CN110705465B - Hand type classification method based on image processing - Google Patents
Hand type classification method based on image processing Download PDFInfo
- Publication number
- CN110705465B CN110705465B CN201910937840.3A CN201910937840A CN110705465B CN 110705465 B CN110705465 B CN 110705465B CN 201910937840 A CN201910937840 A CN 201910937840A CN 110705465 B CN110705465 B CN 110705465B
- Authority
- CN
- China
- Prior art keywords
- hand
- image
- palm
- classification method
- type classification
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/107—Static hand or arm
- G06V40/113—Recognition of static hand signs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
Abstract
The invention discloses a hand shape classification method based on image processing, and particularly relates to a hand image processing technology. The method comprises the following steps: acquiring a hand image of a tested person; establishing a skin color model and selecting a threshold value to carry out binarization on the image; performing edge detection on the binary image, and storing the obtained palm contour image; performing open operation on the palm contour image to obtain a smooth palm contour image; finding out the convex hull of the palm outline, and calculating the length of the palm (L 1 ) Length of middle finger: (L 2 ) And palm width (L 3 ) (ii) a Calculating hand characteristic valuestype(ii) a According to the threshold value, the hand shape of the tested person is determined to be one of three hand shapes of slender, normal and round. The invention solves the problems that the traditional hand type classification method mainly depends on manual measurement or manual observation, the manual measurement wastes time and labor, the error is large, the subjectivity of manual observation is strong, and the standard is not uniform, and provides a rapid, accurate and objective hand type classification method.
Description
Technical Field
The invention relates to a hand shape classification method based on image processing, in particular to an image processing technology and calculation of hand characteristic values.
Background
In the past decades, computer technology has developed rapidly, and in various industries, people rely on computers more and more heavily. Therefore, the information interaction between a person and a computer is increasingly emphasized. At present, the form of human-computer interaction enters a multi-channel and multi-modal stage, and gesture recognition is a mainstream trend at the present stage. However, the gesture recognition algorithms of the prior art directly extract the gesture features of the user, neglect the difference between each hand type, and may make an erroneous judgment when different hand types use the same standard value. If can judge earlier before carrying out gesture recognition which kind of hand type the user is, discern the gesture again, just can reach the effect that improves gesture recognition accuracy. Hand type classification may also be applied in many industries, such as customizing gloves according to hand types of different consumers, customizing specific living goods, and may study differences in the way different hands develop in sports.
Therefore, the application prospect of the hand type classification is very wide, the traditional hand type classification method mainly depends on manual measurement or manual observation, the manual measurement wastes time and labor, the error is large, the subjectivity of the manual observation is strong, and the standard is not uniform, so that the design of a rapid, accurate and objective hand type classification method is very important.
Disclosure of Invention
The invention provides a hand type classification method based on image processing, which can classify the hand type of a person. The method mainly comprises the following steps: the method comprises the steps of collecting a complete hand image of a tested person, carrying out image processing to obtain a palm outline image, calculating a hand characteristic value, and determining the hand shape of the tested person according to a threshold value.
The invention is realized by adopting the following method: a hand type classification method based on image processing comprises the following steps:
step one, carrying out image acquisition on the hand of a tested person to obtain a complete hand image;
establishing a skin color model for the collected hand image, selecting a threshold value for binarization, and storing the obtained hand binarized image;
step three, performing edge detection on the binary image obtained in the step two, and storing the obtained palm contour image;
performing open operation on the palm contour image obtained in the step three to obtain a smooth palm contour image;
step five, respectively calculating the length (L) of the palm1) Length of middle finger (L)2) And width of palm (L)3);
Step six, representing the characteristic value of the hand shape by type, and calculating the value of the type;
step seven, according to the threshold value T1And T2And determining the hand shape of the tested person.
In the first step, five fingers of the right hand of the volunteer are naturally opened, the palm is horizontally placed on a desktop, a camera positioned right above the palm of the tested person vertically shoots downwards, the palm and the wrist can be guaranteed to be shot, a hand image is obtained, and image resolution ratios of the hand of the tested person are unified in size when the hand is subjected to image acquisition.
And in the second step, the hand image is binarized, because the skin color of the hand has better stability, the influence of detail characteristic change of the hand on the skin color is very small, and Cb and Cr values and distribution of the skin color of the hand of the same family are basically consistent in a YCbCr color space and are concentrated in a small range. And converting the shot hand picture from an RGB color space to a YCbCr color space according to a formula (1) to obtain Cb and Cr values of each pixel point in the pictures.
