CN104298358B - A kind of dynamic 3D gesture identification methods based on joint space position data - Google Patents
A kind of dynamic 3D gesture identification methods based on joint space position data Download PDFInfo
- Publication number
- CN104298358B CN104298358B CN201410589578.5A CN201410589578A CN104298358B CN 104298358 B CN104298358 B CN 104298358B CN 201410589578 A CN201410589578 A CN 201410589578A CN 104298358 B CN104298358 B CN 104298358B
- Authority
- CN
- China
- Prior art keywords
- gesture
- point
- position data
- human body
- joint
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
- G06V40/28—Recognition of hand or arm movements, e.g. recognition of deaf sign language
Landscapes
- Engineering & Computer Science (AREA)
- Human Computer Interaction (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- User Interface Of Digital Computer (AREA)
- Image Analysis (AREA)
Abstract
The present invention discloses a kind of dynamic 3D gesture identification methods based on joint space position data, it is characterised in that comprises the following steps：Gathered by the body-sensing camera being connected with computer and identify skeleton positional information, by the spatial position data and body-sensing camera itself attitude information of each joint reference point of human body, positiveization and normalized are carried out to skeleton positional information；According to different gesture motions, different joint reference points is selected；For the motor pattern feature of each gesture, judged according to the time requirement between criterion order and criterion order；When the spatial position data of one section newest in caching each joint reference point of continuous reliable human body meets the time requirement between the criterion order and criterion order of the gesture, it is determined as that human body makes the gesture.The present invention enters the identification of Mobile state 3D gestures based on body-sensing camera, and method simple practical is accurate.
Description
Technical field
The present invention relates to 3D gesture identification fields, specifically a kind of dynamic 3D gestures based on joint space position data
Recognition methods.
Background technology
With the continuous popularization of touch screen technology, user is already adapted to and has become familiar with the interaction with machine.Now, people
Machine interaction technique steps and has gone up higher step, enters the gesture identification epoch, but this is not successful yet.Gesture
Identification occurs in amusement and Game Market, but what kind of influence is this technology will produce to our daily life
It imaginary might as well once take on sofa, only need with a flouriss with regard to that can manipulate near light and TV, or automobile automatic detection
Whether pedestrian is had.As Gesture Recognition supports the development that deepens continuously of human-computer interaction, these and other function will obtain quickly
To realize.Gesture Recognition is studied using 2D visions always for a long time, but with the appearance of 3D sensor technologies,
3D Gesture Recognitions will more be developed.
The content of the invention
It is an object of the invention to provide a kind of dynamic 3D gesture identification methods based on joint space position data.
Technical program of the present invention lies in：A kind of dynamic 3D gesture identification methods based on joint space position data, its
It is characterised by, comprises the following steps：
S01：Gathered by the body-sensing camera being connected with computer and identify skeleton positional information, will gathered
And the skeleton positional information identified is stored in caching in the lump together with the time is obtained, the skeleton positional information includes people
The spatial position data of each joint reference point of body；
S02：The skeleton positional information of deposit is located in advance based on existing skeleton positional information in caching
Reason；
S03：By the spatial position data and body-sensing camera itself attitude information of each joint reference point of human body, to people
Body bone site information carries out positiveization and normalized；
S04：According to different gesture motions, different joint reference points is selected；
S05：The spatial position data of each joint reference point of the continuous reliable human body of newest in caching one section is chosen, for
Different target gestures, sets following criterion：Gesture motion region, gesture motion time, impact point holdup time, target point fortune
Dynamic distance, impact point direction of motio；For the motor pattern feature of each gesture, further according between criterion order and criterion order
Time requirement is judged；
S06：When the spatial position data of one section newest in caching each joint reference point of continuous reliable human body meets this
Gesture criterion order and criterion order between time requirement when, be determined as that human body makes the gesture.
In an embodiment of the present invention, step S07 is also included after the step S06：Pass through what is stored in this section of caching
Skeleton positional information, the spatial position data for choosing each joint reference point of human body calculate the action generation point position of each frame
Put；The motion path of target joint point is obtained for different gesture motion, the characteristics of based on different gestures, by motion path and
The purpose direction and the confidence band in purpose direction that action generation point is acted.
In an embodiment of the present invention, in the step S07, as the right hand waves to act, choose the right hand and appear in left and right area
Ultra-Left, ultra-Right position and period right elbow space average position, calculate 3 formed planes vector as action purpose
Direction, the space average position of shoulder midpoint is chosen as action generation point, the confidence band of purpose is action generation point and the right side
The fan-shaped range that the ultra-Left ultra-Right position of hand is sketched the contours of.
