CN108805021A - The real-time individual tumble behavioral value alarm method of feature based operator - Google Patents
The real-time individual tumble behavioral value alarm method of feature based operator Download PDFInfo
- Publication number
- CN108805021A CN108805021A CN201810393684.4A CN201810393684A CN108805021A CN 108805021 A CN108805021 A CN 108805021A CN 201810393684 A CN201810393684 A CN 201810393684A CN 108805021 A CN108805021 A CN 108805021A
- Authority
- CN
- China
- Prior art keywords
- tumble
- boundary rectangle
- human
- real
- human body
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
- G06V40/23—Recognition of whole body movements, e.g. for sport training
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of real-time individual tumble behavioral value alarm methods of feature based operator, are specifically implemented according to the following steps:To the modeling of video mix Gaussian Background, binaryzation, Canny processing and Morphological scale-space, after extracting movement human target image body motion information is extracted from obtained movement human target image, to moving target into line trace, the tracking information for recording movement human, draws the boundary rectangle of movement human;Calculate gravity center of human body's altitude rate, and judge whether to meet standard, and then judge whether that tumble behavior occurs, calculate the high long ratio of boundary rectangle, and standard when determining whether to meet tumble, when judgement human body is in tumble behavior, record human body is there is a situation where the time of tumble behavior and judges whether to belong to injured and can not stand, and makes alert process;The method of the present invention is by using customized feature operator, to judge human normal behavior and tumble behavior, can be good at ensureing real-time.
Description
Technical field
The invention belongs to abnormal behavior alarm technique fields, and in particular to a kind of real-time individual tumble of feature based operator
Behavioral value alarm method.
Background technology
Child, the elderly and some patients are the groups for being easiest to occur tumble behavior, once and these people have occurred
Tumble behavior, and oneself cannot stand up, it is possible to influence whether the personal safety of these monitoring crowds, and if this
The abnormal behaviour system that can be monitored detects in time, and makes corresponding processing, is just effective to ensure that this kind of monitoring crowd
Personal safety, therefore, the old men such as park, cell, square, hospital, child and patient's frequency of occurrences it is very high place peace
Dress abnormal alarm system tool plays a very important role.
It is shown according to World Health Organization's related data, tumble is dead second largest of accident or Unintentional injury all over the world
Hidden danger.The whole world is estimated to be 42.4 ten thousand people every year leads to death because falling injury.It is numerous dead and non-straight caused by tumble
It connects to be derived from and drops to itself, but the best opportunity is missed because of the timely relief not having after falling.Therefore, a kind of reality is realized
When effective fall detection alarm method be of great significance in monitoring arts.
Invention content
The object of the present invention is to provide a kind of real-time individual tumble behavioral value alarm methods of feature based operator, pass through
With customized feature operator, to judge human normal behavior and tumble behavior, can be good at ensureing real-time.
The technical solution adopted in the present invention is a kind of real-time individual tumble behavioral value alarm side of feature based operator
Method is specifically implemented according to the following steps:
Step 1, to the modeling of video mix Gaussian Background, binaryzation, Canny processing and Morphological scale-space, movement is extracted
Human body target image,
Step 2, extract body motion information in the movement human target image obtained from step 1, to moving target into
Line trace records the tracking information of movement human, draws the boundary rectangle of movement human;
Step 3, according to the body motion information acquired in step 2, gravity center of human body's altitude rate is calculated, and judge whether
Meet standard, and then judge whether that tumble behavior occurs,
Step 4, standard when calculating the high long ratio of boundary rectangle, and determining whether to meet tumble,
Step 5, when step 4 judgement human body is in tumble behavior, the time t of tumble behavior occurs for start recording human body, such as
Fruit t>2min, ginseng then think oneself not restoring to stand in human body long-time, illustrate that human body may be injured, at this point, system
Make alert process;Otherwise, without any processing.
The features of the present invention also characterized in that
The step 1 is specifically, the image acquired in real time using the extraction of mixed Gaussian background modeling method, passes through two-value
Change is handled, and extracts movement human target, is done Morphological scale-space to binary image, is removed the noise in image, utilize simultaneously
Canny edge detections obtain movement human target image.
