CN1123321C - Human hand movement image 3D real time testing method - Google Patents
Human hand movement image 3D real time testing method Download PDFInfo
- Publication number
- CN1123321C CN1123321C CN 00129552 CN00129552A CN1123321C CN 1123321 C CN1123321 C CN 1123321C CN 00129552 CN00129552 CN 00129552 CN 00129552 A CN00129552 A CN 00129552A CN 1123321 C CN1123321 C CN 1123321C
- Authority
- CN
- China
- Prior art keywords
- index point
- human hand
- image
- hand movement
- real time
- 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The present invention relates to a three-dimensional real-time detection method for human hand movement images. The method comprises the following steps: firstly, a signalized point is glued on a joint on a human hand to be detected to make the human hand act accordingly to requirements; human hand movement process images are shot from different angles and collected into a computer in real time, and filtration, mapping, non-linear transformation and fuzzy enhancement are carried out on the images; the images are scanned to find the signalized point; the two-dimensional central coordinates of the signalized point are calculated, then three-dimensional reconstruction is carried out, and the human hand movement images can be obtained. The spatial movement tracks of different human hand action can be detected and recorded in real time by the method in order to research the rule and the control mechanism of human hand movement, or the present invention is used for clinical medicine.
Description
The present invention relates to a kind of human hand movement image 3 D real time testing method, belong to biomedical engineering field.
Staff has 20 degree of freedom, it is the organ the most flexibly that moves in the human body, its motion needs muscle group, skeleton and meticulous control degree more more complex than other motions such as lower limb, staff is a most important operation organ when contacting with external environment in the daily live and work of people, numerous disease adversarys' such as parkinson, apoplexy, cerebral trauma motor function all has certain influence, therefore in clinical rehabilitation usually the important means of the training of human hand movement function as some disease patient rehabilitation.Visible detection is analyzed the motion feature of staff, can help the clinician to carry out disease or the evaluation of rehabilitation process, is clinical diagnosis and the evaluation means that accurate science is provided.
Hand movement detecting method mainly contains three major types: a class is to utilize equipment such as X-ray, CT, MRI to obtain staff skeleton and each cross-sectional image, and it is higher that this class methods detect cost, and generally can only detect the situation of hands when static; Second class is to utilize clinometer to detect the variation of joint angles, as the data glove that in virtual reality, is widely adopted, these class methods are owing to the two ends of angular transducer are placed on the joint, not only produce to slide and introduce error with skin surface, and influencing the natural motion of staff, the error in data that promptly utilizes these class methods to obtain is bigger; The 3rd class methods are to utilize camera system, by picked-up human hand movement process, carry out analyzing and processing then, and the precision of these class methods is higher, are the important tool of carrying out human hand movement research and relevant disease diagnosis.
The objective of the invention is to propose a kind of human hand movement image 3 D real time testing method, at present clinical generally the employing by doctor's range estimation carried out the human hand movement functional diagnosis and evaluation provides a kind of scientific method.
A kind of human hand movement image 3 D real time testing method that the present invention proposes is characterized in that may further comprise the steps:
(1) on the joint of measured's hands to be measured, pastes index point, measured's hands is moved on request;
(2) absorb the human hand movement procedural image from different perspectives, and image is collected computer in real time;
(3) image is carried out filtering with median filtering method, filtered image is mapped as fuzzy matrix, fuzzy matrix is made nonlinear transformation, blur enhancing, then the image after strengthening is done inverse mapping, obtain fuzzy enhanced image;
(4) above-mentioned image is scanned, search out all connected domains according to ganmma controller; Search the sealing connected domain and obtain the outline full-size of sealing connected domain; Sealing connected domain outer profile size screened obtain and the consistent index point of former stickup index point size range;
(5) to finding the 1st and the 2nd width of cloth motor process image behind the index point, progressively extrapolate from beginning flag point and to search action back and put corresponding index point with original logo, all images later to the 3rd width of cloth are put corresponding index point with original logo behind the single order extrapolation location lookup subsequent action of the index point that pro-two width of cloth are determined;
(6) calculate the two-dimentional centre coordinate of index point according to the shape of the above-mentioned index point of searching;
(7) according to the above-mentioned two-dimensional coordinate that absorbs the same index point that the human hand movement plane picture obtains from different perspectives, adopt direct linear transformation's (being abbreviated as DLT, Direct Linear Transformation) method to solve the index point three dimensional space coordinate.
(8) can obtain human hand movement rod figure by be linked in sequence joint or the terminal three dimensional space coordinate of going up index point of dissection; The three dimensional space coordinate that connects same index point in the different images by the sampling order just obtains the displacement movement curve in joint; Two straight lines according to per 3 somes decision among the motion rod figure can calculate two angles that straight line is folded, i.e. the angle of arthrogryposis.According to displacement and angular displacement, utilize differential method can obtain the speed of its motion and the Changing Pattern of angular velocity, acceleration and angular acceleration.
