CN109934152A - A kind of small curved boom image partition method of improvement for sign language image - Google Patents
A kind of small curved boom image partition method of improvement for sign language image Download PDFInfo
- Publication number
- CN109934152A CN109934152A CN201910174083.9A CN201910174083A CN109934152A CN 109934152 A CN109934152 A CN 109934152A CN 201910174083 A CN201910174083 A CN 201910174083A CN 109934152 A CN109934152 A CN 109934152A
- Authority
- CN
- China
- Prior art keywords
- image
- sign language
- region
- area
- point
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 40
- 230000006872 improvement Effects 0.000 title claims abstract description 11
- 238000005192 partition Methods 0.000 title claims abstract description 10
- 230000011218 segmentation Effects 0.000 claims abstract description 29
- 238000001914 filtration Methods 0.000 claims abstract description 9
- 238000012545 processing Methods 0.000 claims abstract description 9
- 238000006243 chemical reaction Methods 0.000 claims abstract description 3
- 230000006870 function Effects 0.000 claims description 9
- 238000010586 diagram Methods 0.000 claims description 8
- 238000003709 image segmentation Methods 0.000 claims description 4
- 230000000877 morphologic effect Effects 0.000 claims description 4
- 230000014759 maintenance of location Effects 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 2
- 238000009499 grossing Methods 0.000 claims description 2
- 230000009466 transformation Effects 0.000 claims 1
- 230000000694 effects Effects 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 3
- 238000005286 illumination Methods 0.000 description 3
- 238000005452 bending Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000013480 data collection Methods 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 230000033228 biological regulation Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 210000000707 wrist Anatomy 0.000 description 1
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a kind of small curved boom image partition methods of improvement for sign language image.Input sign language image is read, colour of skin cluster is carried out after the conversion of YCbCr color space and divides the area of skin color obtained in image；It is smoothed using median filtering, removes the burr or white point around area of skin color, carry out max-thresholds binaryzation, using the cavity in image-region filling algorithm filling sign language region；Processing is carried out using area operator and center coordination and obtains hand-arm region, judges that the method for cutting line position carries out the positioning of hand-arm region cutting line using lateral distance, realizes the full segmentation of hand region.The present invention proposes that being judged that cutting line position is finely divided using lateral distance is cut, the cutting line between arm is positioned, the full segmentation of hand region is realized, solves the segmentation problem of pair of arm regions, suitable for the segmentation of small curved boom, which has preferable robustness.
Description
Technical field
The present invention relates to a kind of image partition methods, more particularly, to a kind of small curved boom image of improvement for sign language image
Dividing method.
Background technique
Hand Gesture Segmentation is the committed step during gesture identification, guarantees that its good segmentation effect is to realize that sign language is accurate
The essential condition of identification.How gesture accurately to be split always from complex background is a difficult point, this be mainly because
It is easy to be illuminated by the light etc. the influence of factors for gesture environment.At present common Hand Gesture Segmentation technology have segmentation based on profile information,
Segmentation based on Skin Color Information, segmentation based on motion information etc..
The method of Hand Gesture Segmentation is varied, but in practical cutting procedure, single method is difficult to be completed at the same time speed
The requirement of degree, precision and fitness, the comprehensive each algorithm advantage of usual connected applications scene is joined using multi-method in practical applications
Close segmentation.The significant challenge of current Hand Gesture Segmentation is as follows: 1) the class area of skin color in images of gestures can be to the result of Hand Gesture Segmentation
It impacts；2) existing arm minimizing technology is mainly the palm and arm is wider designs according to relatively narrow at wrist.This method
There are upper arm situations in sign language image, and when arm bending degree is larger and are not suitable for.In view of the above problems, the present invention proposes
A kind of improvement for sign language image small curved boom image partition method.
Summary of the invention
For small curved boom to the interference problem of sign language information, for area of skin color present in gesture background (hand-arm, neck
Son, arm etc.) and class area of skin color, it is an object of the invention to propose a kind of small curved boom image of the improvement for sign language image
Dividing method.The present invention proposes that being judged that cutting line position is finely divided using lateral distance is cut, and the cutting line between arm is positioned,
Realize the full segmentation of hand region.
