seed points and not at any one
In the seed region of son point;
Center of gravity calculation module 405, for each seed region for described human body, by this seed
In region each point probability be considered as this point density and according in this seed region each point close
Degree, and each some position coordinates in described body image, calculate the center of gravity of this seed region;
Prediction probability computing module 406, for the center of gravity for each seed region, according to described basis
Point concentrates the probability less than or equal to the point of the first predeterminable range of the pixel distance with this center of gravity, and calculating this center of gravity is
The prediction probability of the central point of described human body;
Central point determines module 407, nearest for the center of gravity that basic point is concentrated range prediction maximum probability
Point, is defined as the central point of described human body.
Wherein, in one embodiment, described seed region forms module, specifically for from described basic point
Concentrating, the acquisition probability pixel distance more than the 3rd predetermined threshold value and with described seed points is preset less than second
The point of distance, forms seed region;Wherein, described 3rd predetermined threshold value is less than described second predetermined threshold value.
Wherein, in one embodiment, as it is shown in figure 5, described device also includes:
Filter module 408, determine that module is concentrated from described basic point for described growth point set, acquisition probability
More than the point of the second predetermined threshold value, before forming growth point set, filter described basic point and concentrate, with described people
Other of body region puts non-conterminous point.
Wherein, in one embodiment, as it is shown in figure 5, described prediction probability computing module 406, specifically
Including:
Probability acquiring unit 409, for the center of gravity for each seed region, concentrates from described basic point
Obtain the probability less than or equal to the point of the first predeterminable range of the pixel distance with this center of gravity;
Prediction probability computing unit 410, presets less than or equal to first for calculating the pixel distance with this center of gravity
The sum of the probability of the point of distance, using that calculate and as this center of gravity prediction probability.
About the device in above-described embodiment, wherein modules performs the concrete mode of operation relevant
The embodiment of the method is described in detail, explanation will be not set forth in detail herein.
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 use complete hardware embodiment, complete software implementation or knot
The form of the embodiment in terms of conjunction software and hardware.And, the present invention can use and wherein wrap one or more
Computer-usable storage medium containing computer usable program code (include but not limited to disk memory,
CD-ROM, optical memory etc.) form of the upper computer program implemented.
The present invention is with reference to method, equipment (system) and computer program product according to embodiments of the present invention
The flow chart of product and/or block diagram describe.It should be understood that can by computer program instructions flowchart and
/ or block diagram in each flow process and/or 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, embedding
The processor of formula datatron or other programmable data processing device is to produce a machine so that by calculating
The instruction that the processor of machine or other programmable data processing device performs produces for realizing at flow chart one
The device of the function specified 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 the process of other programmable datas can be guided to set
In the standby computer-readable memory worked in a specific way so that be stored in this computer-readable memory
Instruction produce and include the manufacture of command device, this command device realizes in one flow process or multiple of flow chart
The function specified 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
Sequence of operations step must be performed to produce computer implemented place on computer or other programmable devices
Reason, thus the instruction performed on computer or other programmable devices provides for realizing flow chart one
The step of the function specified in flow process or multiple flow process and/or one square frame of block diagram or multiple square frame.
Although preferred embodiments of the present invention have been described, but those skilled in the art once know base
This creativeness concept, then can make other change and amendment to these embodiments.So, appended right is wanted
Ask and be intended to be construed to include preferred embodiment and fall into all changes and the amendment of the scope of the invention.
Obviously, those skilled in the art can carry out various change and modification without deviating from this to the present invention
Bright spirit and scope.So, if the present invention these amendment and modification belong to the claims in the present invention and
Within the scope of its equivalent technologies, then the present invention is also intended to comprise these change and modification.
Claims (8)
1. the extracting method of characteristics of human body in an image procossing, it is characterised in that described method includes:
For the body image obtained in advance, each pixel calculated in described body image belongs to each individual
The probability of body region;
For the human body of arbitrary central point to be extracted, by belonging to the probability of described human body more than first
The point of predetermined threshold value forms the basic point set of described human body;
Concentrating from described basic point, acquisition probability, more than the point of the second predetermined threshold value, forms growth point set, its
Described in the second predetermined threshold value more than described first predetermined threshold value;
Any point concentrated for described growing point, when this point meets pre-conditioned, the following operation of execution:
With this point as seed points, obtain the seed region corresponding with this seed points by region-growing method;Described default
Condition includes: be not selected as seed points and not in the seed region of any one sub-point;
For each seed region of described human body, by the probability of each point in this seed region
It is considered as the density of this point and according to the density of each point in this seed region, and each point is described
Position coordinates in body image, calculates the center of gravity of this seed region;
For the center of gravity of each seed region, concentrate little with the pixel distance of this center of gravity according to described basic point
In the probability of the point equal to the first predeterminable range, calculate the prediction of the central point that this center of gravity is described human body
Probability;
Basic point is concentrated the point that the center of gravity of range prediction maximum probability is nearest, is defined as described human body
Central point.
Method the most according to claim 1, it is characterised in that described obtained by region-growing method
The seed region corresponding with this seed points, specifically includes:
Concentrating from described basic point, acquisition probability is more than the 3rd predetermined threshold value and the pixel with described seed points
Distance, less than the point of the second predeterminable range, forms seed region;Wherein, described 3rd predetermined threshold value is less than institute
State the second predetermined threshold value.
Method the most according to claim 2, it is characterised in that described concentrate from described basic point,
Acquisition probability is more than the point of the second predetermined threshold value, and before forming growth point set, described method also includes:
Filter described basic point to concentrate, with other non-conterminous point of point of described human body.
Method the most according to claim 1, it is characterised in that described for each seed region
Center of gravity, concentrate with the pixel distance of this center of gravity less than or equal to the point of the first predeterminable range according to described basic point
Probability, calculate the prediction probability of the central point that this center of gravity is described human body, specifically include:
For the center of gravity of each seed region, concentrate the pixel distance obtained with this center of gravity from described basic point
Probability less than or equal to the point of the first predeterminable range;And,
Calculate with the pixel distance of this center of gravity less than or equal to the point of the first predeterminable range probability and, will calculating
And prediction probability as this center of gravity.
5. the extraction element of characteristics of human body in an image procossing, it is characterised in that described device includes:
Probability evaluation entity, for for the body image obtained in advance, calculate in described body image is every
Individual pixel belongs to the probability of each human body;
Basis point set acquisition module, for the human body for arbitrary central point to be extracted, described by belonging to
The probability of human body forms the basic point set of described human body more than the point of the first predetermined threshold value;
Growth point set determines module, and for concentrating from described basic point, acquisition probability is more than the second predetermined threshold value
Point, formed growth point set, wherein said second predetermined threshold value be more than described first predetermined threshold value;
Seed region forms module, for any point concentrated for described growing point, presets when this point meets
During condition, perform following operation: with this point as seed points, obtained and this seed points pair by region-growing method
The seed region answered;Described pre-conditioned include: be not selected as seed points and not at any one sub-point
Seed region in;
Center of gravity calculation module, for each seed region for described human body, by this seed region
In each point probability be considered as this point density and according in this seed region each point density,
And each some position coordinates in described body image, calculate the center of gravity of this seed region;
Prediction probability computing module, for the center of gravity for each seed region, according to described basis point set
In with the pixel distance of this center of gravity less than or equal to the probability of the point of the first predeterminable range, it is described for calculating this center of gravity
The prediction probability of the central point of human body;
Central point determines module, the point that center of gravity for basic point is concentrated range prediction maximum probability is nearest,
It is defined as the central point of described human body.
Device the most according to claim 5, it is characterised in that described seed region forms module,
Specifically for concentrating from described basic point, acquisition probability more than the 3rd predetermined threshold value and with described seed points
Pixel distance, less than the point of the second predeterminable range, forms seed region;Wherein, described 3rd predetermined threshold value is little
In described second predetermined threshold value.
Device the most according to claim 6, it is characterised in that described device also includes:
Filtering module, determine that module is concentrated from described basic point for described growth point set, acquisition probability is more than
The point of the second predetermined threshold value, before forming growth point set, filters described basic point and concentrates, with described people's body
Other of position puts non-conterminous point.
Device the most according to claim 5, it is characterised in that described prediction probability computing module,
Specifically include:
Probability acquiring unit, for the center of gravity for each seed region, concentrates from described basic point and obtains
The probability of the point of the first predeterminable range it is less than or equal to the pixel distance of this center of gravity;
Prediction probability computing unit, for calculating the pixel distance with this center of gravity less than or equal to the first predeterminable range
The sum of probability of point, using that calculate and as this center of gravity prediction probability.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510149262.9A CN106156700B (en) | 2015-03-31 | 2015-03-31 | The extracting method and device of characteristics of human body in a kind of image procossing |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510149262.9A CN106156700B (en) | 2015-03-31 | 2015-03-31 | The extracting method and device of characteristics of human body in a kind of image procossing |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106156700A true CN106156700A (en) | 2016-11-23 |
CN106156700B CN106156700B (en) | 2019-06-18 |
Family
ID=57338089
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510149262.9A Active CN106156700B (en) | 2015-03-31 | 2015-03-31 | The extracting method and device of characteristics of human body in a kind of image procossing |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106156700B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
GB2575153A (en) * | 2018-06-28 | 2020-01-01 | Adobe Inc | Determining image handle locations |
US10621764B2 (en) | 2018-07-05 | 2020-04-14 | Adobe Inc. | Colorizing vector graphic objects |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101251898A (en) * | 2008-03-25 | 2008-08-27 | 腾讯科技(深圳)有限公司 | Skin color detection method and apparatus |
CN102005034A (en) * | 2010-12-01 | 2011-04-06 | 南京大学 | Remote sensing image segmentation method based on region clustering |
JP2012003358A (en) * | 2010-06-15 | 2012-01-05 | Yahoo Japan Corp | Background determination device, method, and program |
CN103093215A (en) * | 2013-02-01 | 2013-05-08 | 北京天诚盛业科技有限公司 | Eye location method and device |
-
2015
- 2015-03-31 CN CN201510149262.9A patent/CN106156700B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101251898A (en) * | 2008-03-25 | 2008-08-27 | 腾讯科技(深圳)有限公司 | Skin color detection method and apparatus |
JP2012003358A (en) * | 2010-06-15 | 2012-01-05 | Yahoo Japan Corp | Background determination device, method, and program |
CN102005034A (en) * | 2010-12-01 | 2011-04-06 | 南京大学 | Remote sensing image segmentation method based on region clustering |
CN103093215A (en) * | 2013-02-01 | 2013-05-08 | 北京天诚盛业科技有限公司 | Eye location method and device |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
GB2575153A (en) * | 2018-06-28 | 2020-01-01 | Adobe Inc | Determining image handle locations |
CN110660117A (en) * | 2018-06-28 | 2020-01-07 | 奥多比公司 | Determining position of image control key |
US10672174B2 (en) | 2018-06-28 | 2020-06-02 | Adobe Inc. | Determining image handle locations |
AU2019202799B2 (en) * | 2018-06-28 | 2021-12-23 | Adobe Inc. | A method to automatically predict handle locations for skinningtransformations using deep learning |
GB2575153B (en) * | 2018-06-28 | 2022-01-12 | Adobe Inc | Determining image handle locations |
US11302053B2 (en) | 2018-06-28 | 2022-04-12 | Adobe Inc. | Determining image handle locations |
US10621764B2 (en) | 2018-07-05 | 2020-04-14 | Adobe Inc. | Colorizing vector graphic objects |
Also Published As
Publication number | Publication date |
---|---|
CN106156700B (en) | 2019-06-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105976400B (en) | Method for tracking target and device based on neural network model | |
CN104268591B (en) | A kind of facial critical point detection method and device | |
CN103218605B (en) | A kind of fast human-eye positioning method based on integral projection and rim detection | |
Wu et al. | Detection and counting of banana bunches by integrating deep learning and classic image-processing algorithms | |
CN107909107A (en) | Fiber check and measure method, apparatus and electronic equipment | |
CN107403168A (en) | A kind of facial-recognition security systems | |
CN108154105A (en) | Aquatic organism detects and recognition methods, device, server and terminal device | |
CN108224895B (en) | Article information input method and device based on deep learning, refrigerator and medium | |
CN106250874A (en) | A kind of dress ornament and the recognition methods of carry-on articles and device | |
CN108416266A (en) | A kind of video behavior method for quickly identifying extracting moving target using light stream | |
CN107944435A (en) | A kind of three-dimensional face identification method, device and processing terminal | |
CN105427346B (en) | A kind of motion target tracking method and system | |
CN113269737B (en) | Fundus retina artery and vein vessel diameter calculation method and system | |
CN110047081A (en) | Example dividing method, device, equipment and the medium of chest x-ray image | |
CN104182984B (en) | Method and system for rapidly and automatically collecting blood vessel edge forms in dynamic ultrasonic image | |
CN109858476A (en) | The extending method and electronic equipment of label | |
CN109146868A (en) | 3D Lung neoplasm generation method, device and electronic equipment | |
CN109117746A (en) | Hand detection method and machine readable storage medium | |
CN110271006A (en) | Mechanical arm visual guide method and device | |
CN109583331A (en) | Human wrist arteries and veins mouth position precise positioning method based on deep learning | |
CN105741326B (en) | A kind of method for tracking target of the video sequence based on Cluster-Fusion | |
CN107169996A (en) | Dynamic human face recognition methods in a kind of video | |
EP3671635B1 (en) | Curvilinear object segmentation with noise priors | |
CN106156700A (en) | The extracting method of characteristics of human body and device in a kind of image procossing | |
CN104463896B (en) | Image corner point detection method and system based on kernel similar region distribution characteristics |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |