CN106127188A - A kind of Handwritten Digit Recognition method based on gyroscope - Google Patents
A kind of Handwritten Digit Recognition method based on gyroscope Download PDFInfo
- Publication number
- CN106127188A CN106127188A CN201610524482.XA CN201610524482A CN106127188A CN 106127188 A CN106127188 A CN 106127188A CN 201610524482 A CN201610524482 A CN 201610524482A CN 106127188 A CN106127188 A CN 106127188A
- Authority
- CN
- China
- Prior art keywords
- section
- zero
- angular velocity
- gesture
- 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
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
- G06V40/30—Writer recognition; Reading and verifying signatures
- G06V40/37—Writer recognition; Reading and verifying signatures based only on signature signals such as velocity or pressure, e.g. dynamic signature recognition
- G06V40/376—Acquisition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/011—Arrangements for interaction with the human body, e.g. for user immersion in virtual reality
- G06F3/014—Hand-worn input/output arrangements, e.g. data gloves
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/03—Arrangements for converting the position or the displacement of a member into a coded form
- G06F3/033—Pointing devices displaced or positioned by the user, e.g. mice, trackballs, pens or joysticks; Accessories therefor
- G06F3/0346—Pointing devices displaced or positioned by the user, e.g. mice, trackballs, pens or joysticks; Accessories therefor with detection of the device orientation or free movement in a 3D space, e.g. 3D mice, 6-DOF [six degrees of freedom] pointers using gyroscopes, accelerometers or tilt-sensors
-
- 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
- G06V40/30—Writer recognition; Reading and verifying signatures
- G06V40/37—Writer recognition; Reading and verifying signatures based only on signature signals such as velocity or pressure, e.g. dynamic signature recognition
- G06V40/382—Preprocessing; Feature extraction
-
- 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
- G06V40/30—Writer recognition; Reading and verifying signatures
- G06V40/37—Writer recognition; Reading and verifying signatures based only on signature signals such as velocity or pressure, e.g. dynamic signature recognition
- G06V40/394—Matching; Classification
Abstract
The present invention provides a kind of Handwritten Digit Recognition method based on gyroscope, by the gyroscope acquisition angle speed being worn in wrist；The method searched for the most backward using three threshold values intercepts effective gesture data；Selected characteristic amount also calculates；The characteristic quantity calculated is compared with 0 ~ 9 these 10 the digital gesture datas preset, it is judged which numeral what effectively gesture data section correspondence was write is.Utilize the inventive method, it is only necessary at weared on wrist gyroscope, there is wearable effective, low cost, without other sensor, advantage low in energy consumption.
Description
Technical field
The invention belongs to gesture identification field, be specifically related to a kind of Handwritten Digit Recognition method based on gyroscope.
Background technology
Along with the development of science and technology, public lives and amusement develop progressively towards intelligence, easily direction.Hands
Gesture is the commonly used operation of people and exchange way, uses sensing equipment and the mode identification technology of advanced person, carries out gesture
Identify, it is possible to achieve exchange and remote control.Particularly with handicapped crowd, utilizing gesture to carry out remote control becomes urgent
The demand cut.
Arabic numerals are that a class has the symbol determining specification stroke, are known by most people, it is easy to operate, conveniently
Memory, and most equipment is integrated with the control interface (such as: television set) being input with Arabic numerals.By handwritten numeral rail
Mark, as interaction gesture, is accustomed to the routine use of user and mental model is consistent.
Current gesture identification method there is also several obvious shortcoming: 1) gesture identification mainly has two kinds of technology: based on
The gesture identification of computer vision and sensor-based gesture identification.Gesture identification based on computer vision is by external disturbance
Relatively big (as illumination condition changes), and affected (such as actor's dressing or carry weight) by actor's profile, even movable
Scope is also limited by photographic head.2) various sensors are used for gesture identification, but still very difficult solution comfort level is poor, knowledge
The problems such as the high power consumption that other scope is limited, discrimination is low, expensive, number of sensors is many and causes.3) algorithm complex is high,
It is difficult under conditions of hardware is limited carry out computing.
Summary of the invention
The technical problem to be solved in the present invention is: providing a kind of Handwritten Digit Recognition method based on gyroscope, having can
Dress effective, low cost, without other sensor, advantage low in energy consumption.