The technical solution adopted by the present invention is that:
1) input sign language image is read, then opponent's sonagram picture carries out colour of skin cluster after carrying out the conversion of YCbCr color space
And segmentation, the class area of skin color in sign language image is removed, the area of skin color in image is obtained；
Towards figure captured by people front when manpower gesticulates gesture before waist or chest when the behaviour of sign language image is stood
Picture is RGB image, and people is located at image middle, wherein gesticulating gesture on the left of image, i.e., manpower compares the hand of gesture for the right side
Hand.
It is embodied under daily illumination condition and is shot with computer camera, to the background of shooting picture without wanting
It asks, includes the neck of sign language person or less and waist area above in the picture of shooting.
In YCbCr color space, luminance signal and carrier chrominance signal are separately separated out.When intensity of illumination changes
When, three components of RGB color can change simultaneously, and Cr component (red chrominance component) and Cb in YCbCr space
It is little that component (chroma blue component) by light intensity is influenced not strong and dependence, therefore the colour of skin is more suitable for clustering in YCbCr space, energy
Enough distributed areas for preferably extracting the colour of skin.
2) the sign language image after step 1) segmentation is smoothed using median filtering, is removed around area of skin color
Burr or white point, then carrying out max-thresholds binaryzation to the sign language image after smoothing processing, (target pixel value is set as 1, back
0) scape pixel value is set as, finally using the cavity in image-region filling algorithm filling sign language region, it is ensured that sign language region in image
Integrality；
3) processing is carried out using area operator and center coordination and obtains hand-arm region, sign language figure after removal colour of skin cluster
As present in neck area and independent arm regions, can remove the colour of skin cluster after sign language image in there are gesture areas
Area of skin color and class area of skin color in addition；
4) judge that the method for cutting line position carries out the positioning of hand-arm region cutting line using lateral distance, realize hand
The full segmentation in portion region.
The step 1) specifically:
1.1) sign language image is converted from rgb space to YCbCr space, linear particular by formula once turns
Bring realization:
Wherein, Y, Cb, Cr respectively indicate the brightness of YCbCr space, chroma blue, red color；R, G, B are respectively indicated
The red light of rgb space, green light, blue light, these three light are known as primaries again；
1.2) then retain Y value between 70-190, Cb is between 77-125, Cr value region model between 138-163
Interior pixel value (as area of skin color) is enclosed, the pixel value by Y, Cb, Cr not within the scope of value region is set as 0, realizes the colour of skin
The correct division in region and background area, has achieved the purpose that image segmentation.
Specific step is as follows for the step 3):
3.1) in sign language image there are three area of skin color (independent arm regions, neck area and hand-arm region) and its
His class area of skin color, hand-arm region include gesture area and arm regions, have multiple skins in sign language image after colour of skin cluster
Color region exists, and is handled using the method for area operator filtering, removes small surface area, and Retention area maximum three
Region is denoted as the first area maximum region maxArea, second area maximum region secArea, third area maximum region respectively
ThiArea retains three regions, other remaining region filled blacks；
3.2) arm regions, neck area are compared, hand-arm region calculates on the unilateral side of sign language image, left side or right side
Trizonal centroid position, centroid calculation first use the zeroth order square of the calculating image of the square moments function module in opencv
(m00) and first moment (m10、m01), m10And m01It respectively indicates, is then calculated using the following equation again:
Wherein,Respectively indicate the abscissa of mass center, the ordinate of mass center；
Then retain the smallest region of abscissa in center-of-mass coordinate to be finally obtained after center coordination as gesture area
Sign language image.
The correct meaning that redundancy in image can express gesture has an impact, and the feature that arm bending will cause
Otherness, the present invention are directed to this phenomenon, propose and judge that the method for cutting line position carries out hand-arm region using lateral distance
The positioning of cutting line realizes the full segmentation of hand region.
Specific step is as follows for the step 4):
4.1) the hand connected domain of the sign language image after center coordination is obtained:
Pass through the cvFindContours function module finding step 3 in opencv software tool) obtain sign language image
In edge contour, then according to edge contour in opencv software tool cvDrawContours function module draw hand
The edge contour figure of sonagram picture, obtains hand connected domain；
4.2) coordinate points on hand connected domain edge are stored in edge array, obtain ultra-left point Pz and most right in image border
The coordinate of point Py calculates the difference Py.x-Pz.x, Pz.x, Py.x point of the image abscissa between ultra-left point Pz and rightest point Py
Not Biao Shi ultra-left point Pz and rightest point Py image abscissa；Difference Py.x-Pz.x and preset lateral threshold value are compared:
As Py.x-Pz.x < 64, then the coordinate of top point Ps in image border is obtained, traversal searches ultra-left point Pz and most
Minimum point Pg between upper Ps, taking the coordinate of minimum reference point Pg1 is (Pg.x, Pg.y+30), by minimum point Pg and minimum ginseng
Line PgPg1 between examination point Pg1 is denoted as cutting line L1；
As Py.x-Pz.x >=64, then judgement is not belonging to small curved boom, without processing, stops pictures subsequent processing, into
The processing of next image of row.
For sign language image used in the present invention, when being small curved boom situation in sign language image, Py.x-Pz.x is smaller than
64, so the present invention judges whether it is small curved boom situation for whether Py.x-Pz.x is used as less than 64.
4.4) cutting line L is used1After cutting, hand region is 8 disconnections for being connected to form with the disconnection of arm regions, with most upper
Point Ps is the morphological reconstruction that seed point carries out four connection forms, from seed point Ps using the profile diagram of sign language image as template
Hand connected domain will be divided into two parts using the first cutting line segment PgPg1 as boundary in the profile diagram of sign language image, will include seed
All pixels point is disposed as white as the hand region in hand-arm region, after being rebuild in the part of point Ps
Image；
4.5) after the reconstruction for subtracting step 4.4) acquisition with the sign language image after the center coordination of step 3.3) acquisition
Image, obtain final sign language image segmentation result.
The zeroth order square of image is calculated in the step 3.2) using the square moments function module in opencv software tool
(m00) and first moment (m10、m01)。
The invention has the advantages that:
(1) present invention determines method using the arm regions based on area operator and center coordination, goes the class unless hand
The colour of skin and area of skin color (such as neck, arm), determine hand-arm region specific location.
(2) present invention judges that the method for cutting line position carries out the positioning of hand-arm region cutting line using lateral distance, should
Method can be realized the accurate segmentation of small curved boom.
Method proposed by the present invention solves the segmentation problem of pair of arm regions, suitable for the segmentation of small curved boom, the arm
Minimizing technology has preferable robustness.
Detailed description of the invention
Fig. 1 is the flow chart of the method for the present invention.
Fig. 2 is partial data collection used in the method for the present invention.
Fig. 3 is the original sign language image and colour of skin cluster result figure of the embodiment of the present invention.
Fig. 4 is the median-filtered result figure of the embodiment of the present invention.
Fig. 5 is that the area operator of the embodiment of the present invention and center coordination carry out hand-arm region definitive result figure.
Fig. 6 is the image border profile diagram of the embodiment of the present invention.
Fig. 7 is the cutting line segment positioning figure and reconstruction result map of the embodiment of the present invention.
Fig. 8 is the segmentation final result figure of the embodiment of the present invention.
Specific embodiment
Present invention will be further explained below with reference to the attached drawings and examples.
Present invention is generally directed to the hand regions of the small curved boom situation in sign language image to divide situation, due to knowledge of the invention
Other object is customized 26 class static state Alphabet Gesture, so self-built static gesture picture library, side of the present invention in specific implementation
Partial data collection used in method as shown in Fig. 2, Fig. 2 (B)~Fig. 2 (X) respectively indicates English alphabet represented by gesture picture,
It is shot, image is saved into .jpg format, last picture size is with calculating camera under daily illumination condition
1920×1080。
As shown in Figure 1, the method for the present invention, which first passes through the picture control in MFC, reads in sign language figure according to the size of regulation
Then picture utilizes area operator and matter using the class area of skin color in the colour of skin clustering method removal background of YCbCr color space
The positioning of hand-arm region is realized in heart positioning, finally judges that the method for cutting line position carries out hand-arm region and cuts using lateral distance
The positioning of secant realizes the full segmentation of hand region.
The embodiment and its implementation process of the method for the present invention are as follows:
1) present invention reads in sign language image using the picture control in MFC, and the image of reading contracts according to formula
It puts.
The present invention is using Cb and Cr in 130 sign language images progress YCbcr color space in sign language image library with brightness
The experimental verification of variation, when Y value is between 70-190, Cr between 138-163 and Cb between 77-125 Clustering Effect compared with
It is good；When Y value is lower than 70, image is very dark, is that image is brighter when Y value is greater than 190, and both of these case need when colour of skin cluster
It is 0 that original pixel value, which is arranged,.As shown in figure 3, Fig. 3 (a) is the original image read, Fig. 3 (b) is effect picture after image clustering
Image after cluster, is denoted as fuse.
2) in order to filter out the irrelevant information in sign language image, enhance area of skin color information, improve picture quality, this is adopted clearly
Image is handled with median filtering.Have experiments verify that carrying out colour of skin cluster to image and being filtered operation again later
Denoising effect well, image can remove noise by the sign language image of median filtering and image is not destroyed.To figure
As fuse carry out median filtering result figure as shown in figure 4, Fig. 4 (a) be cluster after image, Fig. 4 (b) for median filtering it
Image afterwards, is denoted as filter.
3) by sign language image data set it is found that including neck, the colours of skin such as independent arm in the sign language image that the present invention acquires
Region, therefore can have other area of skin color except gesture part in the sign language image after colour of skin cluster.In order to remove gesture area
Region other than domain, the present invention carry out the determination of hand-arm region using area operator and center coordination, removal except gesture area with
Outer area of skin color guarantees that hand-arm region is accurately separated with background, to the figure of segmentation result again such as Fig. 5 of image filter
Shown, Fig. 5 (a) is the sign language image after binaryzation, and Fig. 5 (b) be the sign language image after dividing again, then the picture after dividing is remembered
For src_image.
3.1) area in each region, maximum three regions of Retention area are calculated, maximum three areas are denoted as the respectively
One area maximum region maxArea, second area maximum region secArea, third area maximum region thiArea, by this three
A region retains, other region filled blacks；
3.2) trizonal center-of-mass coordinate is calculated, compares the abscissa size of three region center-of-mass coordinates, then retains
The smallest region of abscissa is as gesture area in center-of-mass coordinate, the sign language image being finally obtained after center coordination.
4) since the correct meaning that the redundancy of arm can express gesture have an impact, and arm length will cause
Feature difference, the present invention are directed to this phenomenon, and the present invention judges that the method for cutting line position carries out hand-arm using lateral distance
The positioning of region cutting line, this method can be realized the accurate segmentation of small curved boom.
4.1) the edge contour figure of image src_image is obtained, profile diagram is as shown in fig. 6, Fig. 6 (a) is after dividing again
Sign language image, Fig. 6 (b) are the edge contour figure of image, and the edge contour seal of image is contoursImage.Especially by
CvFindContours () function tool in opencv searches the edge contour figure of the sign language image after dividing again, then
The edge contour figure that sign language image is drawn with cvDrawContours () function tool in opencv, obtains hand connected domain.
4.2) coordinate points on image border are stored in Boundary array, obtain ultra-left point Pz and most right in image border
The coordinate of point Py calculates the difference Py.x-Pz.x of the image abscissa between ultra-left point Pz and rightest point Py, for institute of the present invention
The sign language image used, when being small curved boom situation in sign language image, Py.x-Pz.x is smaller than 64, so the present invention is by Py.x-
Whether Pz.x is used as less than 64 judges whether it is small curved boom situation.
4.3) as Py.x-Pz.x < 64, the coordinate of top point Ps in image border is obtained, traversal is searched between Pz and Ps
Minimum point, be denoted as Pg, take minimum reference point Pg1 coordinate be (Pg.x, Pg.y+30), minimum point Pg and minimum reference point Pg1
Between line PgPg1 be denoted as cutting line L1, the positioning of line segment is cut as shown in fig. 7, Fig. 7 (a) is the sign language figure after dividing again
Picture, Fig. 7 (b) are the positioning result figure of figure cutting line segment.
4.4) cutting line L is used1After cutting, hand region is 8 disconnections for being connected to form with the disconnection of arm regions.So
The morphological reconstruction of four connection forms is carried out using Ps as seed point.From seed point Ps using the profile diagram of sign language image as template
By in the profile diagram of sign language image with cutting line L1Hand connected domain is divided into two parts for boundary, will include the part of seed point Ps
Middle all pixels point is disposed as white；Shown in effect picture such as Fig. 7 (c) after morphological reconstruction.Image after note reconstruction
For chongjian.
