CN108345867A - Gesture identification method towards Intelligent household scene - Google Patents
Gesture identification method towards Intelligent household scene Download PDFInfo
- Publication number
- CN108345867A CN108345867A CN201810193484.4A CN201810193484A CN108345867A CN 108345867 A CN108345867 A CN 108345867A CN 201810193484 A CN201810193484 A CN 201810193484A CN 108345867 A CN108345867 A CN 108345867A
- Authority
- CN
- China
- Prior art keywords
- image
- lbp
- pixel
- gesture
- gestures
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
- G06V40/28—Recognition of hand or arm movements, e.g. recognition of deaf sign language
-
- 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
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/50—Extraction of image or video features by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Human Computer Interaction (AREA)
- Image Analysis (AREA)
Abstract
Gesture identification method towards Intelligent household scene, includes the following steps:S1, initial pictures are pre-processed;S2, pretreated image is converted to hsv color space from RGB color, each pixel in the image after conversion is compared with colour of skin threshold value, obtains mask figure;Using rectangle fitting mask figure, rectangle part is intercepted, obtains images of gestures;S3, the three chrominance channel image of red, green, blue for detaching images of gestures, images of gestures is divided into 9 units, calculate the LBP values of each pixel in each unit, the LBP values of each unit are normalized, count the LBP distribution histograms of each unit, it is connected into an image LBP distribution histogram again, obtains the LBP features of image;S4, LBP features are trained by SVM, obtain gesture model;The present invention converts image to hsv color space and carries out images of gestures segmentation, and rectangle frame is recycled to be fitted mask figure, extracts the LBP features of gesture, and training gesture model fast and accurately can position and identify gesture in image.
Description
Technical field
The invention belongs to intelligent identification technology fields, and in particular to a kind of gesture identification side towards Intelligent household scene
Method.
Background technology
Gesture identification refers to obtaining the images of gestures of personage in the picture, feature is extracted from images of gestures, by knowing
The feature indescribably taken obtains the meaning of gesture.It can reach control house electric by identifying user gesture under household scene
The effect of equipment.User can control the operation of the equipment such as TV, air-conditioning, light using simple gesture.
According to the applicant understood, now in the market, the smart home device with gesture identification function is more rare, most hand
Gesture identification product can only identify limb action, gesture can not be identified, and recognition effect is unsatisfactory, meanwhile, these productions
Product excessively rely on networking cloud server, can not work in the case of no network, can not to a variety of electrical equipments into
Row effectively control.
SVM in the present invention refers to support vector machines, is a kind of common method of discrimination.In machine learning field, lead to
It is commonly used to carry out pattern-recognition, classification and regression analysis.SVM methods are that sample space is reflected by a Nonlinear Mapping p
It is mapped in a higher-dimension or even infinite dimensional feature space so that the problem of Nonlinear separability converts in original sample space
For linear separability in feature space the problem of.Briefly, it exactly rises peacekeeping linearisation and rises dimension, be exactly sample to height
Dimension space, which is done, to be mapped.But as classification, return the problems such as, it is likely that low-dimensional sample space can not linear process sample
This collection can but realize linear partition (or recurrence) in high-dimensional feature space by a linear hyperplane.
Common kernel function has following 4 kinds:
Linear kernel function K (x, y)=xy;
Polynomial kernel function K (x, y)=[(xy)+1] ^d;
Radial basis function K (x, y)=exp (- | x-y | ^2/d^2);
Two layers of neural network kernel function K (x, y)=tanh (a (xy)+b);
Invention content
It is an object of the invention to:A kind of gesture identification method towards Intelligent household scene is provided, by turning image
It turns to hsv color space and carries out images of gestures segmentation, recycle rectangle frame to be fitted mask figure, extract the LBP features of gesture, training
Gesture model fast and accurately can position and identify gesture in image.