In an embodiment of the present invention, the body-sensing camera is Kinect or Asus Asus Xtion.
In an embodiment of the present invention, described pre-process is specially：By closing on a few frame skeleton positional informations in sky
Between the variance that is distributed judge whether the skeleton positional information of deposit reliable, and to each joint reference point of human body of authentic data
Spatial position data carries out moving average processing.
In an embodiment of the present invention, each joint reference point of the human body includes：Shoulder midpoint, left point of shoulder point, right point of shoulder
Point, belly midpoint, left finesse, left elbow, right finesse, right elbow, head.
In an embodiment of the present invention, in the step S04, wave to act such as the right hand and choose right finesse, right elbow, right point of shoulder
Point, shoulder midpoint, head, belly midpoint are as a reference point.
The present invention enters the identification of Mobile state 3D gestures based on body-sensing camera, and method simple practical is accurate.
Brief description of the drawings
Fig. 1 is the method flow diagram of one embodiment of the invention.
Fig. 2 is the method flow diagram of another embodiment of the present invention.
Embodiment
For features described above of the invention and advantage can be become apparent, special embodiment below, and coordinate accompanying drawing, make detailed
Carefully it is described as follows, but the present invention is not limited thereto.
As shown in figure 1, the present invention provides a kind of dynamic 3D gesture identification methods based on joint space position data, including
Following steps：
S01：Gathered by the body-sensing camera being connected with computer and identify skeleton positional information, will gathered
And the skeleton positional information identified is stored in Computer Cache, the skeleton positional information in the lump together with the time is obtained
Include the spatial position data of each joint reference point of human body；
S02：The skeleton positional information of deposit is entered based on existing skeleton positional information in Computer Cache
Row pretreatment；
S03：Pass through the spatial position data and body-sensing camera itself attitude information of each joint reference point of human body（Body-sensing
Itself attitude information such as " upright ", " level ", " transverse direction " measured by camera gravity sensor）, to skeleton position
Confidence breath carries out positiveization and normalized；
S04：According to different gesture motions, different joint reference points is selected；
S05：The spatial position data of each joint reference point of the continuous reliable human body of newest in caching one section is chosen, for
Different target gestures, set following criterion and admissible error：When gesture motion region, gesture motion time, target point are detained
Between, impact point move distance, impact point direction of motio etc.；For the motor pattern feature of each gesture, further according to criterion order
And the time requirement between criterion order is judged；
S06：When the spatial position data of one section newest in caching each joint reference point of continuous reliable human body meets this
Gesture criterion order and criterion order between time requirement when, be determined as that human body makes the gesture.
As shown in Fig. 2 in another embodiment of the present invention, also include step S07 after the step S06：Delayed by this section
The skeleton positional information of middle storage is deposited, the spatial position data for choosing each joint reference point of human body calculates the action of each frame
Generation point position；The motion path of target joint point is obtained for different gesture motion, the characteristics of based on different gestures, by transporting
The purpose direction and the confidence band in purpose direction that dynamic path and action generation point are acted.
In the step S07, as the right hand waves to act, choose the right hand appear in left and right area ultra-Left, ultra-Right position and
The space average position of period right elbow, the vector of 3 formed planes is calculated as action purpose direction, chooses shoulder midpoint
Sketched the contours as action generation point, the confidence band of purpose by action generation point and the ultra-Left ultra-Right position of the right hand space average position
The fan-shaped range gone out.
The body-sensing camera model Kinect body-sensing cameras or Asus's Asus Xtion body-sensing cameras.
The pretreatment is specially：Judge deposit in the variance of spatial distribution by closing on a few frame skeleton positional informations
Skeleton positional information it is whether reliable, and line slip is entered to each joint reference point spatial position data of human body of authentic data
Average treatment.
Each joint reference point of human body includes：Shoulder midpoint, left point of shoulder point, right point of shoulder point, belly midpoint, left finesse,
Left elbow, right finesse, right elbow, head etc..
In the step S04, wave to act such as the right hand and choose right finesse, right elbow, right point of shoulder point, shoulder midpoint, head, abdomen
Portion midpoint is as a reference point.
In the step S05 and step S06, as the right hand waves to act, the judgement of gesture motion region is right finesse on the right side
1/3 arm length degree above elbow, so that straight up for symmetry axis, left area is 8 centimeters to the left of right elbow abscissa, right area is that right elbow is horizontal
8 centimeters to the right of coordinate, when right hand abscissa appears in left area-right area-left area or right area-left area-right area, and switch every time when
Between be spaced within 1 second, then be determined as that the right hand makes gesture of waving；Ergonomics feature is considered simultaneously, except straight up
Outside symmetry axis, several symmetry axis of increase deflection to the right to adapt to user action, predicting relation under any one symmetry axis into
It is vertical then represent the right hand make gesture of waving.
The foregoing is only presently preferred embodiments of the present invention, all equivalent changes done according to scope of the present invention patent with
Modification, it should all belong to the covering scope of the present invention.
Claims (6)
1. a kind of dynamic 3D gesture identification methods based on joint space position data, it is characterised in that comprise the following steps：
S01：Gathered by the body-sensing camera being connected with computer and identify skeleton positional information, will gathered and know
The skeleton positional information not gone out is stored in caching in the lump together with the time is obtained, and it is each that the skeleton positional information includes human body
The spatial position data of joint reference point；
S02：The skeleton positional information of deposit is pre-processed based on existing skeleton positional information in caching；
S03：By the spatial position data and body-sensing camera itself attitude information of each joint reference point of human body, to human body bone
Bone positional information carries out positiveization and normalized；
S04：According to different gesture motions, different joint reference points is selected；
S05：The spatial position data of each joint reference point of the continuous reliable human body of newest in caching one section is chosen, for difference
Target gesture, set following criterion：Gesture motion region, gesture motion time, impact point holdup time, target point motion away from
From, impact point direction of motio；For the motor pattern feature of each gesture, further according to the time between criterion order and criterion order
It is required that judged；
S06：When the spatial position data of one section newest in caching each joint reference point of continuous reliable human body meets the gesture
Criterion order and criterion order between time requirement when, be determined as that human body makes the gesture；
S07：Pass through the skeleton positional information stored in this section of caching, the locus number of selection each joint reference point of human body
According to the action generation point position for calculating each frame；The motion path of target joint point is obtained for different gesture motions, is based on
The characteristics of different gestures, the purpose direction acted by motion path and action generation point and the reliable model in purpose direction
Enclose.
2. a kind of dynamic 3D gesture identification methods based on joint space position data according to claim 1, its feature
It is：
In the step S07, when the right hand waves action, ultra-Left, ultra-Right position and phase that the right hand appears in left and right area are chosen
Between right elbow space average position, calculate the vector of 3 formed planes as action purpose direction, choose the sky of shoulder midpoint
Between mean place sketched the contours of as action generation point, the confidence band of purpose by action generation point and the ultra-Left ultra-Right position of the right hand
Fan-shaped range.
3. a kind of dynamic 3D gesture identification methods based on joint space position data according to claim 1, its feature
It is：The body-sensing camera is Kinect or Asus AsusXtion.
4. a kind of dynamic 3D gesture identification methods based on joint space position data according to claim 1, its feature
It is, the pretreatment is specially：Judge deposit in the variance of spatial distribution by closing on a few frame skeleton positional informations
Whether skeleton positional information is reliable, and enters line slip to each joint reference point spatial position data of human body of authentic data and put down
Handle.
5. a kind of dynamic 3D gesture identification methods based on joint space position data according to claim 1, its feature
It is：Each joint reference point of human body includes：Shoulder midpoint, left point of shoulder point, right point of shoulder point, belly midpoint, left finesse, a left side
Elbow, right finesse, right elbow, head.