4.5) the image src_image subtracted image chongjian after dividing again is used, final result figure is obtained
Chongjian_dst, as depicted in figure 8, Fig. 8 (a) is the sign language image after dividing again to result figure, and Fig. 8 (b) is final result figure.
It can be seen from the above, this example realizes the positioning of hand-arm region using area operator and center coordination, solves hand
Area of skin color is excessive in sonagram picture, can not carry out hand-arm region orientation problem, has established heavily fortified point for the segmentation of subsequent hand region
Real basis.This example is split small curved boom using lateral distance, solves the problems, such as arm information redundancy, to gesture area
The feature extraction in domain has great importance.This example can realize that the gesture of small curved boom situation is accurately divided, and ensure that hand area
The integrality in domain.
Above-mentioned specific embodiment is used to illustrate the present invention, rather than limits the invention, of the invention
In spirit and scope of protection of the claims, to any modifications and changes that the present invention makes, protection model of the invention is both fallen within
It encloses.
Claims (5)
1. a kind of small curved boom image partition method of improvement for sign language image, it is characterised in that the step of this method is as follows:
1) input sign language image is read, then opponent's sonagram picture carries out colour of skin cluster after carrying out the conversion of YCbCr color space and divides
It cuts, obtains the area of skin color in image；
2) the sign language image after step 1) segmentation is smoothed using median filtering, removes the burr around area of skin color
Or white point, max-thresholds binaryzation then is carried out to the sign language image after smoothing processing, is finally calculated using image-region filling
The cavity in method filling sign language region；
3) processing is carried out using area operator and center coordination and obtains hand-arm region, after removal colour of skin cluster in sign language image
Existing neck area and independent arm regions；
4) judge that the method for cutting line position carries out the positioning of hand-arm region cutting line using lateral distance, realize hand area
The full segmentation in domain.
2. the small curved boom image partition method of a kind of improvement for sign language image according to claim 1, it is characterised in that:
The step 1) specifically:
1.1) sign language image is converted from rgb space to YCbCr space, is come particular by the linear transformation of formula once
It realizes:
Wherein, Y, Cb, Cr respectively indicate the brightness of YCbCr space, chroma blue, red color；R, G, B respectively indicate RGB
The red light in space, green light, blue light；
1.2) then retain Y value between 70-190, Cb between 77-125, Cr is between 138-163 within the scope of value region
Pixel value (as area of skin color), the pixel value by Y, Cb, Cr not within the scope of value region is set as 0, realizes area of skin color
With the division of background area, image segmentation is achieved the purpose that.
3. the small curved boom image partition method of a kind of improvement for sign language image according to claim 1, it is characterised in that:
Specific step is as follows for the step 3):
3.1) it is handled using the method for area operator filtering, maximum three regions of Retention area are denoted as the first face respectively
Product maximum region maxArea, second area maximum region secArea, third area maximum region thiArea, by three regions
Retain, other remaining region filled blacks；
3.2) trizonal centroid position is calculated, centroid calculation first calculates the zeroth order square (m of image00) and first moment (m10、
m01), m10And m01It respectively indicates, is then calculated using the following equation again:
Wherein,Respectively indicate the abscissa of mass center, the ordinate of mass center；
Then the smallest region of abscissa is as gesture area in reservation center-of-mass coordinate, the hand being finally obtained after center coordination
Sonagram picture.
4. the small curved boom image partition method of a kind of improvement for sign language image according to claim 1, it is characterised in that:
Specific step is as follows for the step 4):
4.1) the hand connected domain of the sign language image after center coordination is obtained:
Pass through the cvFindContours function module finding step 3 in opencv software tool) obtain sign language image in
Then edge contour draws sign language figure according to cvDrawContours function module of the edge contour in opencv software tool
The edge contour figure of picture, obtains hand connected domain；
4.2) coordinate points on hand connected domain edge are stored in edge array, obtain ultra-left point Pz and rightest point Py in image border
Coordinate, calculate the difference Py.x-Pz.x of the image abscissa between ultra-left point Pz and rightest point Py, Pz.x, Py.x distinguish table
Show the image abscissa of ultra-left point Pz and rightest point Py；Difference Py.x-Pz.x and preset lateral threshold value are compared:
As Py.x-Pz.x < 64, then the coordinate of top point Ps in image border is obtained, traversal searches ultra-left point Pz and top point
Minimum point Pg between Ps, taking the coordinate of minimum reference point Pg1 is (Pg.x, Pg.y+30), by minimum point Pg and minimum reference point
Line PgPg1 between Pg1 is denoted as cutting line L1；
As Py.x-Pz.x >=64, then judgement is not belonging to small curved boom, without processing.