The method detection record staff work difference action in real time time space movement locus that utilize the present invention to propose, and it is carried out various statistics, with research human hand movement rule and control mechanism, or on clinical medicine quantitative assessment human hand movement function and rehabilitation degree thereof.
Description of drawings:
Fig. 1 staff is to pinching motion three dimensions rod figure, and dimension is millimeter among the figure.
Angular displacement (scalar) curve in Fig. 2 forefinger the 2nd joint,
Fig. 3 forefinger the 2nd joint angle speed change curves
Fig. 4 forefinger the 2nd joint angle acceleration change curve
Below in conjunction with accompanying drawing, introduce content of the present invention in detail:
(1) index point of embodiment is selected black ball for use, and paste position is thumb and forefinger tip, each joint of thumb and forefinger, and make the measured make thumb and forefinger is done pinching;
(2) absorb the human hand movement procedural image from different perspectives, and image real-time acquisition is arrived computer;
(3) image is carried out filtering with median filtering method, filtered image is mapped as fuzzy matrix, fuzzy matrix is made nonlinear transformation, blur enhancing, then the image after strengthening is done inverse mapping, obtain fuzzy enhanced image;
Its concrete enhancing process is as follows:
With the image representation of width of cloth L gray level M * N is X=[l
Mn], m=1,2 ... M, n=1,2 ... N (l wherein
MnBe visual pixel (m, gray level n)) that corresponding fuzzy matrix is
(μ wherein
MnBe l
MnDegree of membership with respect to certain particular gray level).Getting membership function is:
Get
Fuzzy enhancing function be:
Wherein
Be fuzzy enhanced number of times, μ
cBe fuzzy enhanced threshold value, its corresponding ganmma controller value is l
cThe fuzzy image that strengthens back matrix correspondence of note is
L wherein
Mn' for fuzzy strengthen the back pixel (m, gray level n), then
Wherein
Can finish the fuzzy enhancing of image by (1), (2) two formulas.In the specific implementation of algorithm, for reducing amount of calculation, it is as follows in advance formula (2) to be made form Table1, Table2:
Table1[i]=(int)(i×i/l
c+0.5) i=0、1、2、…l
c
Table2[i]=(int)(L-1-(L-1-i)
2/(L-1-l
c)+0.5) i=l
c+1、l
c+2、…L-1
1) search index point:
Pointwise scans image and makes comparisons with threshold value, if the continuity point that satisfies threshold condition is arranged, then write down its parameter line (row number), row (initial point range number) and long (continuity point number), such continuity point is exactly " bar ", and by following 3 structure of transvers plate connected domains.
Template 1: if certain " bar " do not link to each other with any existing territory, think that then this " bar " belongs to neofield, and set up a neofield and deposit this " bar " parameter (row, column, length) in.
Template 2:, think that then this " bar " belongs to coupled territory, and should " bar " add its continuous territory if certain " bar " only links to each other with an existing territory.
Template 3: if certain " bar " links to each other with a plurality of existing territories, think that then these territories are a neofield, they are merged into a territory, wherein each " bar " recombinated by the row, column order and should be discharged each original territory simultaneously by " bar " adding neofield.
Each " bar " of scanning and above-mentioned 3 templates are compared, find the template and of conforming to, can realize the automatic detection of arbitrary shape connected domain by template operation.Owing to remain each " bar " in the whole process by the row, column sequence arrangement in the territory in, when therefore searching, need make comparisons with " bar " of each territory last column as template, amount of calculation is very little.In the specific implementation of algorithm, be further to reduce amount of calculation and to save internal memory, the definition of " bar " has been added stricter restriction, have only those length to surpass MIN continuity point and just be considered to " bar ".When such qualification can avoid image " pit " to occur, identify the connected domain of too much null(NUL).As long as connected domain contour dimension and known markers point gabarit size coupling think that promptly this connected domain is an index point.
2) near tracking mark point:, beginning flag point, search index point to the 1st and 2 images; To the later image of the 3rd width of cloth figure, search index point near the outer push position of the single order of index point among pro-two width of cloth figure, find out the same index point in the different images, promptly carry out index point and follow the tracks of.
3) calculate the geometric center coordinate of index point according to the shape of index point;
4) three-dimensionalreconstruction
According to the two-dimensional coordinate that absorbs the same index point that the human hand movement plane picture obtains from different perspectives, adopt direct linear transformation's (being abbreviated as DLT, Direct Linear Transformation) method to calculate the index point three dimensional space coordinate.Its principle is: suppose object coordinates (X, Y is Z) with picture coordinate (η, mapping relations are the function that contains some undetermined coefficients ζ), according to one group of known (X, Y, Z, η ζ) can solve each undetermined parameter in the mapping function, thereby obtain (X, Y, Z) with (this step is demarcation for η, mapping relations ζ).Adopt the multi-section camera, demarcate the image mapping function that obtains separately respectively, according to the object space point (X, Y, Z) the picture coordinate of each camera (η, ζ) and the image mapping function that has calibrated can simultaneous solve object coordinates (X, Y Z), promptly realize three-dimensionalreconstruction.
(1) is linked in sequence by dissection and sticks on the joint or the terminal three dimensional space coordinate of going up index point can obtain human hand movement rod figure, as connecting successively on forefinger tip, each joint of forefinger, and the three dimensional space coordinate of index point on the thumbtip, each joint of thumb, promptly obtain the three-dimensional rod figure of motion of Fig. 1; Can obtain the space displacement curve movement (see figure 1) in joint by the three dimensional space coordinate of sampling order connection forefinger tip index point; Two straight lines according to per 3 somes decision among the motion rod figure, can calculate two angles that straight line is folded, as coordinate according to index point on forefinger the 1st, 2 and 3 joints, can calculate the angle in the 2nd joint, chronological order by sampling uniformly-spaced connects the angle in the 2nd joint, promptly obtains Fig. 2 the 2nd joint angle displacement (scalar) change curve.According to displacement and angular displacement, utilize differential method can obtain the speed of its motion and the Changing Pattern of angular velocity, acceleration and angular acceleration, for example Fig. 2 the 2nd joint angle displacement (scalar) change curve is obtained Fig. 3 angular velocity varies curve as differential, Fig. 2 the 2nd joint angle speed change curves is obtained Fig. 4 angular acceleration change curve as differential.
(2) aforementioned calculation result and curve are carried out statistical disposition, can obtain staff when making different typical action, the characteristics of motion of different dactylus.By various computing result and the curve that obtains carried out various combinative analysiss, adjuster's hands movement function quantitatively.
The measured is a healthy male among the embodiment, and Fig. 1-Fig. 4 is the result that the human hand movement image 3 D real time testing method that utilizes the present invention to propose obtains.
Claims (1)
1, a kind of human hand movement image 3 D real time testing method is characterized in that this method may further comprise the steps:
On the joint of measured's hands to be measured, paste index point, measured's hands is moved on request;
Absorb the human hand movement procedural image from different perspectives, and image is collected computer in real time;
Image is carried out filtering with median filtering method, filtered image is mapped as fuzzy matrix, fuzzy matrix is made nonlinear transformation, blur enhancing, then the image after strengthening is done inverse mapping, obtain fuzzy enhanced image;
Above-mentioned image is scanned, search out all connected domains according to ganmma controller; Search the sealing connected domain and obtain the outline full-size of sealing connected domain; Sealing connected domain outer profile size screened obtain and the consistent index point of former stickup index point size range;
To finding the 1st and the 2nd width of cloth motor process image behind the index point, progressively extrapolate from beginning flag point and to search action back and put corresponding index point with original logo, all images later to the 3rd width of cloth are put corresponding index point with original logo behind the single order extrapolation location lookup subsequent action of the index point that pro-two width of cloth are determined;
Calculate the two-dimentional centre coordinate of index point according to the shape of the above-mentioned index point of searching;
According to the above-mentioned two-dimensional coordinate that absorbs the same index point that the human hand movement plane picture obtains from different perspectives, adopt direct linear transformation's method to solve the index point three dimensional space coordinate;
Promptly obtain human hand movement rod figure by be linked in sequence joint or the terminal three dimensional space coordinate of going up index point of dissection; The three dimensional space coordinate that connects same index point in the different images by the sampling order just obtains the displacement movement curve in joint; Two straight lines according to per 3 somes decision among the motion rod figure calculate two angles that straight line is folded, i.e. the angle of arthrogryposis.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 00129552 CN1123321C (en) | 2000-10-09 | 2000-10-09 | Human hand movement image 3D real time testing method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 00129552 CN1123321C (en) | 2000-10-09 | 2000-10-09 | Human hand movement image 3D real time testing method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN1285504A CN1285504A (en) | 2001-02-28 |
CN1123321C true CN1123321C (en) | 2003-10-08 |
Family
ID=4593583
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN 00129552 Expired - Fee Related CN1123321C (en) | 2000-10-09 | 2000-10-09 | Human hand movement image 3D real time testing method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN1123321C (en) |
Families Citing this family (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
SG115546A1 (en) * | 2003-06-23 | 2005-10-28 | Affineon Technologies Pte Ltd | Computer input device tracking six degrees of freedom |
CN100409156C (en) * | 2006-01-23 | 2008-08-06 | 宏达国际电子股份有限公司 | Movement judging device |
CN100403313C (en) * | 2006-09-14 | 2008-07-16 | 浙江大学 | Extraction method of key frame of 3d human motion data |
CN100454335C (en) * | 2006-10-23 | 2009-01-21 | 华为技术有限公司 | Realizing method for forming three dimension image and terminal device |
CN101832756B (en) * | 2009-03-10 | 2014-12-10 | 深圳迈瑞生物医疗电子股份有限公司 | Method and device for measuring displacement of targets in images and carrying out strain and strain rate imaging |
CN102227616B (en) * | 2009-05-27 | 2015-04-15 | 松下电器产业株式会社 | Behavior recognition device |
CN104887238A (en) * | 2015-06-10 | 2015-09-09 | 上海大学 | Hand rehabilitation training evaluation system and method based on motion capture |
CN105824417B (en) * | 2016-03-16 | 2019-12-10 | 成都电锯互动科技有限公司 | human-object combination method adopting virtual reality technology |
CN105945945B (en) * | 2016-05-20 | 2018-12-11 | 哈尔滨工业大学 | Finger module division methods based on human hand movement functional analysis |
CN109658627A (en) * | 2018-12-13 | 2019-04-19 | 深圳桓轩科技有限公司 | A kind of Intelligent logistics pickup system based on block chain |
CN109584423A (en) * | 2018-12-13 | 2019-04-05 | 佛山单常科技有限公司 | A kind of intelligent unlocking system |
CN113255462A (en) * | 2021-04-29 | 2021-08-13 | 深圳大学 | Gait scoring method, system, computer program product and readable storage medium |
CN113838219B (en) * | 2021-09-26 | 2023-09-12 | 琼台师范学院 | Virtual dance training method and device based on human motion capture |
CN113807323B (en) * | 2021-11-01 | 2022-12-09 | 北京大学 | Accurate hand function evaluation system and method based on image recognition |
-
2000
- 2000-10-09 CN CN 00129552 patent/CN1123321C/en not_active Expired - Fee Related
Also Published As
Publication number | Publication date |
---|---|
CN1285504A (en) | 2001-02-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN1123321C (en) | Human hand movement image 3D real time testing method | |
CN111432733B (en) | Apparatus and method for determining motion of an ultrasound probe | |
Meyer et al. | Semiautomatic registration of volumetric ultrasound scans | |
DE112004002219B4 (en) | Method for driving a robot | |
EP2227703B1 (en) | Method for detecting motion | |
Parameswaran et al. | View independent human body pose estimation from a single perspective image | |
DE112004000722T5 (en) | Method and apparatus for fast automatic centerline extraction for virtual endoscopy | |
AlZu’bi et al. | Multi-orientation geometric medical volumes segmentation using 3d multiresolution analysis | |
CN107714082A (en) | A kind of ultrasonic probe caliberating device and method based on optical alignment | |
CN101033963A (en) | Location system of video finger and location method based on finger tip marking | |
CN109410157A (en) | The image interfusion method with PCNN is decomposed based on low-rank sparse | |
CN1251145C (en) | Pyramid image merging method being integrated with edge and texture information | |
CN106327479A (en) | Apparatus and method for identifying blood vessels in angiography-assisted congenital heart disease operation | |
Pinzke et al. | Marker-less systems for tracking working postures—results from two experiments | |
CN104732586B (en) | A kind of dynamic body of 3 D human body and three-dimensional motion light stream fast reconstructing method | |
CN1950849A (en) | Pharmacokinetic image registration | |
CN1286962A (en) | Real-time body gait image detecting method | |
CN106709867A (en) | Medical image registration method based on improved SURF and improved mutual information | |
CN113033501A (en) | Human body classification method and device based on joint quaternion | |
CN1275205C (en) | Human limbs three-dimensional model construction based on muscle volume invariance | |
CN1581233A (en) | Method for drawing and image-forming using the body to be conducted simulation operation | |
CN110276762A (en) | A kind of full-automatic bearing calibration of respiratory movement of the diffusion-weighted Abdominal MRI imaging of more b values | |
Shi et al. | Driving motion capture based driver behavior analysis | |
CN113091693A (en) | Monocular vision long-range distance measurement method based on image super-resolution technology | |
Shin et al. | 3D tracking of surgical instruments using a single camera for laparoscopic surgery simulation |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C06 | Publication | ||
PB01 | Publication | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
C19 | Lapse of patent right due to non-payment of the annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |