CN112949542A - Wrist division line determining method based on convex hull detection - Google Patents
Wrist division line determining method based on convex hull detection Download PDFInfo
- Publication number
- CN112949542A CN112949542A CN202110284548.3A CN202110284548A CN112949542A CN 112949542 A CN112949542 A CN 112949542A CN 202110284548 A CN202110284548 A CN 202110284548A CN 112949542 A CN112949542 A CN 112949542A
- Authority
- CN
- China
- Prior art keywords
- points
- wrist
- point
- distance
- convex hull
- 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.)
- Withdrawn
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
- G06F18/23—Clustering techniques
-
- 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
- G06V10/267—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 by performing operations on regions, e.g. growing, shrinking or watersheds
-
- 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
- G06V10/56—Extraction of image or video features relating to colour
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Human Computer Interaction (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
A method for determining a wrist division line based on convex hull detection belongs to the field of image processing; the method for determining the wrist division line comprises the following steps: color space conversion and sobel edge detection; detecting convex hulls and marking convex points and concave points; calculating the distance between the convex points and the concave points; and analyzing the distance data to obtain a wrist division line. The complete palm area can be extracted by utilizing the wrist segmentation line, so that sufficient preparation is made for extracting the subsequent gesture characteristics, and the accuracy of gesture segmentation is greatly improved.
Description
Technical Field
The invention relates to the field of image processing, in particular to a method for determining a wrist division line based on convex hull detection.
Background
The wrist is a dividing line of a palm area and an arm area of a human body, in the research process of static gesture recognition, the palm area is an area required by a researcher, the arm area is a redundant area interfering with gesture feature extraction, how to completely reserve the palm area and remove the arm area directly influences accuracy of gesture recognition, and the wrist dividing line can completely separate the palm area and the arm area.
At present, the main wrist segmentation is mainly divided into two types, one type is that a depth camera is used for collecting skeleton lines of a human body, and the position of the wrist of the human body is judged according to skeleton points, so that the method is high in cost and low in accuracy; the other method is to scan the width of the human hand and determine the position of the wrist according to the fact that the width of the wrist is smaller than the width of the palm and the arm, and the method has the defects of large calculation amount and poor applicability. Therefore, the simple and efficient wrist segmentation method is provided by combining convex hull detection.
Disclosure of Invention
The invention aims to disclose a method for determining a wrist segmentation line based on convex hull detection, which improves the accuracy and the real-time property of wrist segmentation.
A method for determining a wrist segmentation line based on convex hull detection comprises the following steps:
firstly, carrying out skin color threshold segmentation on an input RGB gesture image A, and obtaining a binary image B by using sobel edge detection;
step two, carrying out convex hull detection on the binary image B, and marking all the salient points and points with the farthest distance from the image contour between the adjacent salient points to the connecting line segment of the two salient points;
step three, calculating the distance between the mark points to obtain distance data;
and step four, judging the distance data, and determining two end points of the wrist division line, namely determining the wrist division line.
In the first step, the skin color threshold segmentation is carried out on the input RGB gesture image, and the specific method comprises the following steps:
converting the gesture image A into a YCbCr color space, and converting the color space according to the formula (1):
the palm region and the connected arm region can be divided by making Y > 80, Cr < 133 and < 173, and Cb < 133, wherein Y represents brightness, and Cb and Cr represent blue component and red component.
And in the second step, performing convex hull detection on the binary image B, and marking points with the farthest distance from the image contour between all the salient points and the adjacent salient points to the connecting line segment of the two salient points, wherein the specific steps of the convex hull detection are as follows:
step two, establishing a two-dimensional rectangular coordinate system, selecting one point with the minimum y coordinate from all points on the image as a base point, and recording the point as p0If the y coordinates of a plurality of points are all the minimum value, selecting a point with the minimum x coordinate, removing one point with the same coordinate, and then according to p0The included angles between the vectors formed with other points and the x axis are sorted and are marked as pleft={p1,p2…pnSequencing the included angles from large to small according to clockwise scanning, and otherwise, carrying out anticlockwise scanning;
step two, when scanning clockwise, p0And pleftAll the points in the vector are connected to form a vector pritht={p0p1,p1p2…pn-1pnComparing cross products of adjacent vectors in sequence, if the cross product of the previous vector and the next vector is positive (anticlockwise is negative), removing the intersection point of the two vectors, otherwise, judging the intersection point of the two vectors as a convex point;
step two and step three, sequentially traversing prightRepeating the step two and the step three to finish the detection of the salient points of the whole image;
and step two, sequentially calculating the distance from the point on the outline between the adjacent salient points to the connecting line segment of the two salient points, and marking the point with the maximum distance.