4.4) cutting line L is used1After cutting, the morphological reconstruction of four connection forms is carried out using top point Ps as seed point, from seed point
Ps, which sets out by template of the profile diagram of sign language image, using the first cutting line segment PgPg1 to be boundary by hand in the profile diagram of sign language image
Connected domain is divided into two parts, by include seed point Ps part in all pixels point be disposed as white as hand-arm region
In hand region, the image after being rebuild；
4.5) the sign language image after the center coordination obtained with step 3.3) subtracts the figure after the reconstruction of step 4.4) acquisition
Picture obtains final sign language image segmentation result.
5. the small curved boom image partition method of a kind of improvement for sign language image according to claim 1, it is characterised in that:
Zeroth order square (the m of image is calculated in the step 3.2) using the square moments function module in opencv software tool00) and one
Rank square (m10、m01)。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910174083.9A CN109934152B (en) | 2019-03-08 | 2019-03-08 | Improved small-bent-arm image segmentation method for sign language image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910174083.9A CN109934152B (en) | 2019-03-08 | 2019-03-08 | Improved small-bent-arm image segmentation method for sign language image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109934152A true CN109934152A (en) | 2019-06-25 |
| CN109934152B CN109934152B (en) | 2021-02-09 |
Family
ID=66986599
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910174083.9A Active CN109934152B (en) | 2019-03-08 | 2019-03-08 | Improved small-bent-arm image segmentation method for sign language image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109934152B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110599553A (en) * | 2019-09-10 | 2019-12-20 | 江南大学 | Skin color extraction and detection method based on YCbCr |
| CN111160194A (en) * | 2019-12-23 | 2020-05-15 | 浙江理工大学 | Static gesture image recognition method based on multi-feature fusion |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102982315A (en) * | 2012-11-05 | 2013-03-20 | 中国科学院计算技术研究所 | Gesture segmentation recognition method capable of detecting non-gesture modes automatically and gesture segmentation recognition system |
| CN104102335A (en) * | 2013-04-15 | 2014-10-15 | 中兴通讯股份有限公司 | Gesture control method, device and system |
| CN104766055A (en) * | 2015-03-26 | 2015-07-08 | 济南大学 | Method for removing wrist image in gesture recognition |
| CN104809425A (en) * | 2014-01-24 | 2015-07-29 | 上海联影医疗科技有限公司 | Method and device of extracting region of interest of hand |
| CN109190516A (en) * | 2018-08-14 | 2019-01-11 | 东北大学 | A kind of static gesture identification method based on volar edge contour vectorization |
-
2019
- 2019-03-08 CN CN201910174083.9A patent/CN109934152B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102982315A (en) * | 2012-11-05 | 2013-03-20 | 中国科学院计算技术研究所 | Gesture segmentation recognition method capable of detecting non-gesture modes automatically and gesture segmentation recognition system |
| CN104102335A (en) * | 2013-04-15 | 2014-10-15 | 中兴通讯股份有限公司 | Gesture control method, device and system |
| CN104809425A (en) * | 2014-01-24 | 2015-07-29 | 上海联影医疗科技有限公司 | Method and device of extracting region of interest of hand |
| CN104766055A (en) * | 2015-03-26 | 2015-07-08 | 济南大学 | Method for removing wrist image in gesture recognition |
| CN109190516A (en) * | 2018-08-14 | 2019-01-11 | 东北大学 | A kind of static gesture identification method based on volar edge contour vectorization |
Non-Patent Citations (3)
| Title |
|---|
| 李亚兰: ""基于视觉的实时静态手势识别技术研究"", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 汪镜秋: ""基于肤色分割的嵌入式手势识别算法研究"", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 郑方梅 等: ""基于YCRCb色彩空间的手语图像分割"", 《电脑知识与技术》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110599553A (en) * | 2019-09-10 | 2019-12-20 | 江南大学 | Skin color extraction and detection method based on YCbCr |
| CN110599553B (en) * | 2019-09-10 | 2021-11-02 | 江南大学 | Skin color extraction and detection method based on YCbCr |
