CN104821010A - Binocular-vision-based real-time extraction method and system for three-dimensional hand information - Google Patents
Binocular-vision-based real-time extraction method and system for three-dimensional hand information Download PDFInfo
- Publication number
- CN104821010A CN104821010A CN201510222345.6A CN201510222345A CN104821010A CN 104821010 A CN104821010 A CN 104821010A CN 201510222345 A CN201510222345 A CN 201510222345A CN 104821010 A CN104821010 A CN 104821010A
- Authority
- CN
- China
- Prior art keywords
- staff
- real
- binocular
- time
- camera
- 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
Landscapes
- Image Analysis (AREA)
Abstract
The invention relates to a binocular-vision-based real-time extraction method and system for three-dimensional hand information. The method comprises: obtaining image information collected by a left camera and a right camera of a binocular camera system in real time; detecting hands in the images collected by the left camera and the right camera in real time; detecting a hand at a first frame by using a histogram of oriented gradient (HOG) and a support vector machine (SVM) linear classifier and detecting hands at follow-up frames by using a target tracking algorithm; extracting information of centers of palms and fingertips of the detected hands in real time and using the information as feature points of the hands; and according to the binocular-vision principle, calculating three-dimension coordinates of the feature points based on the extracted feature points so as to obtain real-time three-dimensional hand information. In addition, the system is composed of a binocular camera system and an information processing device; and the information processing device includes a hand detection module, a feature point extraction module, and a three-dimensional reconstruction module. According to the invention, real-time reliable detection and tracking of hands under a complex background can be realized; and thus reliable reconstruction of the three-dimensional hand information can be realized based on the detection and tracking.
Description
Technical field
The present invention relates to real time extracting method and the system of staff three-dimensional information, particularly relate to a kind of staff three-dimensional information real time extracting method based on binocular vision and system, relate to the technical field such as Digital Image Processing and pattern-recognition, space three-dimensional acquisition of information and man-machine interaction.
Background technology
In recent years, along with developing rapidly of computer vision technique, the gesture interaction of view-based access control model, because of the convenience that it is exclusive, obtains increasing investigation and application in field of human-computer interaction.Simultaneously, developing rapidly and people's improving constantly the requirement of the authenticity in man-machine interaction and the property immersed of 3D technology, 3D interaction technique obtains research and the favor of increasing people, correlation technique also achieves the development of advancing by leaps and bounds, 3D becomes the Main Trends of The Development of a lot of interaction technique alternately gradually, also some novel 3D body sense interactive devices have been there is on the market, as Kinect, Leap Motion, Google Glass etc., illustrate that the mutual of view-based access control model, particularly Three-Dimensional Interaction Technology have wide market outlook.
The detection of hand, to follow the tracks of and the extraction of three-dimensional information is the basis realizing three-dimension interaction.At present, the detection method of hand has multiple, is generally speaking mainly divided into two classes: the detection based on the colour of skin and the detection based on machine learning.Based on the colour of skin detection to illumination and shade more responsive, simultaneously other positions such as arm, face etc. and the close object of other and skin color of human body all the detection of opponent can produce and disturb, and therefore the simple detection carrying out staff based on the colour of skin is not the very reliable method of one.And be combined with ADABOOST scheduling algorithm the detection carrying out training the sorter obtaining staff to be used for staff based on other features such as the features such as HAAR, LBP, HOG except the colour of skin that the method for machine learning mainly depends on staff.Because the detailed information comprised in hand is not very abundant, the detection that HAAR characteristic sum LBP feature is used for staff is not very effective, and false drop rate is higher.Therefore, the staff detection and tracking under first we need to find effective method to realize complex background, we could carry out the extraction of staff three-dimensional information further like this.
Summary of the invention
The object of this invention is to provide a kind of staff three-dimensional information real time extracting method based on binocular vision and system, to realize the detection and tracking to Complex Background hand, and realize the reliable reconstruction to staff three-dimensional information on this basis.
Concrete technical scheme of the present invention is:
Based on a staff three-dimensional information real time extracting method for binocular vision, this extracting method comprises:
The image information of the left and right camera acquisition of Real-time Obtaining binocular camera system;
The left and right camera acquisition of real-time detection to image in staff, concrete, employing HOG (Histogram of OrientedGradient) characteristic sum SVM (Support Vector Machine) linear classifier detects the staff in first frame, adopts the staff in target tracking algorism detection postorder frame;
The centre of the palm of the staff that extract real-time detects and finger tip, as the unique point of staff; And
According to Binocular Vision Principle, calculate the three-dimensional coordinate of described unique point by the described unique point extracted, thus obtain real-time staff three-dimensional information.
Preferably, the described left and right camera acquisition of real-time detection to image in staff step in, the method for the staff adopting HOG characteristic sum SVM linear classifier to detect in first frame comprises:
Target image is carried out to the HOG feature extraction of different scale; And
Detect by the HOG feature of the SVM linear classifier trained to the target image extracted, obtain staff region;
And, described in the SVM linear classifier that trains obtain by the following method:
Extract the HOG feature of staff sample and background sample respectively;
The HOG feature of the staff sample extracted and background sample is input in the linear training aids of SVM and carries out training and obtain initial SVM linear classifier; And
Train together with original sample as difficult example with the flase drop sample that some collected by described initial SVM linear classifier, the SVM linear classifier trained described in obtaining.
Preferably, the described left and right camera acquisition of real-time detection to image in staff step in, described target tracking algorism is CamShift algorithm.
Preferably, the centre of the palm of the staff that described extract real-time detects and finger tip, the step as the unique point of staff comprises:
Use oval complexion model to carry out Face Detection to detection window, obtain staff region;
Medium filtering is done to the staff region after Face Detection, obtains pure staff region;
Range conversion is carried out to the staff region obtained after medium filtering;
The staff region of adjusting the distance after converting is carried out thresholding operation and is obtained palm area;
Calculate the barycenter of described palm area, using this barycenter as the centre of the palm; And
Calculate the distance of the centre of the palm to staff profile, using the finger tip of its local maximum point as staff.
Preferably, described according to Binocular Vision Principle, calculate the three-dimensional coordinate of described unique point by the described unique point extracted, thus obtain in the step of real-time staff three-dimensional information, the systematic parameter of the binocular camera system of employing obtains by the following method:
Left and right video camera synchronous acquisition 10 to the 20 width different distance of binocular camera system, the cross-hatch pattern picture of different azimuth;
Extract left and right camera acquisition to cross-hatch pattern picture in the angle point of gridiron pattern pattern, respectively left and right video camera is carried out to the demarcation of single camera, obtains inside and outside parameter and the distortion correction coefficient of left and right video camera; And
Carry out demarcating the translation matrix between the left and right video camera of acquisition and rotation matrix to binocular camera system.
Based on a staff three-dimensional information extract real-time system for binocular vision, this extraction system comprises:
Binocular camera system; And
Signal conditioning package, this signal conditioning package comprises:
Staff detection module, for detect in real time left and right camera acquisition to image in staff, concrete, adopt HOG characteristic sum SVM linear classifier to detect staff in first frame, adopt target tracking algorism to detect staff in postorder frame;
Feature point extraction module, for the unique point of the staff that extract real-time detects; And
Three-dimensional reconstruction module, for according to Binocular Vision Principle, calculates the three-dimensional coordinate of the centre of the palm and finger tip by the unique point extracted.
Preferably, this extraction system also comprises system calibrating device, and this system calibrating device is used for binocular camera system calibrating, obtains the systematic parameter of binocular camera system.
Preferably, this extraction system also comprises signal transmitting apparatus, this signal transmitting apparatus be used for by binocular camera system acquisition to image information be sent to described system calibrating device and described signal conditioning package.
The inventive method and system can realize real-time, the reliable detection and tracking of Complex Background hand, and realize the reliable reconstruction to staff three-dimensional information on this basis.
Accompanying drawing explanation
Fig. 1 is the process flow diagrams of some embodiments based on the staff three-dimensional information real time extracting method of binocular vision;
Fig. 2 is the process flow diagram that in some embodiments, staff detects;
Fig. 3 is the process flow diagram of staff feature point extraction in some embodiments;
Fig. 4 is the block diagrams of some embodiments based on the staff three-dimensional information extract real-time system of binocular vision.
Embodiment
Below in conjunction with schematic case study on implementation, embodiments of the present invention are described.For the sake of clarity, in this manual the actual all features implemented are not described.It should be understood that, in the development of any this actual embodiment, must make the specific objective that many embodiments specifically determine to realize developer, such as conform to the constraint relating to business with design system, the difference according to implementation process changes by described constraint.In addition, it should be understood that this development may be complicated and time-consuming, but will be engaged in routine work for the those of ordinary skill in the art benefiting from present disclosure.
Although the present invention allows various amendment and alternative form, show particular of the present invention by the example in accompanying drawing and be described in detail at this.But, it should be understood that, in order to the understanding of the present invention is more thorough comprehensively to the explanation of particular herein, and be not intended to limit the invention to particular forms disclosed, but contrary, covering drops on as by all modifications in appended spirit and scope defined in claim of the present invention, equivalents thereto and replacement scheme by the present invention.
As shown in Figure 1, some embodiments comprise the following steps based on the staff three-dimensional information real time extracting method of binocular vision:
The image information of the left and right camera acquisition of step S1, Real-time Obtaining binocular camera system.
Step S2, in real time detect left and right camera acquisition to image in staff, concrete, adopt HOG characteristic sum SVM linear classifier to detect staff in first frame, adopt target tracking algorism to detect staff in postorder frame.
In order to the staff that can realize under complex background detects, adopt the method for HOG feature+SVM linear classifier (namely employing HOG characteristic sum SVM linear classifier detects the staff in first frame) here.With reference to Fig. 2, the concrete testing process adopting the method for HOG feature+SVM linear classifier to detect staff comprises: step S21, target image is carried out to the HOG feature extraction of different scale; And step S22, to detect by the HOG feature of the SVM linear classifier trained to the target image extracted, obtain staff region.For SVM linear classifier, first need to collect enough staff samples and background sample making training set, extract the HOG feature of staff sample and background sample respectively, the HOG feature extracted is input in the linear training aids of SVM and carries out training to obtain a staff linear classifier based on HOG feature.
Have very high recall rate by staff sample and background sample training SVM linear classifier out merely although use, its false drop rate is also quite high simultaneously.In order to obtain the detection for staff of recall rate is high, false drop rate is low SVM linear classifier, in some preferred embodiment, further use staff sample and background sample training SVM linear classifier are out optimized, concrete, with reference to Fig. 2, described in the SVM linear classifier that trains obtain by the following method: the HOG feature extracting staff sample and background sample respectively; The HOG feature of the staff sample extracted and background sample is input in the linear training aids of SVM and carries out training and obtain initial SVM linear classifier; And train together with original sample as difficult example with the flase drop sample that some collected by described initial SVM linear classifier, the SVM linear classifier trained described in obtaining.
The method of carrying out staff detection due to HOG feature+SVM linear classifier is a kind of detection method comparatively consuming time, under present circumstances, this method of simple use is difficult to realize the real-time detection of opponent, in order to the extract real-time to staff three-dimensional information finally can be realized, HOG characteristic sum SVM linear classifier is adopted to detect staff in first frame in the present invention, and for postorder frame, then adopt target tracking algorism to detect staff.In the preferred embodiment, CamShift algorithm is adopted to follow the tracks of staff.CamShift algorithm is a kind of target tracking algorism based on color space.The thought of algorithm does MeanShift computing to all frames of video image, and using the initial value of the result of previous frame as the search window of next frame MeanShift algorithm, carry out interative computation, thus realize the tracking to staff.
The centre of the palm of the staff that step S3, extract real-time detect and finger tip, as the unique point of staff.
Realizing under the prerequisite to Complex Background hand detection and tracking, we need the unique point extracting staff to carry out coupling and the reconstruction of 3 D stereo.The centre of the palm is the unique point commonly used the most, but usually to the extraction in the centre of the palm or adopt staff circumscribed circle as the barycenter in the centre of the palm or direct calculating staff region as the centre of the palm, the centre of the palm that these two kinds of methods calculate gained is all more rough, have a strong impact on the precision of extraction to finger tip and coupling, and then affect the precision of three-dimensional reconstruction.
Therefore, we provide a kind of more reliable and stable centre of the palm extracting method, and on this basis, extract the finger tip point of staff.With reference to Fig. 3, the extracting method of this centre of the palm and finger tip comprises the following steps: step S31, use oval complexion model to carry out Face Detection to detection window, obtains staff region; Step S32, medium filtering is done to the staff region after Face Detection, obtain pure staff region; Step S33, range conversion is carried out to the staff region obtained after medium filtering; Staff region after step S34, conversion of adjusting the distance is carried out thresholding operation and is obtained palm area; Step S35, calculate the barycenter of described palm area, using this barycenter as the centre of the palm; Step S36, extraction staff profile; And step S37, calculate the centre of the palm to the distance of staff profile, using its local maximum point as finger tip.
Step S4, according to Binocular Vision Principle, calculate the three-dimensional coordinate of described unique point by the described unique point extracted, thus obtain real-time staff three-dimensional information.
When realizing three-dimensional reconstruction according to Binocular Vision Principle, need the systematic parameter using binocular camera system.By demarcating binocular camera system, the systematic parameter of binocular camera system can be obtained, comprising: the inside and outside parameter of left and right video camera, the distortion correction coefficient of left and right video camera, the translation matrix between left and right video camera and rotation matrix.
In the preferred embodiment, adopt gridiron pattern Camera Calibration Algorithm to demarcate binocular camera system, concrete grammar comprises: 1, first make a sizeable gridiron pattern scaling board; 2, the cross-hatch pattern picture of binocular camera system left and right video camera difference synchronous acquisition 10 to 20 width different distance, different azimuth; 3, extract left and right camera acquisition to cross-hatch pattern picture in the angle point of gridiron pattern pattern, respectively left and right video camera is carried out to the demarcation of single camera, obtains inside and outside parameter and the distortion correction coefficient of left and right video camera; 4, then carry out demarcating the translation matrix between the left and right video camera of acquisition and rotation matrix to binocular camera system.
Fig. 4 is the block diagrams of some embodiments based on the staff three-dimensional information extract real-time system of binocular vision.As shown in Figure 4, some embodiments comprise based on the staff three-dimensional information extract real-time system of binocular vision: binocular camera system 1 and signal conditioning package 2.
Binocular camera system 1 requirement can picture in real time, in synchronous acquisition Same Scene, requires that left and right two video cameras have similar parameter, and suitable relative position relation.
Signal conditioning package 2 comprises: staff detection module 21, feature point extraction module 22, three-dimensional reconstruction module 23.Staff detection module 21 for detect in real time left and right camera acquisition to image in staff, concrete, adopt HOG characteristic sum SVM linear classifier to detect staff in first frame, adopt target tracking algorism to detect staff in postorder frame.The unique point (comprising the centre of the palm and finger tip) of the staff that feature point extraction module 22 detects for extract real-time.Three-dimensional reconstruction module 23, for according to Binocular Vision Principle, calculates the three-dimensional coordinate of unique point by the unique point extracted.
Further, this extraction system also comprises system calibrating device 3, and this system calibrating device 3, for demarcating binocular camera system 1, obtains the systematic parameter of binocular camera system 1.
Further, this extraction system also comprises signal transmitting apparatus (not shown), this signal transmitting apparatus is used for the image information that binocular camera system 1 collects being sent to described system calibrating device 3 to realize demarcating, and the image information that binocular camera system 1 collects is sent to described signal conditioning package 2, detect and three-dimensional reconstruction to realize staff.
Claims (8)
1., based on the staff three-dimensional information real time extracting method of binocular vision, it is characterized in that, this extracting method comprises:
The image information of the left and right camera acquisition of Real-time Obtaining binocular camera system;
The left and right camera acquisition of real-time detection to image in staff, concrete, adopt HOG characteristic sum SVM linear classifier to detect staff in first frame, adopt target tracking algorism to detect staff in postorder frame;
The centre of the palm of the staff that extract real-time detects and finger tip, as the unique point of staff; And
According to Binocular Vision Principle, calculate the three-dimensional coordinate of described unique point by the described unique point extracted, thus obtain real-time staff three-dimensional information.
2. the staff three-dimensional information real time extracting method based on binocular vision according to claim 1, it is characterized in that, the described left and right camera acquisition of real-time detection to image in staff step in, the method for the staff adopting HOG characteristic sum SVM linear classifier to detect in first frame comprises:
Target image is carried out to the HOG feature extraction of different scale; And
Detect by the HOG feature of the SVM linear classifier trained to the target image extracted, obtain staff region;
The described SVM linear classifier trained obtains by the following method:
Extract the HOG feature of staff sample and background sample respectively;
The HOG feature of the staff sample extracted and background sample is input in the linear training aids of SVM and carries out training and obtain initial SVM linear classifier; And
Train together with original sample as difficult example with the flase drop sample that some collected by described initial SVM linear classifier, the SVM linear classifier trained described in obtaining.
3. the staff three-dimensional information real time extracting method based on binocular vision according to claim 1, it is characterized in that, the described left and right camera acquisition of real-time detection to image in staff step in, described target tracking algorism is CamShift algorithm.
4. the staff three-dimensional information real time extracting method based on binocular vision according to claim 1, it is characterized in that, the centre of the palm of the staff that described extract real-time detects and finger tip, the step as the unique point of staff comprises:
Use oval complexion model to carry out Face Detection to detection window, obtain staff region;
Medium filtering is done to the staff region after Face Detection, obtains pure staff region;
Range conversion is carried out to the staff region obtained after medium filtering;
The staff region of adjusting the distance after converting is carried out thresholding operation and is obtained palm area;
Calculate the barycenter of described palm area, using this barycenter as the centre of the palm;
Calculate the distance of the centre of the palm to staff profile, using its local maximum point as finger tip.
5. the staff three-dimensional information real time extracting method based on binocular vision according to claim 1, it is characterized in that, described according to Binocular Vision Principle, the three-dimensional coordinate of described unique point is calculated by the described unique point extracted, thus obtain in the step of real-time staff three-dimensional information, the systematic parameter of the binocular camera system of employing obtains by the following method:
Left and right video camera synchronous acquisition 10 to the 20 width different distance of binocular camera system, the cross-hatch pattern picture of different azimuth;
Extract left and right camera acquisition to cross-hatch pattern picture in the angle point of gridiron pattern pattern, respectively left and right video camera is carried out to the demarcation of single camera, obtains inside and outside parameter and the distortion correction coefficient of left and right video camera; And
Carry out demarcating the translation matrix between the left and right video camera of acquisition and rotation matrix to binocular camera system.
6., based on a staff three-dimensional information extract real-time system for binocular vision, it is characterized in that, this extraction system comprises:
Binocular camera system; And
Signal conditioning package, this signal conditioning package comprises:
Staff detection module, for detect in real time left and right camera acquisition to image in staff, concrete, adopt HOG characteristic sum SVM linear classifier to detect staff in first frame, adopt target tracking algorism to detect staff in postorder frame;
Feature point extraction module, the centre of the palm of the staff detected for extract real-time and finger tip; And
Three-dimensional reconstruction module, for according to Binocular Vision Principle, calculates the three-dimensional coordinate of unique point by the unique point extracted.
7. the staff three-dimensional information extract real-time system based on binocular vision according to claim 6, it is characterized in that, this extraction system also comprises system calibrating device, and this system calibrating device is used for binocular camera system calibrating, obtains the systematic parameter of binocular camera system.