6. a kind of dynamic 3D gesture identification methods based on joint space position data according to claim 1, its feature
It is：In the step S04, when the right hand waves action, right finesse, right elbow, right point of shoulder point, shoulder midpoint, head, abdomen are chosen
Portion midpoint is as a reference point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410589578.5A CN104298358B (en) | 2014-10-29 | 2014-10-29 | A kind of dynamic 3D gesture identification methods based on joint space position data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410589578.5A CN104298358B (en) | 2014-10-29 | 2014-10-29 | A kind of dynamic 3D gesture identification methods based on joint space position data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104298358A CN104298358A (en) | 2015-01-21 |
| CN104298358B true CN104298358B (en) | 2017-11-21 |
Family
ID=52318120
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410589578.5A Active CN104298358B (en) | 2014-10-29 | 2014-10-29 | A kind of dynamic 3D gesture identification methods based on joint space position data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104298358B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105302310B (en) * | 2015-11-12 | 2018-08-31 | 姚焕根 | A kind of gesture identifying device, system and method |
| CN106914016B (en) * | 2015-12-25 | 2020-12-04 | 北京奇虎科技有限公司 | Game player determination method and device |
| CN108108709B (en) * | 2017-12-29 | 2020-10-16 | 纳恩博（北京）科技有限公司 | Identification method and device and computer storage medium |
| CN108399367B (en) * | 2018-01-31 | 2020-06-23 | 深圳市阿西莫夫科技有限公司 | Hand motion recognition method and device, computer equipment and readable storage medium |
| CN109213189A (en) * | 2018-07-19 | 2019-01-15 | 安徽共生物流科技有限公司 | A kind of unmanned plane inspection system and movement determination method |
| CN112257639A (en) * | 2020-10-30 | 2021-01-22 | 福州大学 | Student learning behavior identification method based on human skeleton |
| CN115775347A (en) * | 2021-11-04 | 2023-03-10 | 中国科学院深圳先进技术研究院 | Taijiquan identification method based on fusion information, terminal device and storage medium |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102509092A (en) * | 2011-12-12 | 2012-06-20 | 北京华达诺科技有限公司 | Spatial gesture identification method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101483713B1 (en) * | 2008-06-30 | 2015-01-16 | 삼성전자 주식회사 | Apparatus and Method for capturing a motion of human |
-
2014
- 2014-10-29 CN CN201410589578.5A patent/CN104298358B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102509092A (en) * | 2011-12-12 | 2012-06-20 | 北京华达诺科技有限公司 | Spatial gesture identification method |
Non-Patent Citations (1)
| Title |
|---|
| 《面向多方式人际交互的肢体动作识别研究》;曹雏清;《中国博士学位论文全文数据库信息科技辑》;20140315(第3期);第29-34页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104298358A (en) | 2015-01-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104298358B (en) | A kind of dynamic 3D gesture identification methods based on joint space position data | |
| EP2674913B1 (en) | Three-dimensional object modelling fitting & tracking. | |
| US9354709B1 (en) | Tilt gesture detection | |
| CN107688342B (en) | The obstruction-avoiding control system and method for robot | |
| CN108475439B (en) | Three-dimensional model generation system, three-dimensional model generation method, and recording medium | |
| CA2804902C (en) | A method circuit and system for human to machine interfacing by hand gestures | |
| KR101738569B1 (en) | Method and system for gesture recognition | |
| US9135502B2 (en) | Method for the real-time-capable, computer-assisted analysis of an image sequence containing a variable pose | |
| CN105362048B (en) | Obstacle information reminding method, device and mobile device based on mobile device | |
| CN103999126B (en) | Method and apparatus for estimating pose | |
| CA2784554C (en) | Head recognition method | |
| CN111095164B (en) | Method and apparatus for detecting user input in accordance with gestures | |
| CN104515992B (en) | A kind of method and device that spacescan positioning is carried out using ultrasonic wave | |
| CN108931983A (en) | Map constructing method and its robot | |
| US20120163723A1 (en) | Classification of posture states | |
| CN106846403A (en) | The method of hand positioning, device and smart machine in a kind of three dimensions | |
| CN105849673A (en) | Human-to-computer natural three-dimensional hand gesture based navigation method | |
| US20100328319A1 (en) | Information processor and information processing method for performing process adapted to user motion | |
| CN109760627A (en) | Exempt to stretch out one's hand formula access method | |
| CN101996311A (en) | Yoga stance recognition method and system | |
| JP6320016B2 (en) | Object detection apparatus, object detection method and program | |
| CN116416518A (en) | Intelligent obstacle avoidance method and device | |
| KR102346294B1 (en) | Method, system and non-transitory computer-readable recording medium for estimating user's gesture from 2d images | |
| CN108108709A (en) | A kind of recognition methods and device, computer storage media | |
| KR101394279B1 (en) | Method for recognition of user's motion by analysis of depth image and apparatus for analyzing user's motion using the same |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information |
Address after: 361000 Fujian province Xiamen Software Park Road No. 31 205 unit two expected Applicant after: The conductor (Xiamen) Technology Co. Ltd. Address before: 361009 Fujian province Xiamen software park two sunrise Road No. 22 Room 102, unit I19 Applicant before: CONDUCTOR (XIAMEN) INTELLIGENT TECHNOLOGY CO., LTD. |
|
| COR | Change of bibliographic data | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |