CN107479693A - Real-time hand recognition methods based on RGB information, storage medium, electronic equipment - Google Patents
Real-time hand recognition methods based on RGB information, storage medium, electronic equipment Download PDFInfo
- Publication number
- CN107479693A CN107479693A CN201710549412.4A CN201710549412A CN107479693A CN 107479693 A CN107479693 A CN 107479693A CN 201710549412 A CN201710549412 A CN 201710549412A CN 107479693 A CN107479693 A CN 107479693A
- Authority
- CN
- China
- Prior art keywords
- hand
- image
- real
- model
- rgb
- 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
- 238000000034 method Methods 0.000 title claims abstract description 26
- 238000003709 image segmentation Methods 0.000 claims abstract description 9
- 238000004590 computer program Methods 0.000 claims description 6
- 238000004422 calculation algorithm Methods 0.000 claims description 5
- 238000004806 packaging method and process Methods 0.000 claims 1
- 230000033001 locomotion Effects 0.000 abstract description 20
- 238000005516 engineering process Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 238000010586 diagram Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 230000002452 interceptive effect Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000005057 finger movement Effects 0.000 description 1
- 210000002478 hand joint Anatomy 0.000 description 1
- 239000003550 marker Substances 0.000 description 1
- 230000000877 morphologic effect Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/017—Gesture based interaction, e.g. based on a set of recognized hand gestures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/06—Topological mapping of higher dimensional structures onto lower dimensional surfaces
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- 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
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/107—Static hand or arm
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Human Computer Interaction (AREA)
- Computer Graphics (AREA)
- Geometry (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
Real-time hand recognition methods of the invention based on RGB information, including obtain real-time RGB image group, two dimensional image segmentation, establish three-dimension gesture model, gesture judgement, output result.The invention further relates to storage medium and electronic equipment.The important joint of the hand of people is equivalent to ellipsoid model by the present invention using anisotropic Gaussian and model, by the image acquisition device hand realtime graphic of multi-angle, realizes the quick identification of hand gestures.The present invention is real-time, and degree of accuracy height, strong robustness, delay is low, can realize the complicated hand motion of identification.
Description
Technical field
The present invention relates to image recognition, more particularly to real-time hand recognition methods based on RGB information, storage medium, electricity
Sub- equipment.
Background technology
In man-machine interactive system, dynamic Gesture Recognition is one of interaction technique important in man-machine interaction.For
Enhancing accuracy and robustness, at this stage real-time dynamic hand gesture recognition be primarily present problems with:
1. the joint freedom degrees of hand are more, the difficulty of Real time identification and tracking is big;
2. the speed of actions of hand is fast, high is required to collecting device;
The overlapping meeting of finger in 3.RGB images impact to gesture identification, due to the skin of palm of hand color of same person
Identical, overlapping part is difficult to be accurately distinguished by the RGB information of 2D pictures again;
4. the hand dynamic rate difference of different people causes specific hand motion recognition inaccurate.
The content of the invention
For overcome the deficiencies in the prior art, it is an object of the invention to provide a kind of real-time hand based on RGB information
Recognition methods, the present invention is real-time, and degree of accuracy height, strong robustness, delay is low, can identify the hand motion of complexity.
The present invention provides the real-time hand recognition methods based on RGB information, comprises the following steps:
Real-time RGB image group is obtained, image acquisition device hand multi-angle image, obtains hand two dimension RGB image
Group;
Two dimensional image is split, and image segmentation is carried out successively to the image in the RGB image group;
Three-dimension gesture model is established, the RGB image after splitting according to image sets up vertical three-dimension gesture model;
Gesture judges, using anisotropic Gaussian and model to the three-dimension gesture model and hand three-dimensional modeling data storehouse
In model judgement is identified, obtain current hand gestures;
Gesture result is exported, exports the current hand gestures.
Further, the real-time hand recognition methods based on RGB information also switchs to two dimension including step three-dimensional hand gestures
Hand gestures, the three-dimension gesture model is projected to two-dimensional projection face and obtains two-dimentional gesture model.