In order to reach object above, a kind of gesture identification method towards Intelligent household scene is provided, is included the following steps:
S1, color space standards are carried out to initial pictures using gammma correction methods, then is obtained by gaussian filtering process
Take pretreated image;
S2, pretreated image is converted to hsv color space from RGB color, it is every in the image after conversion
A pixel is compared with colour of skin threshold value, obtains mask figure;Using rectangle fitting mask figure, the center for obtaining rectangle is sat
Mark, length and width intercept rectangle part, obtain images of gestures, and it is 64 × 64 to be sized;
S3, red channel image, green channel images, the blue channel image for detaching images of gestures, by the hand after separation
Gesture image is divided into 9 units, calculates the LBP values of each pixel in each unit, the LBP values of each unit are normalized, unite
The LBP distribution histograms series connection of 9 units is distributed histogram by the LBP distribution histograms for counting each unit for an image LBP
Figure, as the LBP features of image;
S4, LBP features are trained by SVM, obtain gesture model.
The present invention preferred embodiment be:In step S2, each pixel in the image after conversion is compared with colour of skin threshold value
More specific method is:Colour of skin HSV threshold intervals are set as [95:135,30:255,45:255], judge whether is pixel in image
In colour of skin HSV threshold intervals, if so, the pixel value is set to 255, it is no, then it is set to 0, to obtain mask figure.
Preferably, in step S2, rectangle fitting mask figure be included using a rectangle it is all continuously distributed in mask figure
Pixel value be 255 point.
Preferably, in step S3, the computational methods of the LBP values of each pixel are:The pixel window for establishing 3 × 3, with window
The gray value of mouth central pixel point is threshold value, is compared with the gray value of 8 adjacent pixels, if threshold value is small, the picture
The position of vegetarian refreshments is otherwise 0 by mark 1, in this way, 8 in the neighborhood of pixel points pixel can generate 8 bits, is taken
LBP value of the minimum binary number as the pixel.
Preferably, in step S3, normalization is standardized using (0,1), most by the LBP values of pixel in Traversal Unit lattice
Big value and minimum value, and maximum value-minimum value is normalized as radix:
Wherein, Min indicates that minimum value, Min 0, Max indicate maximum value, Max 1.
Preferably, step S4 is specially:
S41, prepare training sample set, including positive sample collection and negative sample collection;
S42, sample is cut manually, sample-size is made to be consistent;
S43, identical images of gestures in all positive samples after cutting is placed in a file, by the institute after cutting
There is identical images of gestures in negative sample to be placed in another file, all training samples are zoomed into identical size later;
The LBP features of S44, all samples of extraction, and assign sample label;
S45, it will be trained in LBP features and sample label the input SVM of all samples;Obtain gesture identification model.
The present invention has the beneficial effect that:This method carries out hand by the way that image is transformed into hsv color space from RGB color
Gesture is divided, then is fitted mask figure with rectangle frame, reduces influence of the ambient enviroment to gesture identification;It is special by the LBP for extracting gesture
Sign can effectively state the meaning of images of gestures with feature;The LBP features of extraction are trained using SVM, obtain gesture mould
Type is used for gesture identification.Can fast and accurately the gesture in image be positioned and be identified, it is comfortable to improve user experience
Degree.
Description of the drawings
The present invention will be further described below with reference to the drawings.
Fig. 1 is the method flow schematic diagram of the present invention;
Fig. 2 is the present invention to gesture image segmentation schematic process flow diagram;
Fig. 3 is the schematic process flow diagram of the LBP features of the extraction images of gestures of the present invention.
Specific implementation mode
Embodiment one
Referring to Fig. 1, the present embodiment provides a kind of gesture identification methods towards Intelligent household scene, wherein step S1-
The step of S2 is to gesture image segmentation, as shown in Figure 2;The step of step S3 is the LBP features for taking images of gestures, such as Fig. 3 institutes
Show;It is as follows:
S1, color space standardsization acquisition image Image1 is carried out to initial pictures Image0 using gammma correction methods,
Pretreated image Image2 is obtained by gaussian filtering process again;
Wherein, the schools Gamma are exactly based on to the gamma curve of image into edlin, to carry out non-linear tone to image
The method of editor detects dark parts and light-colored part in picture signal, and the two ratio is made to increase, to improve image pair
Than degree effect, the shade for reducing image local and the influence caused by illumination variation, while the interference of noise can be inhibited;Gauss
The main influence for eliminating noise to identification of filtering.
S2, the RGB color of pretreated image Image2 is converted to hsv color space, the image after conversion
Each pixel in Image3 is compared with colour of skin threshold value, obtains mask figure Y_Image0;Using rectangle fitting mask figure, obtain
Center position coordinates (x_cen, y_cen), the length and width (h, w) of rectangle, the rectangle part of interception image Image2 is taken to obtain
Take images of gestures Image4, Image4=Image2 [x_cen-h/2:x_cen+h/2,y_cen-w/2:y_cen+w/2];And it adjusts
Whole size is 64 × 64, obtains images of gestures Image5;
Wherein, HSV (Hue, Saturation, Value) is the intuitive nature of color, the parameter point of color in this model
It is not:Tone (H), saturation degree (S), lightness (V)
S3, red channel image channel_r, green channel images channel_g, the blue channel for detaching images of gestures
Image channel_b, image are made of red, green, blue three elements, and R is red channel, is expressed as 1;G is that green is logical
Road is expressed as 2;B is blue channel, is expressed as 3;White image is then 4, it is formed by 1,2,3 channel color mixing.
Images of gestures after separation is divided into 9 units, wherein the image information per unit is:
Image4[16*x:16* (x+1), 16*y:16*(y+1)],x,y∈[0,1,2]
N=(x+1) * (y+1), n ∈ [1,2,3,4,5,6,7,8]
X, y indicate that the coordinate serial number of each unit, n indicate the serial number of unit.
The LBP values for calculating each pixel in each unit, are normalized the LBP values of each unit, count each unit
LBP distribution histograms, be an image LBP distribution histogram, as image by the LBP distribution histograms series connection of 9 units
LBP features;
S4, LBP features are trained by SVM, obtain gesture model.
In step S2, the specific method that each pixel in the image after conversion is compared with colour of skin threshold value is:Setting
Colour of skin HSV threshold intervals are [95:135,30:255,45:255], judge the pixel in image whether in colour of skin HSV threshold intervals
It is interior, if so, the pixel value is set to 255, it is no, then it is set to 0, to obtain mask figure.
In step S2, rectangle fitting mask figure is to include all continuously distributed pixel values in mask figure using a rectangle
For 255 point.
In step S3, LBP values are to have rotational invariance and gray scale not for describing the operator of image local textural characteristics
Denaturation, the computational methods of the LBP values of each pixel are:The pixel window for establishing 3 × 3, with the gray scale of window center pixel
Value is threshold value, is compared with the gray value of 8 adjacent pixels, if threshold value is small, the position of the pixel by mark 1,
Otherwise it is 0, in this way, 8 pixels in the neighborhood of pixel points can generate 8 bits, takes minimum binary number conduct
The LBP values of the pixel.
In step S3, normalization is standardized using (0,1), by the maximum value of the LBP values of pixel in Traversal Unit lattice and
Minimum value, and maximum value-minimum value is normalized as radix:
Wherein, Min indicates that minimum value, Min 0, Max indicate maximum value, Max 1.
Step S4 is specially:
S41, prepare training sample set, including positive sample collection and negative sample collection;
S42, sample is cut manually, sample-size is made to be consistent;
S43, identical images of gestures in all positive samples after cutting is placed in a file, by the institute after cutting
There is identical images of gestures in negative sample to be placed in another file, all training samples are zoomed into identical size later;
The LBP features of S44, all samples of extraction, and assign sample label;
S45, it will be trained in LBP features and sample label the input SVM of all samples;Obtain gesture identification model.
In addition to the implementation, the present invention can also have other embodiment.It is all to use equivalent substitution or equivalent transformation shape
At technical solution, fall within the scope of protection required by the present invention.
Claims (6)
1. a kind of gesture identification method towards Intelligent household scene, which is characterized in that include the following steps:
S1, color space standards are carried out to initial pictures using gammma correction methods, then is obtained in advance by gaussian filtering process
Treated image;
S2, pretreated image is converted to hsv color space from RGB color, each picture in the image after conversion
Element is compared with colour of skin threshold value, obtains mask figure;Using rectangle fitting mask figure, center position coordinates, the length of rectangle are obtained
Degree and width intercept rectangle part, obtain images of gestures, and it is 64 × 64 to be sized;
S3, red channel image, green channel images, the blue channel image for detaching images of gestures, by the gesture figure after separation
As being divided into 9 units, the LBP values of each pixel in each unit are calculated, the LBP values of each unit are normalized, statistics is every
The LBP distribution histograms of a unit, then it is an image LBP distribution histogram that the LBP distribution histograms of 9 units, which are connected,
The as LBP features of image;
S4, LBP features are trained by SVM, obtain gesture model.