And in the third step, the distance between the mark points is calculated to obtain distance data, and the specific operation is as follows:
step three, traversing all marked points, and selecting two adjacent points as a starting point xiAnd yiWherein x isiIs a left dot, yiFor the right point, the distance L between the two points is calculatedi;
Step three, step two, order xiLook for the next point x in the counterclockwise directioni+1,yiLooking for the next point y in the clockwise directioni+1Calculating the distance L between two pointsi+1;
Step three, repeating the step three one, calculating the distances of all adjacent points, and then repeating the step three two to calculate the distances of other points;
judging the distance data in the fourth step, determining two end points of the wrist division line, and judging according to a formula (2):
Li+1<Li,Li+1<Li+2 (2)
if formula (2) is satisfied, xiAnd yiThe two end points of the wrist division line are connected with each other to form the wrist division line.
The invention has the beneficial effects that: the method for determining the wrist division line based on convex hull detection is provided, so that the interference of an arm to a gesture recognition process is avoided, and the accuracy of gesture division is greatly improved.
Drawings
FIG. 1 is a flowchart illustrating an embodiment of a method for determining a wrist segmentation line based on convex hull detection according to the present invention;
FIG. 2 is a schematic diagram of the end points of the wrist segmentation lines established by the convex hull detection-based wrist segmentation line establishing method of the present invention;
Detailed Description
The invention provides a method for determining a wrist division line for convex hull detection. 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 method of the present invention is described in further detail below with reference to the accompanying drawings:
detailed description of the invention
The invention provides a method for determining a wrist division line for convex hull detection, which comprises the following specific steps as shown in figure 1:
firstly, carrying out skin color threshold segmentation on an input RGB gesture image A, and obtaining a binary image B by using sobel edge detection;
step two, carrying out convex hull detection on the binary image B, and marking all the salient points and points with the farthest distance from the image contour between the adjacent salient points to the connecting line segment of the two salient points;
step three, calculating the distance between the mark points to obtain distance data;
and step four, judging the distance data, and determining two end points of the wrist division line, namely determining the wrist division line.
Detailed description of the invention
On the basis of the first specific embodiment, a method for determining a wrist segmentation line for convex hull detection is provided, wherein in the second step, a skin color threshold segmentation is performed on an input RGB gesture image a, and interference regions such as clothes and the like and hand regions can be segmented by reasonably setting a threshold by using good clustering characteristics of human skin colors in a YCbCr color space, and the specific method comprises the following steps:
converting the gesture image A into a YCbCr color space, and converting the color space according to the formula (1):
dividing the palm region and the connected arm region by enabling Y to be more than 80, enabling Cr to be more than 133 and less than 173, and enabling Cb to be more than 127 and less than 133, wherein Y represents brightness, and Cb and Cr represent blue components and red components; and obtaining a binary image B by using sobel edge detection after obtaining the skin color area.
Detailed description of the invention
On the basis of the first specific embodiment, a method for determining a wrist segmentation line for convex hull detection is described in the second step, where convex hull detection is performed on a binary image B, and points where the distance from the image contour between all the salient points and the adjacent salient points to the connecting line segment of the two salient points is the farthest, that is, a concave part and a wrist concave part between fingers are marked, where the salient points are marked as white circles and the concave points are marked as white rectangles, as shown in fig. 2, where the specific steps of convex hull detection are:
step two, establishing a two-dimensional rectangular coordinate system, selecting one point with the minimum y coordinate from all points on the image as a base point, and recording the point as p0If the y coordinates of a plurality of points are all the minimum value, selecting a point with the minimum x coordinate, removing one point with the same coordinate, and then according to p0With its base point constitutingIs ordered by the angle of the vector of (d) to the x-axis, denoted as pleft={p1,p2…pnSequencing the included angles from large to small according to clockwise scanning, and otherwise, carrying out anticlockwise scanning;
step two, when scanning clockwise, p0And pleftAll the points in the vector are connected to form a vector pritht={p0p1,p1p2…pn-1pnComparing cross products of adjacent vectors in sequence, if the cross product of the previous vector and the next vector is positive (anticlockwise is negative), removing the intersection point of the two vectors, otherwise, judging the intersection point of the two vectors as a convex point;
step two and step three, sequentially traversing prightRepeating the step two and the step three to finish the detection of the salient points of the whole image;
and step two, sequentially calculating the distance from the point on the outline between the adjacent salient points to the connecting line segment of the two salient points, and marking the point with the maximum distance.