Then according to the formula (2), selecting a proper threshold value to complete the binarization of the image, and setting the obtained image as Q:
where Y is the brightness of the color, Cb and Cr represent the concentration offset components of blue and red, respectively, and a1、a2Upper and lower thresholds representing Y, b1、b2Represents the upper and lower thresholds of Cb, c1、c2Represents the upper and lower threshold values of Cr.
Taking the yellow race as an example, the Y value corresponding to the skin color pixel of the hand of the yellow race is mainly distributed in the [50, 255] interval, the Cb value is mainly distributed in the [87, 142] interval, and the Cr value is mainly distributed in the [132, 151] interval, then:
and in the fourth step, by utilizing an open operation method, burrs on the palm contour in the image can be removed through corrosion operation and then expansion operation, so that a smooth palm contour image is obtained.
Measuring L in step five1、L2And L3The method comprises the following steps:
a) and acquiring the peripheral outline convex hull of the whole hand. In image processing, a convex hull can be regarded as a convex set surrounding the outermost layer of an image;
b) determining convex hull defects and defect starting points;
c) finding out the relative position of the palm and the fingers, and calibrating the central point and the outline of the palm;
d) obtaining the coordinates of the lowest point of the palm outline by using the palm center point and the palm radius, and eliminating the wrist partial image below the lowest point;
e) obtaining the middle finger tip A1The lowest point A of the palm contour2Convex hull defect A of middle finger3And palm center point A0The vertical coordinates of the four points are used for calculating L1、L2And L3Then, there are:
wherein y is1Is A1Ordinate, y, of the point2Is A2Ordinate, y, of the point3Is A3Ordinate, y, of the point0Is A0The ordinate of the point.
Through a large number of sample statistics, when L2And L1Multiplying the ratio of (A) by 0.3, L3And L1When the ratio of (2) is multiplied by 0.7, the type values of different hand types are changed maximally, so that the hand types can be reflected more accurately, and a calculation formula of the type in the step six is determined:
and (5) determining that the hand shape of the tested person is one of slender, normal and round hand shapes.
Threshold T for determining hand shape by mass sample statistics1=0.46,T2=0.58。
The invention has the beneficial effects that:
the invention provides a hand type classification method based on image processing, which can determine that the hand type of a tested person is one of a slender hand type, a normal hand type and a round hand type. The method can be applied to the fields of improving gesture recognition accuracy, customizing gloves and articles for daily use, researching hands by medicine and motion and the like. The method solves the problems that the traditional hand type classification method mainly depends on manual measurement or manual observation, the manual measurement wastes time and labor, the error is large, the subjectivity of manual observation is strong, and the standard is not uniform, and provides a rapid, accurate and objective hand type classification method.
Drawings
Fig. 1 is a flowchart of an embodiment of a hand type classification method based on image processing according to the present invention.
Fig. 2 is a schematic diagram of a hand image capturing system for acquiring a hand image based on a hand shape classification method of image processing according to the present invention.
FIG. 3 is a diagram illustrating a hand shape classification method L based on image processing according to the present invention1、L2And L3The specific measuring method and the relative position of each point.
FIG. 4 is a diagram illustrating a hand shape classification method measurement L based on image processing according to the present invention1、L2And L3In the example (1).
Detailed Description
The invention provides an embodiment of a hand type classification method based on image processing. In order to make the technical solutions in the embodiments of the present invention better understood and make the above objects, features, and advantages of the present invention more comprehensible, the technical solutions of the present invention are described in further detail below with reference to the accompanying drawings:
detailed description of the invention
The invention firstly provides a hand type classification method based on image processing, as shown in fig. 1, the specific method is as follows:
step one, carrying out image acquisition on the hand of a tested person to obtain a complete hand image;
establishing a skin color model for the collected hand image, selecting a threshold value for binarization, and storing the obtained hand binarized image;
step three, performing edge detection on the binarized image obtained in the step two, and storing the obtained palm contour image, wherein one obtaining method can be realized through a Canny operator in OpenCV;
performing open operation on the palm contour image obtained in the step three to obtain a smooth palm contour image;
step five, respectively calculating the length (L) of the palm1) Length of middle finger (L)2) And width of palm (L)3);
Step six, representing the characteristic value of the hand shape by type, and calculating the value of the type;
step seven, according to the threshold value T1And T2And determining the hand shape of the tested person.
Detailed description of the invention
Different from the first embodiment, in the first embodiment, a schematic view of a photographing system used for collecting the hand image of the subject in the step one is shown in fig. 2, and the specific photographing method includes: let the volunteer right hand five fingers open naturally, the palm is kept flat downwards in the appointed frame of desktop, and the camera that is located directly over the measurand palm shoots downwards perpendicularly, guarantees that palm and wrist can all be shot, obtains the hand image, adopts the image resolution ratio of uniform size when carrying out image acquisition to measurand's hand.
