CN113361410A - Child detection box filtering algorithm based on grid clustering - Google Patents
Child detection box filtering algorithm based on grid clustering Download PDFInfo
- Publication number
- CN113361410A CN113361410A CN202110633366.2A CN202110633366A CN113361410A CN 113361410 A CN113361410 A CN 113361410A CN 202110633366 A CN202110633366 A CN 202110633366A CN 113361410 A CN113361410 A CN 113361410A
- Authority
- CN
- China
- Prior art keywords
- frame
- grid
- child
- detection
- clustering
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of computer vision, in particular to a child detection frame filtering algorithm based on grid clustering, which comprises the following steps: acquiring video data in a real scene through a camera, sending the video data to a computer through the camera, acquiring and storing a target image, selecting the target camera, and pulling detection frame data corresponding to one week from a production environment; the child detection frame filtering algorithm provided by the invention can directly take the detected human-shaped frame as statistical data, then uses a grid clustering method to obtain the average frame height in a grid range, judges whether the child is detected or not by utilizing the average frame height and the target frame height, and can judge the average frame height of each area in a target image at the same time, thereby avoiding misjudgment caused by shielding, and solving the problems that more shielding and misdetection exist in a real scene, the statistical clustering is directly carried out by using a detection frame, and the abnormal detection frame can disturb the whole data distribution, so that the misjudgment of the child frame is more.
Description
Technical Field
The invention relates to the technical field of computer vision, in particular to a child detection frame filtering algorithm based on grid clustering.
Background
Object detection is one of the most fundamental problems in the field of computer vision and has been extensively discussed and studied. In recent years, the target detection method based on the deep convolutional neural network greatly improves the detection accuracy, but still has more challenges in practical application. For example, in some specific service scenarios (for example, a mall customer flow system analyzes user purchasing power behavior based on target detection), the detection system is required to be able to distinguish "adult" from "child", that is, each human-shaped frame needs to be given a label to determine whether the person in the frame is an adult, so as to conveniently provide filtering conditions for subsequent user behavior statistics. Therefore, it is important and valuable to distinguish between large and small persons in a detection system.
At present, a large number of pre-labeled data sets can be prepared in advance to participate in the training of the neural network, and the probability of whether an output detection frame is a child or not is judged by utilizing the multi-task network. But it is difficult to obtain a sample of a child in a real scene. Particularly, in places with dense personnel, such as markets, restaurants and the like, adults are mainly used, and children have fewer samples. Therefore, in order to avoid the situation of insufficient data quantity, there are some methods for distinguishing adults from children by statistically detecting box features. For example, by counting the aspect ratio of the detection box, setting a threshold value, and regarding the detection box below the threshold value as a child; or the characteristics of the detection boxes are used as input, the detection boxes are clustered by using a k-means clustering algorithm and the like, and then clustering results are analyzed so as to distinguish children from adults.
There are more shelters from and the condition of false retrieval in the reality scene, and the same person stands the formation of image difference in the image when different positions in addition, detects the frame size and also differs thereupon, leads to a lot of humanoid frames not to accord with the normal reason, if directly carry out statistics clustering with detecting the frame, the whole data distribution can be disturbed to unusual detection frame, and the threshold value is not well confirmed, and it is more to filter out child's frame erroneous judgement at last.
The invention provides a grid-clustering-based child detection frame filtering algorithm for detecting the need of filtering child detection frames in public places such as shopping malls, restaurants and the like, fine-grained grids are drawn in an image, an average value frame of each grid is formulated according to the midpoint distribution of the upper edge of each humanoid frame, and whether the child is detected is judged by calculating the difference between a to-be-determined frame and the average value frame. The following problems in the prior art are solved:
1. the method has the advantages that due to the fact that pedestrians are not complete in the detection frame due to the fact that the pedestrians are shielded, whether the detection frame is a child or not cannot be judged directly according to the size of the detection frame;
2. the sizes of frames of the same person at different geographic positions are greatly different due to camera distortion;
3. the sample size of children in the real scene is insufficient, and the children cannot participate in model training on a large scale.
Disclosure of Invention
The invention aims to provide a child detection frame filtering algorithm based on grid clustering to solve the problems in the background technology.
In order to achieve the purpose, the invention provides the following technical scheme: a child detection frame filtering algorithm based on grid clustering comprises the following steps:
(1) acquiring an image to be detected: acquiring video data in a real scene through a camera, and sending the video data to a computer by the camera to acquire and store a target image;
(2) extracting a human shape detection frame: selecting a target camera, pulling detection frame data [ xmin, ymin, xmax, ymax ] corresponding to a circle from a production environment, wherein the detection frame data respectively correspond to the coordinates of the detection frame at the upper left corner and the lower right corner of a target image;
(3) grid clustering calculation average value frame: constructing grids in a target image and calculating an average value frame in each grid, firstly drawing fine-grained grids in the target image, and then counting original detection frames falling into each grid to obtain a mean value frame Box of each grid;
(4) estimating whether the decision block is a child: inputting a to-be-determined block, judging whether the to-be-determined block is child according to the average frame, and judging whether the to-be-determined block is child according to the height difference between the average frame and the to-be-determined block in the grid;
(5) screening and classifying the detection frames: and screening the child frames by combining the mean frame obtained by the grid clustering to finish the distinguishing of the children and the adults.
Preferably, in the step (1), the computer identifies and acquires a target object from the video data, and photographs the target object to obtain a target image.
Preferably, in the step (2), the origin of coordinates of the detection frame data is an upper left corner point of the target image.
Preferably, in the step (3), the size of the grid is customized in advance, a grid range is given, and a grid map with a specified size is generated in the target image.
Preferably, in the step (3), the middle point of the upper edge of each detection frame is used as a base point, and the grid into which the base point falls is calculated to indicate that the corresponding detection frame also falls into the grid, and the original detection frame falling into each grid is counted.
Preferably, in the step (3), the detection frames in each grid are averaged, so as to obtain a mean frame Box of each grid.
Preferably, in the step (4), the target grid into which the midpoint of the upper edge of the decision block falls is calculated by taking the midpoint of the upper edge of the decision block as a base point.
Preferably, in the step (4), the difference criteria are as follows: and measuring the heights of the average frame and the to-be-determined frame, and judging the child if the height of the to-be-determined frame is less than 80% of the average frame.
Preferably, in the step (5), the target image is repeatedly scanned until the child frame is completely screened, and then the scanning is finished.
Preferably, in the step (5), the screened child box can be used as a labeling sample to participate in model training for supervised learning.
Compared with the prior art, the invention has the following beneficial effects:
the child detection frame filtering algorithm provided by the invention can directly take the detected human-shaped frame as statistical data, then uses a grid clustering method to obtain the average frame height in a grid range, judges whether the child is detected or not by utilizing the average frame height and the target frame height, and can judge the average frame height of each area in a target image at the same time, thereby avoiding misjudgment caused by shielding, and solving the problems that more shielding and misdetection exist in a real scene, the statistical clustering is directly carried out by using a detection frame, and the abnormal detection frame can disturb the whole data distribution, so that the misjudgment of the child frame is more.
Drawings
FIG. 1 is a schematic view of the process of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The first embodiment is as follows:
a child detection frame filtering algorithm based on grid clustering comprises the following steps:
(1) acquiring an image to be detected: the method comprises the steps that video data in a real scene are obtained through a camera, the camera sends the video data to a computer, a target image is obtained and stored, the computer identifies and obtains a target object from the video data and photographs the target object to obtain the target image, the target image can be obtained more clearly through screening, and extraction of a detection frame is facilitated;
(2) extracting a human shape detection frame: selecting a target camera, pulling detection frame data [ xmin, ymin, xmax, ymax ] corresponding to one circle from a production environment, wherein the detection frame data respectively correspond to coordinates of a detection frame at the upper left corner and the lower right corner of a target image, and the origin of coordinates of the detection frame data is the upper left corner point of the target image, so that each detection frame can be accurately positioned, and the difference existing between the detection frames is reduced;
(3) grid clustering calculation average value frame: constructing grids in a target image and calculating an average value frame in each grid, firstly drawing fine-grained grids in the target image, and then counting original detection frames falling into each grid to obtain a mean value frame Box of each grid;
(4) estimating whether the decision block is a child: inputting a to-be-determined block, judging whether the to-be-determined block is child according to the average frame, and judging whether the to-be-determined block is child according to the height difference between the average frame and the to-be-determined block in the grid;
(5) screening and classifying the detection frames: and screening the child frames by combining the mean frame obtained by the grid clustering to finish the distinguishing of the children and the adults.