The associated motion information of movement human target for extracting and recording in the step 2, including outside barycentric coodinates
Connect the sum of all pixels that the length of rectangle is 1 and 0 with pixel value in high and boundary rectangle.
The step 2 specifically,
Step 2.1, utilize camshif combination Kalman algorithms into line trace the foreground target that step 1 is extracted, and to carrying
The binary image of the movement human target of taking-up is projected,
Step 2.2, it is traversed once from the left end of image to right end first, records its abscissa X respectivelyminWith
Xmax;It is traversed once from the bottom of image to the top again, finds out the white pixel point of bottom and the top, record respectively
Lower its ordinate YminAnd Ymax;
Step 2.3, the height Height and width Length of boundary rectangle are found out according to formula (1) and formula (2) respectively,
Height=Ymax-Ymin (1)
Length=Xmax-Xmin (2)
XminAnd YminIndicate the minimum abscissa of boundary rectangle and minimum ordinate, XmaxAnd YmaxIt is the maximum of boundary rectangle
Abscissa and maximum ordinate;
Step 2.4, according to the height and width of the calculated boundary rectangle of step 2.3, the external square of movement human is drawn
Shape.
The step 3 specifically,
Step 3.1, the center of gravity that movement human is identified using the gray scale barycenter of boundary rectangle traverses entire boundary rectangle area
Domain, i.e., from XminTo XmaxWhile traversing often row, from YminTo YmaxEach column is traversed, calculates separately out in circumscribed rectangular region and owns
The sum of gray value of pixel ∑(x,y)∈RF (x, y), all pixels point and corresponding x coordinate sum of products x ∑s(x,y)∈Rf(x,y),
All pixels point and corresponding y-coordinate sum of products y ∑s(x,y)∈RWherein, R is external just region to f (x, y), and f (x, y) is external
Gray value in rectangular area at pixel (x, y) coordinate.External rectangle region is calculated separately out using formula (3) and formula (4)
The gray scale barycenter in domain namely the center of gravity X in movement human regionwAnd Yw:
Pinpoint coordinate W (Xw,Yw) it is the gray scale barycenter of circumscribed rectangular region namely the center of gravity of movement human.
Step 3.2, height of C.G. change rate is calculated, and whether preliminary judgement human body has the variation that center of gravity reduces.
The step 3.2 specifically,
Gravity center of human body's ordinate Y is directly used in the decline of center of gravitywReduction reflect, calculated separately according to formula (3) and (4)
Go out the horizontal seat X at movement human center in each frame image of present framew(i) and ordinate Yw(i), it is then calculated according to formula (5)
The difference in height WHeight (i) of gravity center of human body and boundary rectangle bottom.That is,
WHeight (i)=Yw(i)-Ymin(i) (5)
Wherein i refers to the i-th frame image, i=1,2,3 ... ..., n ... ....
Assuming that present frame is n-th frame, the average value of WHeight in all frame images before n-th frame is then sought:
Then, movement human height of C.G. change rate is calculated according to formula (7),
WHeightRate=WHeight/averageWHeight (7)
When height of C.G. change rate is less than 1, it is determined as that mass centre changing occurs for human body, tumble behavior occurs, at this point, turns step
Rapid 4 are further judged.
The step 4 specifically,
Step 4.1, boundary rectangle height is calculated to grow than Y2X,
Step 4.2, judge whether value of the boundary rectangle high length than Y2X is less than 1, if it is, being determined as that human body has occurred
Tumble behavior continues step 5, otherwise, return to step 1.
Boundary rectangle high length in the step 4.1 is calculated than Y2X according to formula (8),
Wherein, Y2X represents the long ratio of height of outer rectangular.
The beneficial effects of the invention are as follows:A kind of real-time individual tumble behavioral value alarm side of feature based operator of the present invention
Method realizes the developing direction acted according to current behavior action prediction, and real-time to the abnormal progress currently having occurred and that
Analysis, processing;Greatly reduce the accident rate that can not fallen and squat down for a long time for a long time etc. and bring;Cloud service is not needed
Device, real-time are high.