Further, described step output gesture result is specially to export the two-dimentional gesture model and or or described
Three-dimension gesture model.
Further, the segmentation of described step two dimensional image is specifically using two-dimentional quaternary tree in the RGB image group
Image carries out image segmentation successively.
Further, described step establishes three-dimension gesture model specifically using action matching energy maximum algorithm to image
The image of the RGB image group after segmentation is matched, and establishes three-dimension gesture model.
Further, described three-dimension gesture model is carried out equivalent using the spheroid of 17 anisotropic Gaussian features.
Further, described image collector is set to video camera, and the frequency acquisition of described video camera is more than per second 20
Frame.
Further, the quantity of described image collecting device is at least 2.
A kind of electronic equipment, including:Processor;Memory;And program, wherein described program are stored in the storage
In device, and it is configured to by computing device, described program includes being used to perform the real-time hard recognition side based on RGB information
Method.
A kind of computer-readable recording medium, is stored thereon with computer program:The computer program is held by processor
Real-time hand recognition methods of the row based on RGB information.
Compared with prior art, the beneficial effects of the present invention are:
Real-time hand recognition methods of the invention based on RGB information, including obtain real-time RGB image group, two dimensional image point
Cut, establish three-dimension gesture model, gesture judges, output result.The invention further relates to storage medium and electronic equipment.The present invention adopts
The important joint of the hand of people is equivalent to ellipsoid model with anisotropic Gaussian and model, passes through the image collector of multi-angle
Collection hand realtime graphic is put, realizes the quick identification of hand gestures.The present invention is real-time, degree of accuracy height, strong robustness, prolongs
When it is low, the complicated hand motion of identification can be realized.
Described above is only the general introduction of technical solution of the present invention, in order to better understand the technological means of the present invention,
And can be practiced according to the content of specification, below with presently preferred embodiments of the present invention and coordinate accompanying drawing describe in detail as after.
The embodiment of the present invention is shown in detail by following examples and its accompanying drawing.
Brief description of the drawings
Accompanying drawing described herein is used for providing a further understanding of the present invention, forms the part of the application, this hair
Bright schematic description and description is used to explain the present invention, does not form inappropriate limitation of the present invention.In the accompanying drawings:
Fig. 1 is the real-time hand recognition methods step schematic diagram based on RGB information of the present invention;
Fig. 2 is the equivalent hand threedimensional model schematic diagram using SoG;
The equivalent hand threedimensional model using SAG that Fig. 3 is the present invention is schematically represented intention;
Fig. 4 is that the step three-dimensional hand gestures of the present invention switch to the principle schematic of two-dimentional hand gestures;
Fig. 5 is using number of cameras under SAG and SoG equivalent hand threedimensional model and error relationship figure;
Fig. 6 is using the error chart under SAG and SoG equivalent hand threedimensional model difference Sub Data Set.
Embodiment
Below, with reference to accompanying drawing and embodiment, the present invention is described further, it is necessary to which explanation is, not
Under the premise of afoul, new implementation can be formed between various embodiments described below or between each technical characteristic in any combination
Example.
Real-time hand recognition methods based on RGB information, as shown in figure 1, comprising the following steps:
Real-time RGB image group is obtained, image acquisition device hand multi-angle image, obtains hand two dimension RGB image
Group;In general, image collector are set to video camera, and the frequency acquisition of video camera is more than 20 frame per second, the number of image collecting device
Measure is at least 2;In one embodiment, lower operation time simultaneously to improve acquisition precision, hand is moved using 4 video cameras
Shooting is tracked, frequency acquisition is 25 frame per second.
Two dimensional image is split, and image segmentation is carried out successively to the image in RGB image group;Specifically use two-dimentional quaternary tree pair
Image in RGB image group carries out image segmentation successively.
Three-dimension gesture model is established, the RGB image after splitting according to image sets up vertical three-dimension gesture model;Specific use is moved
The image for making the RGB image group after matching energy maximum algorithm is split to image matches, by comparing three-dimensional modeling data
Each threedimensional model in storehouse is carried out pair with the two-dimensional signal of different angle camera collection in the two-dimensional projection of different angle
Than finally searching out the three-dimension gesture model most matched.It is the hand motion information definition that some is gathered under camera group
Ca, the 3D model definitions that some in 3D model libraries is acted are Cb, can obtain:
Wherein, EpqIt is the measuring similarity of above-mentioned two two-dimentional hand motions, d (cp, cq) represent two models
The similarity degree of middle feature, wherein,
In formula, E (Ca, Cb) action matching energy is defined as, represent hand motion in true hand motion and 3 d model library
Measuring similarity, E (Ca, Cb) bigger, then illustrate that real hand motion is got over some hand motion model in 3 d model library
It is similar.The RGB information for the hand motion being interpreted as in every pictures that all cameras are collected and
The similarity of two-dimensional projection in some threedimensional model do plus and, as a result maximum be in three-dimension gesture model with true hand
Act that most like model;d(cp, cq)DpqIt is then EpqDifferent hand-characteristics is resolved into.
Gesture judges, using anisotropic Gaussian and model in three-dimension gesture model and hand three-dimensional modeling data storehouse
Judgement is identified in model, obtains current hand gestures;First SoG (Sum of Gaussians gaussian sums model) is said
It is bright, SoG 2011 by C.Stoll in paper " Fast articulated motion tracking using a sums
Itd is proposed first in of Gaussians body model. " by SoG technologies be used for in human body tracking, a paper of 2013
“Interactive marker less articulated hand motion tracking using RGB and depth
Data. " then SoG has been first utilized in hand tracking, but SoG core concept is to use isotropism Gauss model, by people
Body trunk is equivalent to three-dimensional geometry body, so as to which trunk is described and matched.As shown in Fig. 2 use 30 isotropism height
This and model it is equivalent go out hand 3D models, each isotropism gaussian sum model can regard a spheroid as, and Fig. 2 is the hand
Projection of portion's model in two-dimensional space, it is found that reduction to hand motion is simultaneously bad, it is impossible to ideally represent the dynamic of hand
Make feature.
There is significant drawback to be using SoG technologies, it is equivalent into sphere model can not give full expression to hand motion spy
Sign, while the computing of the correlation of individual sphere model more than 30 also considerably increases the operand of system, improves to hardware
Demand, reduce the real-time of system algorithm.
The present invention proposes a kind of anisotropic Gaussian and model, is named as SAG:Sum of Anisotropic
Gaussians, preferably equivalent hand motion, and carry out equivalent hand using less geometric mould, make to hand action schedule
Up to more accurate, while also greatly reduce operand.
In three dimensions, the volume of hand is added approximation by the model of multiple solids and obtained, i.e.,:The space of hand is special
Sign is equivalent to the sum of anisotropic Gaussian characteristic model.So structure database hand three-dimensional space model with following formula
Represent:
Wherein, Gi(*) represents the anisotropic Gaussian model of a non-regularization;
Wherein, μiRepresent Gaussian mean, ∑iRepresent the covariance matrix of i-th of Gauss feature.In hsv color space,
Each Gauss model has the average color vector c of an associationi。
The final Gauss model for establishing three dimensional anisotropic, wherein, x ∈ R3, while can this 3D model is empty in 2D
Between mapped.
As shown in figure 3, by SAG it is equivalent go out hand model be made up of spheroid, by by each spheroid
Central point is connected, and obtains the equivalent skeletal structure of hand;It can see in Fig. 3, using 17 anisotropic Gaussian features with regard to energy
Hand threedimensional model is preferably expressed, and two-dimensional projection, closer to the physical feature of human hand, needing 30 in SoG, (amount of calculation is several
Add 1 times), but either three-dimensional or two dimensional model expression effect is all bad.
Three-dimensional hand gestures switch to two-dimentional hand gestures, and three-dimension gesture model is projected to two-dimensional projection face and obtains two-dimentional hand
Potential model;As shown in figure 4, by three-dimension gesture model (the 3D sphere models of equivalent hand joint are spheroid) at some visual angle
Tripleplane on (image projection surface under visual focus), obtain the two-dimentional gesture model of hand motion;Assuming that video camera
It is a camera matrix in origin (0,0,0), and P=K [I | 0], the parameter of its Gaussian Profile can obtain:
Wherein
In formula, | M | M determinant is represented, | M31| represent to remove the result of the 3rd row and the 1st row, k in Metzler matrix13Represent
That element that the 1st row the 3rd arranges in matrix K, by that analogy.
Gesture result is exported, exports current hand gestures;Specially export two-dimentional gesture model and or or three-dimension gesture
Model.
A kind of electronic equipment, including:Processor;Memory;And program, its Program are stored in memory, and
And be configured to by computing device, program includes being used to perform the real-time hand recognition methods based on RGB information.One kind calculates
Machine readable storage medium storing program for executing, is stored thereon with computer program:Computer program is executed by processor the real time hand based on RGB information
Portion's recognition methods.
In theory, video camera number is The more the better, because number of cameras is more, the information that the same time collects is got over
More, the degree of accuracy is naturally higher, but operand also can accordingly increase.Fig. 5 is varying number camera being averaged under equal ambient
Error.In one embodiment, using CPU:Intel Xeon E5-1620 (3.60GHz), internal memory:16GB, hard disk:512G SSD
Hardware device tested, tested using the data in the dataset of public data collection Dexter 1, respectively to SoG and
The SAG that this patent uses is tested, as a result such as Fig. 6, wherein, first row represents, the error less than how many millimeters just thinks to tie
Fruit is correct, the first row represent in data set different Sub Data Set title be respectively adbadd represent the identification of hand comprehensive morphological,
Fingercount represents that fingers number identification, fingerwave represent that finger movement Path Recognition, flexex1 represent that finger is curved
Curvature identification, pinch represent that grasping movement identification, random hands free movement identification, tigergrasp represent very fast crawl
Identification, numerical value represent accuracy rate, unit %.Overstriking is comparatively speaking more excellent result.By result it can be seen that,
The present invention has very big raising compared to SoG, and is acknowledged as gathering the most accurate algorithm of hand data before SoG.
Real-time hand recognition methods of the invention based on RGB information, including obtain real-time RGB image group, two dimensional image point
Cut, establish three-dimension gesture model, gesture judges, output result.The invention further relates to storage medium and electronic equipment.The present invention adopts
The important joint of the hand of people is equivalent to ellipsoid model with anisotropic Gaussian and model, passes through the image collector of multi-angle
Collection hand realtime graphic is put, realizes the quick identification of hand gestures.The present invention is real-time, degree of accuracy height, strong robustness, prolongs
When it is low, the complicated hand motion of identification can be realized.
More than, only presently preferred embodiments of the present invention, any formal limitation not is made to the present invention;All one's own professions
The those of ordinary skill of industry can swimmingly implement the present invention shown in by specification accompanying drawing and above;But all to be familiar with sheet special
The technical staff of industry without departing from the scope of the present invention, is made a little using disclosed above technology contents
The equivalent variations of variation, modification and evolution, it is the equivalent embodiment of the present invention;Meanwhile all substantial technologicals according to the present invention
Variation, modification and evolution of any equivalent variations made to above example etc., still fall within technical scheme
Within protection domain.
Claims (10)
1. the real-time hand recognition methods based on RGB information, it is characterised in that comprise the following steps:
Real-time RGB image group is obtained, image acquisition device hand multi-angle image, obtains hand two dimension RGB image group;
Two dimensional image is split, and image segmentation is carried out successively to the image in the RGB image group;
Three-dimension gesture model is established, the RGB image after splitting according to image sets up vertical three-dimension gesture model;
Gesture judges, using anisotropic Gaussian and model in the three-dimension gesture model and hand three-dimensional modeling data storehouse
Judgement is identified in model, obtains current hand gestures;
Gesture result is exported, exports the current hand gestures.
2. the real-time hand recognition methods based on RGB information as claimed in claim 1, it is characterised in that also including step 3
Dimension hand gestures switch to two-dimentional hand gestures, and the three-dimension gesture model is projected to two-dimensional projection face and obtains two-dimentional gesture mould
Type.
3. the real-time hand recognition methods based on RGB information as claimed in claim 2, it is characterised in that described step is defeated
It is specially to export the two-dimentional gesture model and or the or three-dimension gesture model to go out gesture result.
4. the real-time hand recognition methods based on RGB information as claimed in claim 1, it is characterised in that described step two
Dimension image segmentation specifically carries out image segmentation successively using two-dimentional quaternary tree to the image in the RGB image group.
5. the real-time hand recognition methods based on RGB information as claimed in claim 1, it is characterised in that described step is built
Vertical three-dimension gesture model is specifically entered using action matching energy maximum algorithm to the image of the RGB image group after image segmentation
Row matching, establishes three-dimension gesture model.
6. the real-time hand recognition methods based on RGB information as claimed in claim 1, it is characterised in that:Described three-dimensional hand
Potential model is carried out equivalent using the spheroid of 17 anisotropic Gaussian features.
7. the real-time hand recognition methods based on RGB information as claimed in claim 1, it is characterised in that:Described image is adopted
Packaging is set to video camera, and the frequency acquisition of described video camera is more than 20 frame per second.
8. the real-time hand recognition methods based on RGB information as claimed in claim 1, it is characterised in that:Described image is adopted
The quantity of acquisition means is at least 2.
9. a kind of electronic equipment, it is characterised in that including:Processor;Memory;And program, wherein described program are stored in
In the memory, and it is configured to by computing device, described program includes being used for perform claim requirement 1-6 any one
Described method.
10. a kind of computer-readable recording medium, is stored thereon with computer program, it is characterised in that:The computer program
It is executed by processor method as claimed in any one of claims 1 to 6.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710549412.4A CN107479693A (en) | 2017-07-07 | 2017-07-07 | Real-time hand recognition methods based on RGB information, storage medium, electronic equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710549412.4A CN107479693A (en) | 2017-07-07 | 2017-07-07 | Real-time hand recognition methods based on RGB information, storage medium, electronic equipment |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107479693A true CN107479693A (en) | 2017-12-15 |
Family
ID=60596142
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710549412.4A Pending CN107479693A (en) | 2017-07-07 | 2017-07-07 | Real-time hand recognition methods based on RGB information, storage medium, electronic equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107479693A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108399367A (en) * | 2018-01-31 | 2018-08-14 | 深圳市阿西莫夫科技有限公司 | Hand motion recognition method, apparatus, computer equipment and readable storage medium storing program for executing |
CN109461203A (en) * | 2018-09-17 | 2019-03-12 | 百度在线网络技术(北京)有限公司 | Gesture three-dimensional image generating method, device, computer equipment and storage medium |
CN109934065A (en) * | 2017-12-18 | 2019-06-25 | 虹软科技股份有限公司 | A kind of method and apparatus for gesture identification |
CN110147767A (en) * | 2019-05-22 | 2019-08-20 | 深圳市凌云视迅科技有限责任公司 | Three-dimension gesture attitude prediction method based on two dimensional image |
WO2020147598A1 (en) * | 2019-01-15 | 2020-07-23 | 北京字节跳动网络技术有限公司 | Model action method and apparatus, speaker having screen, electronic device, and storage medium |
US11500976B2 (en) | 2020-11-03 | 2022-11-15 | Nxp B.V. | Challenge-response method for biometric authentication |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102893182A (en) * | 2010-02-22 | 2013-01-23 | 兰德马克绘图国际公司 | Systems and methods for modeling 3d geological structures |
CN103733226A (en) * | 2011-05-16 | 2014-04-16 | 马普科技促进协会 | Fast articulated motion tracking |
-
2017