Detailed description of the invention
On the basis of the first specific implementation mode, a method for determining a wrist segmentation line for convex hull detection includes the third step of calculating the distance between the marking points to obtain distance data, and the specific operations are as follows:
step three, traversing all marked points, and selecting two adjacent points as a starting point xiAnd yiWherein x isiIs a left dot, yiFor the right point, the distance L between the two points is calculatedi;
Step three, step two, order xiLook for the next point x in the counterclockwise directioni+1,yiLooking for the next point y in the clockwise directioni+1Calculating the distance L between two pointsi+1;
And step three, repeating the step three one, calculating the distances of all adjacent points, and then repeating the step three two to calculate the distances of other points.
Detailed description of the invention
In the first embodiment, the distance data is determined in the fourth step, two end points of the wrist division line are established, and the determination is performed according to the formula (2):
Li+1<Li,Li+1<Li+2 (2)
if formula (2) is satisfied, xiAnd yiTwo end points of the wrist division line, x as shown in FIG. 2iAnd yiThe connecting line is the wrist dividing line.
Claims (5)
1. A method for determining a wrist division line based on convex hull detection is characterized by comprising the following steps:
firstly, carrying out skin color threshold segmentation on an input RGB gesture image A, and obtaining a binary image B by using sobel edge detection;
step two, carrying out convex hull detection on the binary image B, and marking all the salient points and points with the farthest distance from the image contour between the adjacent salient points to the connecting line segment of the two salient points;
step three, calculating the distance between the mark points to obtain distance data;
and step four, judging the distance data, and determining two end points of the wrist division line, namely determining the wrist division line.
2. The method for determining the wrist segmentation line based on convex hull detection as claimed in claim one, wherein in the step one, the input RGB gesture image is subjected to threshold segmentation of skin color, and a binary image B is obtained by using sobel edge detection, and the method specifically comprises the following steps:
converting the gesture image A into a YCbCr color space, and converting the color space according to the formula (1):
the palm region and the connected arm region can be divided by making Y > 80, Cr < 133 and < 173, and Cb < 133, wherein Y represents brightness, and Cb and Cr represent blue component and red component.
3. The method for determining a wrist segmentation line based on convex hull detection as claimed in claim one, wherein the convex hull detection is performed on the binary image B in the second step, and the points with the farthest distance from the image contour between all the salient points and the adjacent salient points to the line segment of the two salient points are marked, wherein the convex hull detection specifically comprises the steps of:
step two, establishing a two-dimensional rectangular coordinate system, selecting one point with the minimum y coordinate from all points on the image as a base point, and recording the point as p0If the y coordinates of a plurality of points are all the minimum value, selecting a point with the minimum x coordinate, removing one point with the same coordinate, and then according to p0The included angles between the vectors formed with other points and the x axis are sorted and are marked as pleft={p1,p2…pnSequencing the included angles from large to small according to clockwise scanning, and otherwise, carrying out anticlockwise scanning;
step two, when scanning clockwise, p0And pleftAll the points in the vector are connected to form a vector pritht={p0p1,p1p2…pn- 1pnComparing cross products of adjacent vectors in sequence, if the cross product of the previous vector and the next vector is positive (anticlockwise is negative), removing the intersection point of the two vectors, otherwise, judging the intersection point of the two vectors as a convex point;
step two and step three, sequentially traversing prightRepeating the step two and the step three to finish the detection of the salient points of the whole image;
and step two, sequentially calculating the distance from the point on the outline between the adjacent salient points to the connecting line segment of the two salient points, and marking the point with the maximum distance.
4. The method for determining the wrist segmentation line based on convex hull detection as claimed in claim one, wherein the step three is to calculate the distance between the marking points to obtain the distance data, and the specific operations are as follows:
step three, traversing all marked points, and selecting two adjacent points as a starting point xiAnd yiWherein x isiIs a left dot, yiFor the right point, the distance L between the two points is calculatedi;
Step three, step two, order xiLook for the next point x in the counterclockwise directioni+1,yiLooking for the next point y in the clockwise directioni+1Calculating the distance L between two pointsi+1;
And step three, repeating the step three one, calculating the distances of all adjacent points, and then repeating the step three two to calculate the distances of other points.