Description of the drawings
Fig. 1 is the algorithm flow chart of the real-time individual tumble behavioral value alarm method the present invention is based on feature operator;
Fig. 2 be feature based operator real-time individual tumble behavioral value alarm method in image be binarized rear image;
Fig. 3 is the real-time individual tumble behavioral value alarm method the present invention is based on feature operator for the row that traces into
People carries out the schematic diagram after Canny edge detections;
Fig. 4 is the rectangle for pedestrian of the real-time individual tumble behavioral value alarm method the present invention is based on feature operator
Collimation mark knows figure;
Fig. 5 is the boundary rectangle gray scale matter of the real-time individual tumble behavioral value alarm method the present invention is based on feature operator
The schematic diagram of heart point;
Fig. 6 is the height of C.G. change rate of the real-time individual tumble behavioral value alarm method the present invention is based on feature operator
Schematic diagram;
Fig. 7 is the high long ratio of boundary rectangle in the real-time individual tumble behavioral value alarm method the present invention is based on feature operator
Schematic diagram;
Fig. 8 is pedestrian's tumble signal of the real-time individual tumble behavioral value alarm method the present invention is based on feature operator
Figure.
Specific implementation mode
The following describes the present invention in detail with reference to the accompanying drawings and specific embodiments.
The real-time individual tumble behavioral value alarm method of a kind of feature based operator of the present invention, as shown in Figure 1, specifically pressing
Implement according to following steps:
Step 1, to the modeling of video mix Gaussian Background, binaryzation, Canny processing and Morphological scale-space, movement is extracted
Human body target image, the specific steps are:
The image acquired in real time using the extraction of mixed Gaussian background modeling method extracts movement by binary conversion treatment
Human body target is done Morphological scale-space to binary image, removes the noise in image, while being obtained using Canny edge detections
Movement human target image;
Step 2, extract body motion information in the movement human target image obtained from step 1, to moving target into
Line trace records the tracking information of movement human, draws the boundary rectangle of movement human.
Wherein, the associated motion information for the movement human target extracted and recorded, including barycentric coodinates boundary rectangle
The sum of all pixels that pixel value is 1 and 0 in long and high and boundary rectangle;
The specific steps are:
Step 2.1, utilize camshif combination Kalman algorithms into line trace the foreground target that step 1 is extracted, and to carrying
The binary image of the movement human target of taking-up is projected, as shown in Figure 3
Step 2.2, it is traversed once from the left end of image to right end first, records its abscissa X respectivelyminWith
Xmax;It is traversed once from the bottom of image to the top again, finds out the white pixel point of bottom and the top, record respectively
Lower its ordinate YminAnd Ymax。
Step 2.3, the height Height and width Length of boundary rectangle are found out according to formula (1) and formula (2) respectively,
Height=Ymax-Ymin (1)
Length=Xmax-Xmin (2)
XminAnd YminIndicate the minimum abscissa of boundary rectangle and minimum ordinate, XmaxAnd YmaxIt is the maximum of boundary rectangle
Abscissa and maximum ordinate.
Step 2.4, according to the height and width of the calculated boundary rectangle of step 2.3, the external square of movement human is drawn
Shape;As shown in Figure 4.
Step 3, according to the body motion information acquired in step 2, gravity center of human body's altitude rate is calculated, and judge whether
Meeting standard, and then judges whether that tumble behavior occurs, specific method is,
Step 3.1, the center of gravity that movement human is identified using the gray scale barycenter of boundary rectangle traverses entire boundary rectangle area
Domain, i.e., from XminTo XmaxWhile traversing often row, from YminTo YmaxEach column is traversed, calculates separately out in circumscribed rectangular region and owns
The sum of gray value of pixel ∑(x,y)∈RF (x, y), all pixels point and corresponding x coordinate sum of products x ∑s(x,y)∈Rf(x,y),
All pixels point and corresponding y-coordinate sum of products y ∑s(x,y)∈RWherein, R is external just region to f (x, y), and f (x, y) is external
Gray value in rectangular area at pixel (x, y) coordinate.External rectangle region is calculated separately out using formula (3) and formula (4)
The gray scale barycenter in domain namely the center of gravity X in movement human regionwAnd Yw:
Pinpoint coordinate W (Xw,Yw) it is the gray scale barycenter of circumscribed rectangular region namely the center of gravity of moving object.Shown in Fig. 5 i.e.
For the gray scale barycenter of label.
Step 3.2, height of C.G. change rate is calculated, and whether preliminary judgement human body has the variation that center of gravity reduces, specific side
Method is,
The decline of center of gravity can directly use gravity center of human body's ordinate YwReduction reflect, according to formula (3) and (4) point
The horizontal seat X at movement human center in each frame image of present frame is not calculatedw(i) and ordinate Yw(i), then according to formula
(5) the difference in height WHeight (i) of gravity center of human body and boundary rectangle bottom are calculated.That is,
WHeight (i)=Yw(i)-Ymin(i) (5)
Wherein i refers to the i-th frame image, i=1,2,3 ... ..., n ... ....
Assuming that present frame is n-th frame, the average value of WHeight in all frame images before n-th frame is then sought:
Then, movement human height of C.G. change rate is calculated according to formula (7),
WHeightRate=WHeight/averageWHeight (7)
Under normal circumstances, when movement human normal walking, the value of height of C.G. change rate should be equal to 1, and when center of gravity has
Apparent reduction should also be had by being decreased obviously the value of height of C.G. change rate, and can decrease below 1 some value always.
Therefore it is different whether to fall and squat down as judgement movement human that the value of movement human height of C.G. change rate may be used
An index of Chang Hangwei, so, when height of C.G. change rate is not equal to 1, it is determined as that mass centre changing occurs for human body, falls
Backward is further to be judged at this point, going to step 4.Fig. 6 indicates height of C.G. change rate schematic diagram.
Step 4, standard when calculating the high long ratio of boundary rectangle, and determining whether to meet tumble,
Step 4.1, the high long ratio of boundary rectangle is calculated,
Height of C.G. change rate makes squatting motion just for the sake of explanation human body, it cannot be said that and whether person of good sense's body falls, this
Need the judgement of other indexs.For case above, whether the present invention is occurred using boundary rectangle high length ratio as movement human
Another criterion of tumble behavior and behavior of squatting down.
The boundary rectangle high length ratio is calculated according to formula (8),
Wherein, Y2X represents the long ratio of height of outer rectangular.Fig. 7 indicates that boundary rectangle high length compares schematic diagram.
Step 4.2, judge whether value of the boundary rectangle high length than Y2X is less than 1, if it is, being determined as that human body has occurred
Tumble behavior continues step 5, otherwise, return to step 1.Fig. 8 shows human body tumble schematic diagrames.
Step 5:When step 4 judgement human body is in tumble behavior, the time t of tumble behavior occurs for start recording human body, such as
Fruit t>2min, ginseng then think oneself not restoring to stand in human body long-time, illustrate that human body may be injured, at this point, system
Make alert process.Otherwise, without any processing.
In order to verify the accuracy and validity of method of the invention, use sensitivity and specificity as assessment we
The index of method, the calculation formula (9) of sensitivity and the calculation formula (10) of specificity are as follows:
Wherein, TP is the correctly predicted number of falls of system.FN is the practical number of falls that system is missed.FP is wrong inspection
The quantity of survey.TN is the quantity of normal activity.
Experiment shows the individual anomaly detection method and other existing detection methods of feature based operator of the invention
Reliable as a result, comparing result is as shown in table 1 compared to achieving, biomechanical approach can reach 100% sensitivity and spy
The opposite sex, but people always need wearable sensors.
1 the method for the present invention of table is compared with the sensibility of other existing detection methods and specificity
The method of the present invention | Chen’s | Chua’s | MHI | Biomechanics | |
Sensibility (%) | 88.43 | 90.9 | 90.5 | 85.7 | 100.0 |
Specific (%) | 95.86 | 93.8 | 93.3 | 80.0 | 100.0 |
2 the method for the present invention of table handles sensibility and specificity contrast table when different cameral data
Camera1 | Camera2 | Camera3 | Average | |
Sensibility (%) | 85.81 | 90.07 | 89.42 | 88.43 |
Specific (%) | 99.34 | 97.23 | 91.02 | 95.86 |
The experimental results showed that the data set of the method for the present invention realized in camera 1 85.81% sensitivity and
99.34% specificity, and it is higher than other methods, as shown in table 2, the sensitivity and 95.86% that mean apparent is 88.43%
Specificity also superior to other methods.
Claims (8)
1. a kind of real-time individual tumble behavioral value alarm method of feature based operator, which is characterized in that specifically according to following
Step is implemented:
Step 1, to the modeling of video mix Gaussian Background, binaryzation, Canny processing and Morphological scale-space, movement human is extracted
Target image,
Step 2, extract body motion information in the movement human target image obtained from step 1, to moving target carry out with
Track records the tracking information of movement human, draws the boundary rectangle of movement human;
Step 3, according to the body motion information acquired in step 2, gravity center of human body's altitude rate is calculated, and judge whether to meet
Standard, and then judge whether that tumble behavior occurs,
Step 4, standard when calculating the high long ratio of boundary rectangle, and determining whether to meet tumble,
Step 5, when step 4 judgement human body is in tumble behavior, the time t of tumble behavior occurs for start recording human body, if t>
2min, ginseng then think oneself not restoring to stand in human body long-time, illustrate that human body may be injured, at this point, system is made
Alert process;Otherwise, without any processing.
2. the real-time individual tumble behavioral value alarm method of feature based operator according to claim 1, feature exist
In the step 1 is specifically, extract the image acquired in real time using mixed Gaussian background modeling method, at binaryzation
Reason, extracts movement human target, and Morphological scale-space is done to binary image, removes the noise in image, while utilizing Canny
Edge detection obtains movement human target image.
3. the real-time individual tumble behavioral value alarm method of feature based operator according to claim 1, feature exist
In, the associated motion information of movement human target for extracting and recording in the step 2, including barycentric coodinates boundary rectangle
Length and high and boundary rectangle in pixel value be 1 and 0 sum of all pixels.
4. the real-time individual tumble behavioral value alarm method of feature based operator according to claim 1, feature exist
In, the step 2 specifically,
Step 2.1, utilize camshif combination Kalman algorithms into line trace the foreground target that step 1 is extracted, and to extracting
The binary image of movement human target projected,
Step 2.2, it is traversed once from the left end of image to right end first, records its abscissa X respectivelyminAnd Xmax;Again
It is traversed once from the bottom of image to the top, finds out the white pixel point of bottom and the top, it is vertical to record it respectively
Coordinate YminAnd Ymax;
Step 2.3, the height Height and width Length of boundary rectangle are found out according to formula (1) and formula (2) respectively,
Height=Ymax-Ymin (1)
Length=Xmax-Xmin (2)
XminAnd YminIndicate the minimum abscissa of boundary rectangle and minimum ordinate, XmaxAnd YmaxIt is the horizontal seat of maximum of boundary rectangle
Mark and maximum ordinate;
Step 2.4, according to the height and width of the calculated boundary rectangle of step 2.3, the boundary rectangle of movement human is drawn.