Example two:
a child detection frame filtering algorithm based on grid clustering comprises the following steps:
(1) acquiring an image to be detected: the method comprises the steps that video data in a real scene are obtained through a camera, the camera sends the video data to a computer, a target image is obtained and stored, the computer identifies and obtains a target object from the video data and photographs the target object to obtain the target image, the target image can be obtained more clearly through screening, and extraction of a detection frame is facilitated;
(2) extracting a human shape detection frame: selecting a target camera, pulling detection frame data [ xmin, ymin, xmax, ymax ] corresponding to one circle from a production environment, wherein the detection frame data respectively correspond to coordinates of a detection frame at the upper left corner and the lower right corner of a target image, and the origin of coordinates of the detection frame data is the upper left corner point of the target image, so that each detection frame can be accurately positioned, and the difference existing between the detection frames is reduced;
(3) grid clustering calculation average value frame: constructing grids in a target image and calculating an average frame in each grid, firstly drawing fine-grained grids in the target image, then counting original detection frames falling into each grid to obtain an average frame boxBox of each grid, customizing the size of each grid in advance, giving a grid range, generating a grid map with a specified size in the target image, taking the middle point of the upper edge of each detection frame as a base point, calculating the grids falling into the base point to represent that the corresponding detection frames also fall into the grids, counting the original detection frames falling into each grid, averaging the detection frames in each grid respectively to obtain the average frame Box of each grid, and clustering the grids to obtain the average frame in each grid, thereby being beneficial to comparison and judgment with a decision block;
(4) estimating whether the decision block is a child: inputting a to-be-determined block, judging whether the to-be-determined block is child according to the average frame, and judging whether the to-be-determined block is child according to the height difference between the average frame and the to-be-determined block in the grid;
(5) screening and classifying the detection frames: and screening the child frames by combining the mean frame obtained by the grid clustering to finish the distinguishing of the children and the adults.
Example three:
a child detection frame filtering algorithm based on grid clustering comprises the following steps:
(1) acquiring an image to be detected: the method comprises the steps that video data in a real scene are obtained through a camera, the camera sends the video data to a computer, a target image is obtained and stored, the computer identifies and obtains a target object from the video data and photographs the target object to obtain the target image, the target image can be obtained more clearly through screening, and extraction of a detection frame is facilitated;
(2) extracting a human shape detection frame: selecting a target camera, pulling detection frame data [ xmin, ymin, xmax, ymax ] corresponding to one circle from a production environment, wherein the detection frame data respectively correspond to coordinates of a detection frame at the upper left corner and the lower right corner of a target image, and the origin of coordinates of the detection frame data is the upper left corner point of the target image, so that each detection frame can be accurately positioned, and the difference existing between the detection frames is reduced;
(3) grid clustering calculation average value frame: constructing grids in a target image and calculating an average frame in each grid, firstly drawing fine-grained grids in the target image, then counting original detection frames falling into each grid to obtain an average frame Box of each grid, customizing the size of each grid in advance, giving a grid range, generating a grid map with a specified size in the target image, taking the middle point of the upper edge of each detection frame as a base point, calculating the grids falling into the base point to represent that the corresponding detection frames also fall into the grids, counting the original detection frames falling into each grid, averaging the detection frames in each grid respectively to obtain the average frame Box of each grid, and obtaining the average frame in each grid through grid clustering so as to be beneficial to comparison and judgment with a decision block;
(4) estimating whether the decision block is a child: inputting a to-be-determined block, judging whether the to-be-determined block is child according to the average frame, judging whether the to-be-determined block is child according to the height difference between the average frame in the grid and the to-be-determined block, calculating a target grid into which the midpoint of the upper edge of the determination block falls by taking the midpoint of the upper edge of the determination block as a base point, wherein the difference standard is as follows: the heights of the average frame and the to-be-determined frame are measured, if the height of the to-be-determined frame is less than 80% of the average frame, the child frame is determined to be child, the child frame and the adult frame are conveniently distinguished, the detection precision of the child frame can be effectively improved, and the child frame is prevented from being misjudged;
(5) screening and classifying the detection frames: the method includes the steps that a child frame is screened by combining a mean frame obtained through grid clustering, a child and an adult are distinguished, a target image is repeatedly scanned, scanning is finished until the child frame is completely screened, the screened child frame can be used as a labeling sample, supervision learning is conducted by participating in model training, omission in scanning can be prevented, the purpose of increasing screening precision is achieved, and meanwhile model training of a detection frame can be participated.
The child detection frame filtering algorithm provided by the invention can directly take the detected human-shaped frame as statistical data, then uses a grid clustering method to obtain the average frame height in a grid range, judges whether the child is detected or not by utilizing the average frame height and the target frame height, and can judge the average frame height of each area in a target image at the same time, thereby avoiding misjudgment caused by shielding, and solving the problems that more shielding and misdetection exist in a real scene, the statistical clustering is directly carried out by using a detection frame, and the abnormal detection frame can disturb the whole data distribution, so that the misjudgment of the child frame is more.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (10)
1. A child detection frame filtering algorithm based on grid clustering is characterized in that: the method comprises the following steps:
(1) acquiring an image to be detected: acquiring video data in a real scene through a camera, and sending the video data to a computer by the camera to acquire and store a target image;
(2) extracting a human shape detection frame: selecting a target camera, pulling detection frame data [ xmin, ymin, xmax, ymax ] corresponding to a circle from a production environment, wherein the detection frame data respectively correspond to the coordinates of the detection frame at the upper left corner and the lower right corner of a target image;
(3) grid clustering calculation average value frame: constructing grids in a target image and calculating an average value frame in each grid, firstly drawing fine-grained grids in the target image, and then counting original detection frames falling into each grid to obtain a mean value frame Box of each grid;
(4) estimating whether the decision block is a child: inputting a to-be-determined block, judging whether the to-be-determined block is child according to the average frame, and judging whether the to-be-determined block is child according to the height difference between the average frame and the to-be-determined block in the grid;
(5) screening and classifying the detection frames: and screening the child frames by combining the mean frame obtained by the grid clustering to finish the distinguishing of the children and the adults.
2. The child detection box filtering algorithm based on grid clustering as claimed in claim 1, wherein: in the step (1), the computer identifies and acquires the target object from the video data, and photographs the target object to obtain the target image.
3. The child detection box filtering algorithm based on grid clustering as claimed in claim 1, wherein: in the step (2), the origin of coordinates of the detection frame data is the upper left corner point of the target image.
4. The child detection box filtering algorithm based on grid clustering as claimed in claim 1, wherein: in the step (3), the size of the grid is customized in advance, a grid range is given, and a grid map with the specified size is generated in the target image.
5. The child detection box filtering algorithm based on grid clustering as claimed in claim 1, wherein: in the step (3), the middle point of the upper edge of each detection frame is used as a base point, and the grid into which the base point falls is calculated to represent that the corresponding detection frame also falls into the grid, and the original detection frame falling into each grid is counted.
6. The child detection box filtering algorithm based on grid clustering as claimed in claim 1, wherein: in the step (3), the detection frames in each grid are averaged to obtain a mean frame Box of each grid.
7. The child detection box filtering algorithm based on grid clustering as claimed in claim 1, wherein: in the step (4), the target grid into which the midpoint of the upper edge of the decision block falls is calculated by taking the midpoint of the upper edge of the decision block as a base point.
8. The child detection box filtering algorithm based on grid clustering as claimed in claim 1, wherein: in the step (4), the difference criteria are as follows: and measuring the heights of the average frame and the to-be-determined frame, and judging the child if the height of the to-be-determined frame is less than 80% of the average frame.
9. The child detection box filtering algorithm based on grid clustering as claimed in claim 1, wherein: and (5) repeatedly scanning the target image until the child frame is completely screened, and then finishing scanning.