8. the staff three-dimensional information extract real-time system based on binocular vision according to claim 7, it is characterized in that, this extraction system also comprises signal transmitting apparatus, this signal transmitting apparatus be used for by binocular camera system acquisition to image information be sent to described system calibrating device and described signal conditioning package.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510222345.6A CN104821010A (en) | 2015-05-04 | 2015-05-04 | Binocular-vision-based real-time extraction method and system for three-dimensional hand information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510222345.6A CN104821010A (en) | 2015-05-04 | 2015-05-04 | Binocular-vision-based real-time extraction method and system for three-dimensional hand information |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104821010A true CN104821010A (en) | 2015-08-05 |
Family
ID=53731293
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510222345.6A Pending CN104821010A (en) | 2015-05-04 | 2015-05-04 | Binocular-vision-based real-time extraction method and system for three-dimensional hand information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104821010A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106408579A (en) * | 2016-10-25 | 2017-02-15 | 华南理工大学 | Video based clenched finger tip tracking method |
| CN106778452A (en) * | 2015-11-24 | 2017-05-31 | 沈阳新松机器人自动化股份有限公司 | Service robot is based on human testing and the tracking of binocular vision |
| CN106920251A (en) * | 2016-06-23 | 2017-07-04 | 阿里巴巴集团控股有限公司 | Staff detecting and tracking method and device |
| CN107205145A (en) * | 2016-03-17 | 2017-09-26 | 中航华东光电（上海）有限公司 | Terminal guidance video image three dimensional data collection system |
| CN107813310A (en) * | 2017-11-22 | 2018-03-20 | 浙江优迈德智能装备有限公司 | One kind is based on the more gesture robot control methods of binocular vision |
| CN108122280A (en) * | 2017-12-20 | 2018-06-05 | 北京搜狐新媒体信息技术有限公司 | The method for reconstructing and device of a kind of three-dimensional point cloud |
| CN108256504A (en) * | 2018-02-11 | 2018-07-06 | 苏州笛卡测试技术有限公司 | A kind of Three-Dimensional Dynamic gesture identification method based on deep learning |
| CN108305321A (en) * | 2018-02-11 | 2018-07-20 | 谢符宝 | A kind of three-dimensional human hand 3D skeleton patterns real-time reconstruction method and apparatus based on binocular color imaging system |
| CN109272519A (en) * | 2018-09-03 | 2019-01-25 | 先临三维科技股份有限公司 | Determination method, apparatus, storage medium and the processor of nail outline |
| CN109544692A (en) * | 2017-09-22 | 2019-03-29 | 江苏智谋科技有限公司 | Three Dimensions Inspection System based on VR equipment |
| CN109670396A (en) * | 2018-11-06 | 2019-04-23 | 华南理工大学 | A kind of interior Falls Among Old People detection method |
| CN117037272A (en) * | 2023-08-08 | 2023-11-10 | 深圳市震有智联科技有限公司 | Method and system for monitoring fall of old people |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103268482A (en) * | 2013-05-31 | 2013-08-28 | 清华大学 | Low-complexity gesture extracting and gesture depth acquiring method |
-
2015
- 2015-05-04 CN CN201510222345.6A patent/CN104821010A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103268482A (en) * | 2013-05-31 | 2013-08-28 | 清华大学 | Low-complexity gesture extracting and gesture depth acquiring method |
Non-Patent Citations (5)
| Title |
|---|
| 任 彧等: "基于HOG 特征和 SVM 的手势识别", 《科技通报》 * |
| 廖赟: "基于裸手的自然人机交互关键算法研究", 《中国博士学位论文全文数据库信息科技辑》 * |
| 谭同德等: "基于双目视觉的人手定位与手势识别系统研究", 《计算机工程与设计》 * |
| 邬铎: "基于双目视觉的手势识别算法研究与实现", 《中国硕士学位论文全文数据库信息科技辑》 * |
| 闫贺: "基于共生梯度方向直方图的实时人手检测系统", 《中国硕士学位论文全文数据库信息科技辑》 * |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106778452A (en) * | 2015-11-24 | 2017-05-31 | 沈阳新松机器人自动化股份有限公司 | Service robot is based on human testing and the tracking of binocular vision |
| CN107205145A (en) * | 2016-03-17 | 2017-09-26 | 中航华东光电（上海）有限公司 | Terminal guidance video image three dimensional data collection system |
| CN106920251A (en) * | 2016-06-23 | 2017-07-04 | 阿里巴巴集团控股有限公司 | Staff detecting and tracking method and device |
| WO2017219875A1 (en) * | 2016-06-23 | 2017-12-28 | 阿里巴巴集团控股有限公司 | Hand detecting and tracking method and device |
| US10885639B2 (en) | 2016-06-23 | 2021-01-05 | Advanced New Technologies Co., Ltd. | Hand detection and tracking method and device |
| US10885638B2 (en) | 2016-06-23 | 2021-01-05 | Advanced New Technologies Co., Ltd. | Hand detection and tracking method and device |