2. a kind of gesture identification method towards Intelligent household scene according to claim 1, which is characterized in that the step
In rapid S2, the specific method that each pixel in the image after conversion is compared with colour of skin threshold value is:Set colour of skin HSV threshold values
Section is [95:135,30:255,45:255], the pixel in image is judged whether in colour of skin HSV threshold intervals, if so,
The pixel value is set to 255, no, then 0 is set to, to obtain mask figure.
3. a kind of gesture identification method towards Intelligent household scene according to claim 1, which is characterized in that the step
In rapid S2, rectangle fitting mask figure is that a rectangle is used to include all continuously distributed pixel values in mask figure as 255 point.
4. a kind of gesture identification method towards Intelligent household scene according to claim 1, which is characterized in that the step
In rapid S3, the computational methods of the LBP values of each pixel are:The pixel window for establishing 3 × 3, with the ash of window center pixel
Angle value is threshold value, is compared with the gray value of 8 adjacent pixels, if threshold value is small, the position of the pixel is by mark
1, it is otherwise 0, in this way, 8 in the neighborhood of pixel points pixel can generate 8 bits, minimum binary number is taken to make
For the LBP values of the pixel.
5. a kind of gesture identification method towards Intelligent household scene according to claim 1, which is characterized in that the step
In rapid S3, normalization is standardized using (0,1), by the maximum value and minimum value of the LBP values of pixel in Traversal Unit lattice, and will
Maximum value-minimum value is normalized as radix:
Wherein, Min indicates that minimum value, Min 0, Max indicate maximum value, Max 1.
6. a kind of gesture identification method towards Intelligent household scene according to claim 1, which is characterized in that the step
Suddenly S4 is specially:
S41, prepare training sample set, including positive sample collection and negative sample collection;
S42, sample is cut manually, sample-size is made to be consistent;
S43, identical images of gestures in all positive samples after cutting is placed in a file, it will be all negative after cutting
Identical images of gestures is placed in another file in sample, and all training samples are zoomed to identical size later;
The LBP features of S44, all samples of extraction, and assign sample label;
S45, it will be trained in LBP features and sample label the input SVM of all samples;Obtain gesture identification model.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810193484.4A CN108345867A (en) | 2018-03-09 | 2018-03-09 | Gesture identification method towards Intelligent household scene |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810193484.4A CN108345867A (en) | 2018-03-09 | 2018-03-09 | Gesture identification method towards Intelligent household scene |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108345867A true CN108345867A (en) | 2018-07-31 |
Family
ID=62957526
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810193484.4A Pending CN108345867A (en) | 2018-03-09 | 2018-03-09 | Gesture identification method towards Intelligent household scene |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108345867A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109961016A (en) * | 2019-02-26 | 2019-07-02 | 南京邮电大学 | The accurate dividing method of more gestures towards Intelligent household scene |
CN111368767A (en) * | 2020-03-09 | 2020-07-03 | 广东三维家信息科技有限公司 | Method and device for identifying home material color tone and electronic equipment |
CN113112502A (en) * | 2021-05-11 | 2021-07-13 | 上海非夕机器人科技有限公司 | Cable detection method, robot and device with storage function |
CN114972367A (en) * | 2021-02-25 | 2022-08-30 | 上海复旦微电子集团股份有限公司 | Method, device, equipment and computer readable storage medium for segmenting image |
CN117392759A (en) * | 2023-12-11 | 2024-01-12 | 成都航空职业技术学院 | Action recognition method based on AR teaching aid |
CN114972367B (en) * | 2021-02-25 | 2024-06-07 | 上海复旦微电子集团股份有限公司 | Method, apparatus, device and computer readable storage medium for segmenting images |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104680127A (en) * | 2014-12-18 | 2015-06-03 | 闻泰通讯股份有限公司 | Gesture identification method and gesture identification system |
CN105320248A (en) * | 2014-06-03 | 2016-02-10 | 深圳Tcl新技术有限公司 | Mid-air gesture input method and device |