| CN111160194A (en) * | 2019-12-23 | 2020-05-15 | 浙江理工大学 | Static gesture image recognition method based on multi-feature fusion |
| CN111160194B (en) * | 2019-12-23 | 2022-06-24 | 浙江理工大学 | Static gesture image recognition method based on multi-feature fusion |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109934152B (en) | 2021-02-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2017084204A1 (en) | Method and system for tracking human body skeleton point in two-dimensional video stream | |
| CN103914699A (en) | Automatic lip gloss image enhancement method based on color space | |
| CN109948461B (en) | Sign language image segmentation method based on centroid positioning and distance transformation | |
| CN106097354B (en) | A kind of hand images dividing method of combining adaptive Gauss Face Detection and region growing | |
| JPH0877334A (en) | Automatic feature point extracting method for face image | |
| CN105184216A (en) | Cardiac second region palm print digital extraction method | |
| WO2020038312A1 (en) | Multi-channel tongue body edge detection device and method, and storage medium | |
| CN103581571A (en) | Video image matting method based on three elements of color | |
| CN103186904A (en) | Method and device for extracting picture contours | |
| CN109446982A (en) | A kind of power screen cabinet pressing plate state identification method and system based on AR glasses | |
| CN107767379A (en) | Pcb board marks print quality inspection method | |
| CN109214297A (en) | A kind of static gesture identification method of combination depth information and Skin Color Information | |
| US9239962B2 (en) | Nail region detection method, program, storage medium, and nail region detection device | |
| JP2007272435A (en) | Face feature extraction device and face feature extraction method | |
| CN105046199A (en) | Finger tip point extraction method based on pixel classifier and ellipse fitting | |
| CN111160194B (en) | Static gesture image recognition method based on multi-feature fusion | |
| TW201734954A (en) | Data processing device, data processing method, and data processing program | |
| CN107154058A (en) | A kind of method for guiding user to reduce magic square | |
| CN107527343A (en) | A kind of agaricus bisporus stage division based on image procossing | |
| CN109934152A (en) | A kind of small curved boom image partition method of improvement for sign language image | |
| CN105869155B (en) | The extracting method of skin of face flushing areas | |
| CN108711160A (en) | A kind of Target Segmentation method based on HSI enhancement models | |
| CN110287894A (en) | A kind of gesture identification method and system for ultra-wide angle video | |
| CN113344836A (en) | Face image processing method and device, computer readable storage medium and terminal | |
| Ueda et al. | AR food changer using deep learning and cross-modal effects |
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 | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20211216 Address after: 230000 Anhui Hefei high tech Zone Innovation Industrial Park two phase J2 District C block 18 floor. Patentee after: HEFEI LONGTUTEM INFORMATION TECHNOLOGY Co.,Ltd. Address before: No.928, No.2 street, Jianggan Economic Development Zone, Hangzhou City, Zhejiang Province, 310018 Patentee before: ZHEJIANG SCI-TECH University |
|
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20220302 Address after: 200000 room 5g11, No. 1, Lane 129, Datian Road, Jing'an District, Shanghai Patentee after: Shanghai Huashi Investment Co.,Ltd. Address before: 230000 Anhui Hefei high tech Zone Innovation Industrial Park two phase J2 District C block 18 floor. Patentee before: HEFEI LONGTUTEM INFORMATION TECHNOLOGY Co.,Ltd. |
|
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20240106 Address after: Room 1260, South Building of Film and Television Park, No. 4915 Beisong Road, Songjiang District, Shanghai, 200000 Patentee after: SHANGHAI KANGAO MEDICAL TECHNOLOGY Co.,Ltd. Address before: 200000 room 5g11, No. 1, Lane 129, Datian Road, Jing'an District, Shanghai Patentee before: Shanghai Huashi Investment Co.,Ltd. |