5. the real-time individual tumble behavioral value alarm method of feature based operator according to claim 1, feature exist
In, the step 3 specifically,
Step 3.1, the center of gravity that movement human is identified using the gray scale barycenter of boundary rectangle traverses entire circumscribed rectangular region,
I.e. from XminTo XmaxWhile traversing often row, from YminTo YmaxEach column is traversed, all pictures in circumscribed rectangular region are calculated separately out
The sum of gray value of vegetarian refreshments ∑(x,y)∈RF (x, y), all pixels point and corresponding x coordinate sum of products x ∑s(x,y)∈RF (x, y), institute
There are pixel and corresponding y-coordinate sum of products y ∑s(x,y)∈RWherein, R is external just region to f (x, y), and f (x, y) is external square
Gray value in shape region at pixel (x, y) coordinate;Circumscribed rectangular region is calculated separately out using formula (3) and formula (4)
Gray scale barycenter namely movement human region center of gravity XwAnd Yw:
Pinpoint coordinate W (Xw,Yw) it is the gray scale barycenter of circumscribed rectangular region namely the center of gravity of moving object;
Step 3.2, height of C.G. change rate is calculated, and whether preliminary judgement human body has the variation that center of gravity reduces.
6. the real-time individual tumble behavioral value alarm method of feature based operator according to claim 5, feature exist
In, the step 3.2 specifically,
Gravity center of human body's ordinate Y is directly used in the decline of center of gravitywReduction reflect, according to formula (3) and (4) calculate separately out and work as
The horizontal seat X at movement human center in each frame image of previous framew(i) and ordinate Yw(i), human body is then calculated according to formula (5)
The difference in height WHeight (i) of center of gravity and boundary rectangle bottom;That is,
WHeight (i)=Yw(i)-Ymin(i) (5)
Wherein i refers to the i-th frame image, i=1,2,3 ... ..., n ... ...;
Assuming that present frame is n-th frame, the average value of WHeight in all frame images before n-th frame is then sought:
Then, movement human height of C.G. change rate is calculated according to formula (7),
WHeightRate=WHeight/averageWHeight (7)
When height of C.G. change rate is not equal to 1, it is determined as that mass centre changing occurs for human body, tumble behavior occurs, at this point, going to step
4 are further judged.
7. the real-time individual tumble behavioral value alarm method of feature based operator according to claim 1, feature exist
In, the step 4 specifically,
Step 4.1, boundary rectangle height is calculated to grow than Y2X,
Step 4.2, judge whether value of the boundary rectangle high length than Y2X is less than 1, if it is, being determined as that human body is fallen
Behavior continues step 5, otherwise, return to step 1.
8. the real-time individual tumble behavioral value alarm method of feature based operator according to claim 1, feature exist
In, the boundary rectangle high length in the step 4.1 is calculated than Y2X according to formula (8),
Wherein, Y2X represents the long ratio of height of outer rectangular.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810393684.4A CN108805021A (en) | 2018-04-27 | 2018-04-27 | The real-time individual tumble behavioral value alarm method of feature based operator |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810393684.4A CN108805021A (en) | 2018-04-27 | 2018-04-27 | The real-time individual tumble behavioral value alarm method of feature based operator |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108805021A true CN108805021A (en) | 2018-11-13 |
Family
ID=64093529
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810393684.4A Pending CN108805021A (en) | 2018-04-27 | 2018-04-27 | The real-time individual tumble behavioral value alarm method of feature based operator |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108805021A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111428546A (en) * | 2019-04-11 | 2020-07-17 | 杭州海康威视数字技术股份有限公司 | Method and device for marking human body in image, electronic equipment and storage medium |
CN113740848A (en) * | 2021-09-30 | 2021-12-03 | 中科芯集成电路有限公司 | Tracking type fall detection device for solitary old people based on millimeter wave radar |
CN113920537A (en) * | 2021-10-15 | 2022-01-11 | 江阴市立青染整机械有限公司 | State judgment platform using image detection |
CN115273401A (en) * | 2022-08-03 | 2022-11-01 | 浙江慧享信息科技有限公司 | Method and system for automatically sensing falling of person |
US11604254B2 (en) | 2019-08-16 | 2023-03-14 | Fujitsu Limited | Radar-based posture recognition apparatus and method and electronic device |
CN116091983A (en) * | 2023-04-10 | 2023-05-09 | 四川弘和通讯集团有限公司 | Behavior detection method and device, electronic equipment and storage medium |
-
2018
- 2018-04-27 CN CN201810393684.4A patent/CN108805021A/en active Pending
Non-Patent Citations (1)
Title |
---|
FACUN ZHANG ET AL: "《Research on Children’s Fall Detection by Characteristic Operator》", 《ICAIP 2017》 * |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111428546A (en) * | 2019-04-11 | 2020-07-17 | 杭州海康威视数字技术股份有限公司 | Method and device for marking human body in image, electronic equipment and storage medium |
CN111428546B (en) * | 2019-04-11 | 2023-10-13 | 杭州海康威视数字技术股份有限公司 | Method and device for marking human body in image, electronic equipment and storage medium |
US11604254B2 (en) | 2019-08-16 | 2023-03-14 | Fujitsu Limited | Radar-based posture recognition apparatus and method and electronic device |
CN113740848A (en) * | 2021-09-30 | 2021-12-03 | 中科芯集成电路有限公司 | Tracking type fall detection device for solitary old people based on millimeter wave radar |
CN113740848B (en) * | 2021-09-30 | 2023-11-17 | 中科芯集成电路有限公司 | Solitary old person tracking type falling detection device based on millimeter wave radar |
CN113920537A (en) * | 2021-10-15 | 2022-01-11 | 江阴市立青染整机械有限公司 | State judgment platform using image detection |
CN115273401A (en) * | 2022-08-03 | 2022-11-01 | 浙江慧享信息科技有限公司 | Method and system for automatically sensing falling of person |
CN115273401B (en) * | 2022-08-03 | 2024-06-14 | 浙江慧享信息科技有限公司 | Method and system for automatically sensing falling of person |
CN116091983A (en) * | 2023-04-10 | 2023-05-09 | 四川弘和通讯集团有限公司 | Behavior detection method and device, electronic equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108805021A (en) | The real-time individual tumble behavioral value alarm method of feature based operator | |
CN109919132B (en) | Pedestrian falling identification method based on skeleton detection | |
US9996739B2 (en) | System and method for automatic gait cycle segmentation | |
CN105574501B (en) | A kind of stream of people's video detecting analysis system | |
Pediaditis et al. | Vision-based motion detection, analysis and recognition of epileptic seizures—a systematic review | |
Dubois et al. | Human activities recognition with RGB-Depth camera using HMM | |
Shoaib et al. | View-invariant fall detection for elderly in real home environment | |
CN110781733B (en) | Image duplicate removal method, storage medium, network equipment and intelligent monitoring system | |
Yang et al. | Fall detection for multiple pedestrians using depth image processing technique | |
CN109271918B (en) | Method for distinguishing people with balance ability disorder based on gravity center shift model | |
US20220036056A1 (en) | Image processing apparatus and method for recognizing state of subject | |
CN114469076B (en) | Identity-feature-fused fall identification method and system for solitary old people | |
SG188111A1 (en) | Condition detection methods and condition detection devices | |
CN111259718A (en) | Escalator retention detection method and system based on Gaussian mixture model | |
CN111243230B (en) | Human body falling detection device and method based on two depth cameras | |
Zhang et al. | Visual surveillance for human fall detection in healthcare IoT | |
CN115116127A (en) | Fall detection method based on computer vision and artificial intelligence | |
CN111144174A (en) | System for identifying falling behavior of old people in video by using neural network and traditional algorithm | |
Nguyen et al. | Extracting silhouette-based characteristics for human gait analysis using one camera | |
Fleyeh et al. | Extracting body landmarks from videos for parkinson gait analysis | |
Suriani et al. | Sudden fall classification using motion features | |
Dorgham et al. | Improved elderly fall detection by surveillance video using real-time human motion analysis | |
Lee et al. | Automated abnormal behavior detection for ubiquitous healthcare application in daytime and nighttime | |
Christodoulidis et al. | Near real-time human silhouette and movement detection in indoor environments using fixed cameras | |
An et al. | Support vector machine algorithm for human fall recognition kinect-based skeletal data |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181113 |
|
RJ01 | Rejection of invention patent application after publication |