| CN106408579A (en) * | 2016-10-25 | 2017-02-15 | 华南理工大学 | Video based clenched finger tip tracking method |
| CN106408579B (en) * | 2016-10-25 | 2019-01-29 | 华南理工大学 | A kind of kneading finger tip tracking based on video |
| CN109544692A (en) * | 2017-09-22 | 2019-03-29 | 江苏智谋科技有限公司 | Three Dimensions Inspection System based on VR equipment |
| CN107813310A (en) * | 2017-11-22 | 2018-03-20 | 浙江优迈德智能装备有限公司 | One kind is based on the more gesture robot control methods of binocular vision |
| CN108122280A (en) * | 2017-12-20 | 2018-06-05 | 北京搜狐新媒体信息技术有限公司 | The method for reconstructing and device of a kind of three-dimensional point cloud |
| CN108305321A (en) * | 2018-02-11 | 2018-07-20 | 谢符宝 | A kind of three-dimensional human hand 3D skeleton patterns real-time reconstruction method and apparatus based on binocular color imaging system |
| CN108256504A (en) * | 2018-02-11 | 2018-07-06 | 苏州笛卡测试技术有限公司 | A kind of Three-Dimensional Dynamic gesture identification method based on deep learning |
| CN108305321B (en) * | 2018-02-11 | 2022-09-30 | 牧星天佑(北京)科技文化发展有限公司 | Three-dimensional human hand 3D skeleton model real-time reconstruction method and device based on binocular color imaging system |
| CN109272519A (en) * | 2018-09-03 | 2019-01-25 | 先临三维科技股份有限公司 | Determination method, apparatus, storage medium and the processor of nail outline |
| CN109272519B (en) * | 2018-09-03 | 2021-06-01 | 先临三维科技股份有限公司 | Nail contour determination method and device, storage medium and processor |
| CN109670396A (en) * | 2018-11-06 | 2019-04-23 | 华南理工大学 | A kind of interior Falls Among Old People detection method |
| CN117037272A (en) * | 2023-08-08 | 2023-11-10 | 深圳市震有智联科技有限公司 | Method and system for monitoring fall of old people |
| CN117037272B (en) * | 2023-08-08 | 2024-03-19 | 深圳市震有智联科技有限公司 | Method and system for monitoring fall of old people |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104821010A (en) | Binocular-vision-based real-time extraction method and system for three-dimensional hand information | |
| Zhou et al. | A novel finger and hand pose estimation technique for real-time hand gesture recognition | |
| CN103941866B (en) | Three-dimensional gesture recognizing method based on Kinect depth image | |
| CN101593022B (en) | Method for quick-speed human-computer interaction based on finger tip tracking | |
| CN105739702B (en) | Multi-pose finger tip tracking for natural human-computer interaction | |
| Sawant et al. | Real time sign language recognition using pca | |
| CN104317391A (en) | Stereoscopic vision-based three-dimensional palm posture recognition interactive method and system | |
| JP7015152B2 (en) | Processing equipment, methods and programs related to key point data | |
| CN110555408B (en) | Single-camera real-time three-dimensional human body posture detection method based on self-adaptive mapping relation | |
| CN103530599A (en) | Method and system for distinguishing real face and picture face | |
| CN110084192B (en) | Rapid dynamic gesture recognition system and method based on target detection | |
| CN103839277A (en) | Mobile augmented reality registration method of outdoor wide-range natural scene | |
| CN106204658A (en) | Moving image tracking and device | |
| KR20170036747A (en) | Method for tracking keypoints in a scene | |
| CN112861808B (en) | Dynamic gesture recognition method, device, computer equipment and readable storage medium | |
| CN108305321A (en) | A kind of three-dimensional human hand 3D skeleton patterns real-time reconstruction method and apparatus based on binocular color imaging system | |
| Zhou et al. | A study on attention-based LSTM for abnormal behavior recognition with variable pooling | |
| CN104898971A (en) | Mouse pointer control method and system based on gaze tracking technology | |
| CN104866826A (en) | Static gesture language identification method based on KNN algorithm and pixel ratio gradient features | |
| Yashas et al. | Hand gesture recognition: a survey | |
| Liang et al. | Egocentric hand pose estimation and distance recovery in a single RGB image | |
| Dai et al. | Federated filter algorithm with positioning technique based on 3D sensor | |
| CN108694348B (en) | Tracking registration method and device based on natural features | |
| CN112099330B (en) | Holographic human body reconstruction method based on external camera and wearable display control equipment | |
| Shih | Hand gesture recognition using color-depth association for smart home |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150805 |