CN107609454A (en) * | 2016-07-11 | 2018-01-19 | 北京君正集成电路股份有限公司 | A kind of method and device of gesture in identification image |
-
2018
- 2018-03-09 CN CN201810193484.4A patent/CN108345867A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105320248A (en) * | 2014-06-03 | 2016-02-10 | 深圳Tcl新技术有限公司 | Mid-air gesture input method and device |
CN104680127A (en) * | 2014-12-18 | 2015-06-03 | 闻泰通讯股份有限公司 | Gesture identification method and gesture identification system |
CN107609454A (en) * | 2016-07-11 | 2018-01-19 | 北京君正集成电路股份有限公司 | A kind of method and device of gesture in identification image |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109961016A (en) * | 2019-02-26 | 2019-07-02 | 南京邮电大学 | The accurate dividing method of more gestures towards Intelligent household scene |
CN109961016B (en) * | 2019-02-26 | 2022-10-14 | 南京邮电大学 | Multi-gesture accurate segmentation method for smart home scene |
CN111368767A (en) * | 2020-03-09 | 2020-07-03 | 广东三维家信息科技有限公司 | Method and device for identifying home material color tone and electronic equipment |
CN111368767B (en) * | 2020-03-09 | 2024-02-02 | 广东三维家信息科技有限公司 | Household material tone identification method and device and electronic equipment |
CN114972367A (en) * | 2021-02-25 | 2022-08-30 | 上海复旦微电子集团股份有限公司 | Method, device, equipment and computer readable storage medium for segmenting image |
CN114972367B (en) * | 2021-02-25 | 2024-06-07 | 上海复旦微电子集团股份有限公司 | Method, apparatus, device and computer readable storage medium for segmenting images |
CN113112502A (en) * | 2021-05-11 | 2021-07-13 | 上海非夕机器人科技有限公司 | Cable detection method, robot and device with storage function |
CN113112502B (en) * | 2021-05-11 | 2023-10-20 | 上海非夕机器人科技有限公司 | Cable detection method, robot and device with storage function |
CN117392759A (en) * | 2023-12-11 | 2024-01-12 | 成都航空职业技术学院 | Action recognition method based on AR teaching aid |
CN117392759B (en) * | 2023-12-11 | 2024-03-12 | 成都航空职业技术学院 | Action recognition method based on AR teaching aid |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Liu et al. | A robust skin color based face detection algorithm | |
CN108345867A (en) | Gesture identification method towards Intelligent household scene | |
WO2019169772A1 (en) | Picture processing method, electronic apparatus, and storage medium | |
Ajmal et al. | A comparison of RGB and HSV colour spaces for visual attention models | |
CN106909884B (en) | Hand region detection method and device based on layered structure and deformable part model | |
Mythili et al. | Color image segmentation using ERKFCM | |
CN110287787B (en) | Image recognition method, image recognition device and computer-readable storage medium | |
Baek et al. | Vehicle color classification based on the support vector machine method | |
JP6779688B2 (en) | Image processing equipment, image processing method, computer program | |
CN108038458B (en) | Method for automatically acquiring outdoor scene text in video based on characteristic abstract diagram | |
CN104866826B (en) | A kind of static sign Language Recognition Method based on KNN and pixel ratio Gradient Features | |
CN104392211A (en) | Skin recognition method based on saliency detection | |
CN108830908A (en) | A kind of magic square color identification method based on artificial neural network | |
Fernando et al. | Low cost approach for real time sign language recognition | |
CN208689542U (en) | Gesture recognition control system towards Intelligent household scene | |
Chen et al. | An integrated color and hand gesture recognition approach for an autonomous mobile robot | |
CN112101260B (en) | Method, device, equipment and storage medium for identifying safety belt of operator | |
Fitriyah et al. | Traffic sign recognition using edge detection and eigen-face: Comparison between with and without color pre-classification based on Hue | |
McFeely et al. | Shadow identification for digital imagery using colour and texture cues | |
CN108549855A (en) | Real-time humanoid detection method towards Intelligent household scene | |
CN115601690A (en) | Edible fungus environment detection method based on intelligent agriculture | |
CN111985482B (en) | Multi-mode multi-scene automatic license plate recognition method based on deep learning | |
CN110348530B (en) | Method for identifying lipstick number | |
CN117152787A (en) | Character clothing recognition method, device, equipment and readable storage medium | |
WO2021184718A1 (en) | Card border recognition method, apparatus and device, and computer storage medium |
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: 20180731 |