10. The child detection box filtering algorithm based on grid clustering as claimed in claim 1, wherein: and (5) the screened child frames can be used as marking samples to participate in model training for supervised learning.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110633366.2A CN113361410B (en) | 2021-06-07 | 2021-06-07 | Child detection box filtering algorithm based on grid clustering |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110633366.2A CN113361410B (en) | 2021-06-07 | 2021-06-07 | Child detection box filtering algorithm based on grid clustering |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113361410A true CN113361410A (en) | 2021-09-07 |
| CN113361410B CN113361410B (en) | 2022-08-26 |
Family
ID=77532910
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110633366.2A Active CN113361410B (en) | 2021-06-07 | 2021-06-07 | Child detection box filtering algorithm based on grid clustering |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113361410B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103646250A (en) * | 2013-09-13 | 2014-03-19 | 魏运 | Pedestrian monitoring method and device based on distance image head and shoulder features |
| CN106022235A (en) * | 2016-05-13 | 2016-10-12 | 中国人民解放军国防科学技术大学 | Missing child detection method based on human body detection |
| CN109484935A (en) * | 2017-09-13 | 2019-03-19 | 杭州海康威视数字技术股份有限公司 | A kind of lift car monitoring method, apparatus and system |
| CN111639616A (en) * | 2020-06-05 | 2020-09-08 | 上海一由科技有限公司 | Heavy identity recognition method based on deep learning |
| CN111854963A (en) * | 2020-06-11 | 2020-10-30 | 浙江大华技术股份有限公司 | Temperature detection method, device, equipment and computer equipment |
| CN112513870A (en) * | 2018-08-08 | 2021-03-16 | 索博客科技有限公司 | System and method for detecting, tracking and counting human objects of interest with improved altimetry |
-
2021
- 2021-06-07 CN CN202110633366.2A patent/CN113361410B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103646250A (en) * | 2013-09-13 | 2014-03-19 | 魏运 | Pedestrian monitoring method and device based on distance image head and shoulder features |
| CN106022235A (en) * | 2016-05-13 | 2016-10-12 | 中国人民解放军国防科学技术大学 | Missing child detection method based on human body detection |
| CN109484935A (en) * | 2017-09-13 | 2019-03-19 | 杭州海康威视数字技术股份有限公司 | A kind of lift car monitoring method, apparatus and system |
| CN112513870A (en) * | 2018-08-08 | 2021-03-16 | 索博客科技有限公司 | System and method for detecting, tracking and counting human objects of interest with improved altimetry |
| CN111639616A (en) * | 2020-06-05 | 2020-09-08 | 上海一由科技有限公司 | Heavy identity recognition method based on deep learning |
| CN111854963A (en) * | 2020-06-11 | 2020-10-30 | 浙江大华技术股份有限公司 | Temperature detection method, device, equipment and computer equipment |
Non-Patent Citations (2)
| Title |
|---|
| TONGYU ZHU 等: "A Grid-based Approach to Continuous Clustering of Moving Objects", 《ADVCOMP 2010》 * |
| 刘弋锋: "基于浅层学习引导深度学习的行人检测", 《中国优秀博硕士学位论文全文数据库(博士)信息科技辑》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113361410B (en) | 2022-08-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110458061B (en) | Method for identifying old people falling down and accompanying robot | |
| CN102803991B (en) | Object detection device | |
| CN103632158B (en) | Forest fire prevention monitor method and forest fire prevention monitor system | |
| CN109670396A (en) | A kind of interior Falls Among Old People detection method | |
| CN105469105A (en) | Cigarette smoke detection method based on video monitoring | |
| JP5287333B2 (en) | Age estimation device | |
| CN111191576A (en) | Personnel behavior target detection model construction method, intelligent analysis method and system | |
| CN111832489A (en) | Subway crowd density estimation method and system based on target detection | |
| CN106503643B (en) | Tumble detection method for human body | |
| CN110390229B (en) | Face picture screening method and device, electronic equipment and storage medium | |
| CN108564577A (en) | Solar cell segment grid defect inspection method based on convolutional neural networks | |
| CN112183472A (en) | Method for detecting whether test field personnel wear work clothes or not based on improved RetinaNet | |
| KR20150108118A (en) | Remote automatic metering system based image recognition | |
| CN114972356A (en) | Plastic product surface defect detection and identification method and system | |
| CN112257549B (en) | Floor danger detection early warning method and system based on computer vision | |
| CN104036485A (en) | Method about image resampling tampering detection | |
| CN109583307A (en) | A kind of Cashmere and Woolens fiber recognition method based on local feature Yu word packet model | |
| CN112257615A (en) | Clustering-based customer number statistical method | |
| CN106408563A (en) | Snow noise detection method based on variation coefficient | |
| CN111127433A (en) | Method and device for detecting flame | |
| CN113361410B (en) | Child detection box filtering algorithm based on grid clustering | |
| CN105427303B (en) | A kind of vision measurement and method of estimation of substation's legacy | |
| CN103077403B (en) | pedestrian counting method and device | |
| CN114648683B (en) | Neural network performance improving method and device based on uncertainty analysis | |
| CN113706815B (en) | Vehicle fire identification method combining YOLOv3 and optical flow 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 |