5. The method for determining wrist segmentation lines based on convex hull detection as claimed in claim one, wherein the step four is to determine the distance data, establish two end points of the wrist segmentation lines, and determine according to formula (2):
Li+1<Liand L isi+1<Li+2 (2)
If formula (2) is satisfied, xiAnd yiThe two end points of the wrist division line are connected with each other to form the wrist division line.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110284548.3A CN112949542A (en) | 2021-03-17 | 2021-03-17 | Wrist division line determining method based on convex hull detection |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110284548.3A CN112949542A (en) | 2021-03-17 | 2021-03-17 | Wrist division line determining method based on convex hull detection |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112949542A true CN112949542A (en) | 2021-06-11 |
Family
ID=76228721
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110284548.3A Withdrawn CN112949542A (en) | 2021-03-17 | 2021-03-17 | Wrist division line determining method based on convex hull detection |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112949542A (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104899591A (en) * | 2015-06-17 | 2015-09-09 | 吉林纪元时空动漫游戏科技股份有限公司 | Wrist point and arm point extraction method based on depth camera |
CN107767375A (en) * | 2017-11-02 | 2018-03-06 | 广东电网有限责任公司电力科学研究院 | One kind point cloud method for evaluating quality and device |
CN109919039A (en) * | 2019-02-14 | 2019-06-21 | 上海磐启微电子有限公司 | A kind of static gesture identification method based on palm and finger characteristic |
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 |
CN112270292A (en) * | 2020-11-11 | 2021-01-26 | 上海磐启微电子有限公司 | Static gesture recognition system and method |
-
2021
- 2021-03-17 CN CN202110284548.3A patent/CN112949542A/en not_active Withdrawn
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104899591A (en) * | 2015-06-17 | 2015-09-09 | 吉林纪元时空动漫游戏科技股份有限公司 | Wrist point and arm point extraction method based on depth camera |
CN107767375A (en) * | 2017-11-02 | 2018-03-06 | 广东电网有限责任公司电力科学研究院 | One kind point cloud method for evaluating quality and device |
CN109919039A (en) * | 2019-02-14 | 2019-06-21 | 上海磐启微电子有限公司 | A kind of static gesture identification method based on palm and finger characteristic |
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 |
CN112270292A (en) * | 2020-11-11 | 2021-01-26 | 上海磐启微电子有限公司 | Static gesture recognition system and method |
Non-Patent Citations (3)
Title |
---|
周凯: "基于肤色和 SVM 的手势识别及其应用研究", 《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》 * |
张迪: "基于机器学习的手势识别技术研究", 《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》 * |
董旭德 等: "基于质心分水岭算法的静态手势分割算法模型", 《智能系统学报》 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104751142B (en) | A kind of natural scene Method for text detection based on stroke feature | |
CN108121991B (en) | Deep learning ship target detection method based on edge candidate region extraction | |
CN103971102A (en) | Static gesture recognition method based on finger contour and decision-making trees | |
CN109919039B (en) | Static gesture recognition method based on palm and finger characteristics | |
EP3678046B1 (en) | Hand detection method and system, image detection method and system, hand segmentation method, storage medium, and device | |
CN107330354B (en) | Natural gesture recognition method | |
CN108154151B (en) | Rapid multi-direction text line detection method | |
JPH0877334A (en) | Automatic feature point extracting method for face image | |
CN105260720B (en) | fingerprint identification method and device | |
CN108399644A (en) | A kind of wall images recognition methods and its device | |
CN114863492B (en) | Method and device for repairing low-quality fingerprint image | |
CN109961016B (en) | Multi-gesture accurate segmentation method for smart home scene | |
CN112906550A (en) | Static gesture recognition method based on watershed transformation | |
CN102831428B (en) | Method for extracting quick response matrix code region in image | |
CN107194402B (en) | Parallel refined skeleton extraction method | |
CN109241975B (en) | License plate character segmentation method based on character center point positioning | |
CN108182700B (en) | Image registration method based on two-time feature detection | |
CN113989308A (en) | Polygonal target segmentation method based on Hough transform and template matching | |
CN113723314A (en) | Sugarcane stem node identification method based on YOLOv3 algorithm | |
Lin et al. | License plate location method based on edge detection and mathematical morphology | |
CN109615626B (en) | Quality evaluation method for chip surface printing symbol structure defects | |
CN109977892B (en) | Ship detection method based on local saliency features and CNN-SVM | |
CN112949542A (en) | Wrist division line determining method based on convex hull detection | |
CN109766860B (en) | Face detection method based on improved Adaboost algorithm | |
CN115588208A (en) | Full-line table structure identification method based on digital image processing technology |
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 | ||
WW01 | Invention patent application withdrawn after publication |
Application publication date: 20210611 |
|
WW01 | Invention patent application withdrawn after publication |