- 2017-07-07 CN CN201710549412.4A patent/CN107479693A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102893182A (en) * | 2010-02-22 | 2013-01-23 | 兰德马克绘图国际公司 | Systems and methods for modeling 3d geological structures |
CN103733226A (en) * | 2011-05-16 | 2014-04-16 | 马普科技促进协会 | Fast articulated motion tracking |
Non-Patent Citations (2)
Title |
---|
SRIDHAR1 ETC.: "Real-time Hand Tracking Using a Sum of Anisotropic Gaussians Model", 《2014 SECOND INTERNATIONAL CONFERENCE ON 3D VISION》 * |
魏诗白: "基于混合高斯模型的三维手部关节跟踪算法", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109934065A (en) * | 2017-12-18 | 2019-06-25 | 虹软科技股份有限公司 | A kind of method and apparatus for gesture identification |
CN109934065B (en) * | 2017-12-18 | 2021-11-09 | 虹软科技股份有限公司 | Method and device for gesture recognition |
CN108399367A (en) * | 2018-01-31 | 2018-08-14 | 深圳市阿西莫夫科技有限公司 | Hand motion recognition method, apparatus, computer equipment and readable storage medium storing program for executing |
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 |
CN109461203A (en) * | 2018-09-17 | 2019-03-12 | 百度在线网络技术(北京)有限公司 | Gesture three-dimensional image generating method, device, computer equipment and storage medium |
WO2020147598A1 (en) * | 2019-01-15 | 2020-07-23 | 北京字节跳动网络技术有限公司 | Model action method and apparatus, speaker having screen, electronic device, and storage medium |
CN110147767A (en) * | 2019-05-22 | 2019-08-20 | 深圳市凌云视迅科技有限责任公司 | Three-dimension gesture attitude prediction method based on two dimensional image |
CN110147767B (en) * | 2019-05-22 | 2023-07-18 | 深圳市凌云视迅科技有限责任公司 | Three-dimensional gesture attitude prediction method based on two-dimensional image |
US11500976B2 (en) | 2020-11-03 | 2022-11-15 | Nxp B.V. | Challenge-response method for biometric authentication |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106682598B (en) | Multi-pose face feature point detection method based on cascade regression | |
CN107479693A (en) | Real-time hand recognition methods based on RGB information, storage medium, electronic equipment | |
Gilani et al. | Deep, dense and accurate 3D face correspondence for generating population specific deformable models | |
Ye et al. | Accurate 3d pose estimation from a single depth image | |
Ren et al. | Depth camera based hand gesture recognition and its applications in human-computer-interaction | |
Ma et al. | Kinect Sensor‐Based Long‐Distance Hand Gesture Recognition and Fingertip Detection with Depth Information | |
CN104392223B (en) | Human posture recognition method in two-dimensional video image | |
Cheng et al. | Image-to-class dynamic time warping for 3D hand gesture recognition | |
CN105759967B (en) | A kind of hand overall situation attitude detecting method based on depth data | |
Qiang et al. | SqueezeNet and fusion network-based accurate fast fully convolutional network for hand detection and gesture recognition | |
Ding et al. | STFC: Spatio-temporal feature chain for skeleton-based human action recognition | |
CN108182397B (en) | Multi-pose multi-scale human face verification method | |
Li et al. | Hand gesture recognition enhancement based on spatial fuzzy matching in leap motion | |
JP2015510203A (en) | Improvements in or related to 3D proximity interaction | |
JP2019096113A (en) | Processing device, method and program relating to keypoint data | |
CN108573231B (en) | Human body behavior identification method of depth motion map generated based on motion history point cloud | |
Weiyao et al. | Human action recognition using multilevel depth motion maps | |
JP2016014954A (en) | Method for detecting finger shape, program thereof, storage medium of program thereof, and system for detecting finger shape | |
CN109543644A (en) | A kind of recognition methods of multi-modal gesture | |
CN108564043B (en) | Human body behavior recognition method based on space-time distribution diagram | |
Ming et al. | A unified 3D face authentication framework based on robust local mesh SIFT feature | |
Yang et al. | 3D character recognition using binocular camera for medical assist | |
Lan et al. | Data fusion-based real-time hand gesture recognition with Kinect V2 | |
Liu et al. | Energy-based global ternary image for action recognition using sole depth sequences | |
CN105955473A (en) | Computer-based static gesture image recognition interactive system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171215 |