Detailed description of the invention
In the first embodiment, L1、L2And L3Fig. 3 shows a specific measurement method, and fig. 4 shows a measurement flowchart.
Measurement L1、L2And L3The method comprises the following steps:
a) and acquiring the peripheral outline convex hull of the whole hand. In image processing, a convex hull can be regarded as a convex set surrounding the outermost layer of an image, wherein one obtaining method can be realized through a covexhull function in OpenCV;
b) determination of convex hull defects (A)2) And the starting point of defect (A)1);
c) Finding the relative position of the palm and the fingers, and calibrating the central point (A) of the palm0) And profile (B)1);
d) Obtaining the coordinates of the lowest point of the palm outline by using the palm center point and the palm radius, and eliminating the wrist partial image below the lowest point;
e) obtaining the middle finger tip A1The lowest point A of the palm contour2Convex hull defect A of middle finger3And palm center point A0The vertical coordinates of the four points are used for calculating L1、L2And L3Then, there are:
wherein y is1Is A1Ordinate, y, of the point2Is A2Ordinate, y, of the point3Is A3Ordinate, y, of the point0Is A0The ordinate of the point.
Claims (9)
1. A hand type classification method based on image processing is characterized in that: the method is realized by the following steps:
step one, carrying out image acquisition on the hand of a tested person to obtain a complete hand image;
establishing a skin color model for the collected hand image, selecting a threshold value for binarization, and storing the obtained hand binarized image;
step three, performing edge detection on the binary image obtained in the step two, and storing the obtained palm contour image;
performing open operation on the palm contour image obtained in the step three to obtain a smooth palm contour image;
step five, respectively calculating the length L of the palm1Length of middle finger L2And palm width L3;
Step six, representing the characteristic value of the hand shape by type, and calculating the value of the type;
step seven, according to the threshold value T1And T2And determining the hand shape of the tested person.
2. The hand type classification method according to claim 1, characterized in that: the hand image acquisition method comprises the following steps:
let the volunteer right hand five fingers open naturally, the palm is kept flat on the desktop downwards to the palm, is located the perpendicular shooting downwards of camera directly over the surveyed palm, guarantees that palm and wrist can all be shot, obtains the hand image.
3. The hand type classification method according to claim 1, characterized in that: the method for binarizing the hand image comprises the following steps:
converting the shot hand image from an RGB color space to a YCbCr color space according to a formula (1) to obtain Cb and Cr values of each pixel point in the hand image;
then according to the formula (2), selecting a proper threshold value to complete the binarization of the image, and setting the obtained image as Q:
where Y is the brightness of the color, Cb and Cr represent the concentration offset components of blue and red, respectively, and a1、a2Upper and lower thresholds representing Y, b1、b2Represents the upper and lower thresholds of Cb, c1、c2Represents the upper and lower threshold values of Cr.
4. The hand type classification method according to claim 1, characterized in that: measurement L1、L2And L3The length method comprises the following steps:
a) acquiring a convex hull of the peripheral outline of the whole hand;
b) determining convex hull defects and defect starting points;
c) finding out the relative position of the palm and the fingers, and calibrating the central point and the outline of the palm;
d) obtaining the coordinates of the lowest point of the palm outline by using the palm center point and the palm radius, and eliminating the wrist partial image below the lowest point;
e) obtaining the middle finger tip A1The lowest point A of the palm contour2Convex hull defect A of middle finger3And palm center point A0The vertical coordinates of the four points are used for calculating L1、L2And L3Then, there are:
wherein y is1Is A1Ordinate, y, of the point2Is A2Ordinate, y, of the point3Is A3Ordinate, y, of the point0Is A0The ordinate of the point.
7. The hand type classification method according to claim 2, characterized in that: and image resolution with uniform size is adopted when the hand of the tested person is subjected to image acquisition.
8. The hand type classification method according to claim 3, characterized in that: take Huang race as an example, then there is a1=50、a2=255、b1=87、b2=142、c1=132、c2=151。
9. The hand type classification method according to claim 6, a threshold T for determining a hand type1Is 0.46, T2Is 0.58.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910937840.3A CN110705465B (en) | 2019-09-30 | 2019-09-30 | Hand type classification method based on image processing |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910937840.3A CN110705465B (en) | 2019-09-30 | 2019-09-30 | Hand type classification method based on image processing |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110705465A CN110705465A (en) | 2020-01-17 |
CN110705465B true CN110705465B (en) | 2021-05-18 |
Family
ID=69197448
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910937840.3A Expired - Fee Related CN110705465B (en) | 2019-09-30 | 2019-09-30 | Hand type classification method based on image processing |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110705465B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111563477A (en) * | 2020-05-21 | 2020-08-21 | 苏州沃柯雷克智能系统有限公司 | Method, device, equipment and storage medium for acquiring qualified hand photos |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101470800A (en) * | 2007-12-30 | 2009-07-01 | 沈阳工业大学 | Hand shape recognition method |
CN105615070A (en) * | 2014-11-04 | 2016-06-01 | 广西北海利博盛安全用品有限公司 | Methods for acquiring sizes of hands and manufacturing gloves |
CN106326860A (en) * | 2016-08-23 | 2017-01-11 | 武汉闪图科技有限公司 | Gesture recognition method based on vision |
CN108921011A (en) * | 2018-05-15 | 2018-11-30 | 合肥岚钊岚传媒有限公司 | A kind of dynamic hand gesture recognition system and method based on hidden Markov model |
CN109190460A (en) * | 2018-07-23 | 2019-01-11 | 南京航空航天大学 | Based on cumulative matches and etc. error rates hand shape arm vein fusion identification method |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8428311B2 (en) * | 2009-02-25 | 2013-04-23 | Honda Motor Co., Ltd. | Capturing and recognizing hand postures using inner distance shape contexts |
CN102760232A (en) * | 2012-08-02 | 2012-10-31 | 成都众合云盛科技有限公司 | Intermediate and long distance online identification system based on palm prints |
-
2019
- 2019-09-30 CN CN201910937840.3A patent/CN110705465B/en not_active Expired - Fee Related
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101470800A (en) * | 2007-12-30 | 2009-07-01 | 沈阳工业大学 | Hand shape recognition method |
CN105615070A (en) * | 2014-11-04 | 2016-06-01 | 广西北海利博盛安全用品有限公司 | Methods for acquiring sizes of hands and manufacturing gloves |
CN106326860A (en) * | 2016-08-23 | 2017-01-11 | 武汉闪图科技有限公司 | Gesture recognition method based on vision |
CN108921011A (en) * | 2018-05-15 | 2018-11-30 | 合肥岚钊岚传媒有限公司 | A kind of dynamic hand gesture recognition system and method based on hidden Markov model |
CN109190460A (en) * | 2018-07-23 | 2019-01-11 | 南京航空航天大学 | Based on cumulative matches and etc. error rates hand shape arm vein fusion identification method |
Non-Patent Citations (1)
Title |
---|
基于分类区分度和相关性的手形特征选择方法;苑玮琦 等;《仪器仪表学报》;20130831;第34卷(第8期);第1787-1794页 * |
Also Published As
Publication number | Publication date |
---|---|
CN110705465A (en) | 2020-01-17 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Aquino et al. | vitisBerry: An Android-smartphone application to early evaluate the number of grapevine berries by means of image analysis | |
JP4997252B2 (en) | How to identify the illumination area in an image | |
Zhang et al. | Tongue color analysis for medical application | |
CN106529559A (en) | Pointer-type circular multi-dashboard real-time reading identification method | |
CN106127108B (en) | A kind of manpower image region detection method based on convolutional neural networks | |
WO2017128606A1 (en) | Educational toy kit and tangram color recognition method therefor | |
CN105574161B (en) | A kind of brand logo key element recognition methods, device and system | |
CN110530875A (en) | A kind of FPCB open defect automatic detection algorithm based on deep learning | |
WO2018032626A1 (en) | Teaching toy kit and number identification method thereof | |
CN103914708A (en) | Food variety detection method and system based on machine vision | |
CN107392151A (en) | Face image various dimensions emotion judgement system and method based on neutral net | |
WO2022267300A1 (en) | Method and system for automatically extracting target area in image, and storage medium | |
CN109658391A (en) | A kind of radius of circle measurement method being fitted based on contour mergence and convex closure | |
CN111126143A (en) | Deep learning-based exercise judgment guidance method and system | |
CN108601509A (en) | Image processing apparatus, image processing method and program | |
CN110705465B (en) | Hand type classification method based on image processing | |
CN109507198A (en) | Mask detection system and method based on Fast Fourier Transform (FFT) and linear Gauss | |
CN102136077B (en) | Method for automatically recognizing lip color based on support vector machine | |
JP5780791B2 (en) | Cell tracking method | |
CN109472257B (en) | Character layout determining method and device | |
CN110263608A (en) | Electronic component automatic identifying method based on image feature space variable threshold value metric | |
CN111738984B (en) | Skin image spot evaluation method and system based on watershed and seed filling | |
JP7383722B2 (en) | Method and system for defect detection in image data of target coatings | |
CN110866917A (en) | Tablet type and arrangement mode identification method based on machine vision | |
CN115343279A (en) | Test paper-based skin oil evaluation method |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20210518 Termination date: 20210930 |