The present invention solves that the technical scheme that above-mentioned technical problem is taked is: a kind of handwritten numeral based on gyroscope is known
Other method, it is characterised in that: it comprises the following steps:
S1, by the gyroscope acquisition angle speed being worn in wrist, the corresponding one group of angular velocity group of the most each sampled point,
Often group angular velocity group includes x-axis angular velocity, y-axis angular velocity and z-axis angular velocity, and x-axis is the horizontal right direction of human body, and z-axis is
Gravity direction, the horizontal forward direction of y-axis human body；Calculate the accumulated angle speed of each sampled point respectively；
S2, the intercepting of effective gesture data: use the method searched for the most backward of three threshold values, the value to accumulated angle speed
ω temporally intercepts, it is judged that the effectively Origin And Destination of gesture, intercepts some sections of effective gesture data sections；Every section of effective hands
Gesture data segment comprises several continuous sampling points and the angular velocity group of correspondence thereof and accumulated angle speed；
S3, the choosing of characteristic quantity:
In effective gesture data section, the angular velocity of each axle is formed by 3 class subsegments: positive section, negative section and zero section；Positive section
Definition be: in this data segment, the value of data is all higher than 0, and maximum is higher than first threshold；The definition of negative section is: at this
In data segment, the value of data is respectively less than 0, and minima is less than Second Threshold；The definition of zero section is: in this data segment, data
Value change between first threshold and Second Threshold；
Choose following 5 characteristic quantities: the number T of (1) effectively gesture data section sampled point1, it is from the effective gesture of this section
In data segment, the beginning sampled point of effective gesture is to the total number terminating sampled point；(2) x-axis angular velocity omegaxSegments T2, for
The positive section of x-axis angular velocity, negative section, the total number of zero section；(3) z-axis angular velocity omegazSegments T3, for z-axis angular velocity positive section,
Negative section, the total number of zero section；(4) the position T of zero section of x-axis angular velocity4, according to the position of zero section of appearance of x-axis angular velocity it is
First section, rear, interlude again without zero section, be identified respectively；(5) the position T of zero section of z-axis angular velocity5, according to z-axis
Headed by the position of zero section of appearance of angular velocity, section, rear, interlude are again without zero section, are identified respectively；
S4, the calculating of characteristic quantity:
According to the definition of each characteristic quantity, calculate respectively from effective gesture data section that S2 obtains；
S5, the identification of handwritten numeral:
The characteristic quantity of calculating is compared with these 10 the digital gesture datas of 0-9 preset, it is judged that effectively gesture number
According to section correspondence write be which numeral.
As stated above, described S5 specifically comprises the steps of
S5-1, the foundation of decision method:
Set up 6 kinds of decision methods, be represented sequentially as by the characteristic vector order calculated: W1=[T2,T4,T3,T5], W2=
[T3,T5,T2,T4], W3=[T2,T3,T5,T4], W4=[T2,T3,T4,T5], W5=[T3,T2,T4,T5], W6=[T3,T2,T5,
T4]；
S5-2, the dynamic selection of decision method:
These 10 digital gesture datas of default 0-9 are divided into 4 classes, respectively S1={ 1}, S2={ 0,6,7}, S3=
{ 2,3,9}, S4={ 4,5,8}；Decision method corresponding to this 4 class is respectively as follows: S1Corresponding aritrary decision method, S2Corresponding W1, S3Right
Answer W2, S4Corresponding W2；
Utilize T1Average, T1Median Q2、T1First quartile Q1、T1The 3rd quartile Q3These four genus
Property, carry out fuzzy clustering according to these four attributes, according to maximum membership grade principle, it is believed that object to be identified T1Relatively it is under the jurisdiction of S1-
S4In which kind of, select the decision method of such correspondence；
S5-3, according to selected decision method, calculate characteristic vector successively, with these 10 digital handss of default 0-9
Gesture data compare, it is judged which numeral what effectively gesture data section correspondence was write is.
As stated above, in described S2: the effectively judgement of the starting point of gesture: choose 3 threshold values from small to largePass throughCarry out the judgement that gesture motion starts, choose and be more thanFirst sampled point, sampled
Point Ab；Search is less than the most forwardFirst sampled point, obtain sampled point As；If AbWith AsThe two sampled point away from
From less than distance threshold DA, then the As starting point as effect gesture is judged；Otherwise, from AbStart search forward to be less thanFirst adopt
Sampling point, obtains time point Am, by AmIt is set to the starting point of effective gesture；
The effectively judgement of the terminal of gesture: choose 3 threshold values from small to largeChoose and meet and N thereafter
The meansigma methods of the accumulated angle speed sum of individual sampled point is less than threshold valueTime point Bs；Then from a BsSearch is more than bigger forward
Threshold valueFirst sampled point, obtain sampled point Bb；If BsWith BbDistance less than distance threshold DB, then by BsIt is set to have
The terminal of effect gesture；Otherwise, from BbStart search backward to be less thanFirst sampled point, obtain sampled point Bm, by BmIt is set to
The effectively terminal of gesture.
As stated above, the position T of zero section of described x-axis angular velocity4, when zero section of x-axis angular velocity occurs in first section,
T4It is designated 1；When zero section of x-axis angular velocity occurs in rear, T4It is designated-1；When zero section of x-axis angular velocity occurs in centre
Section, T4It is designated 0；When x-axis angular velocity does not has zero section, T4It is designated 2；
The position T of zero section of described z-axis angular velocity5The identification method position T with zero section of x-axis angular velocity4Identical.
As stated above, described T2Computational methods as follows:
1. finding approximation zero point: the symbol first with zero point both sides numerical value is different, the sampled point finding zero point existence is interval,
Take point that in two interval point that this sampled point is interval, absolute value is less as approximation zero point, by all approximation zero points according to sampling
Time sequencing sorts, and obtains approximating xero-sequence；
2. head and the tail zero point is removed: the head and the tail from approximation xero-sequence scan for respectively, remove at a distance of the approximation zero less than 5
Point, updates approximation xero-sequence；
3. find non-zero section interval zero section: the first interval of each adjacent approximation zero point composition of pairing approximation xero-sequence
Scan for, if there is the x-axis angular velocity absolute value of certain point more than threshold value Ωt, then this interval is that non-zero section is interval, makes this non-
Zero section of interval is [a, b]；The angular velocity absolute value of first section is continuously less than threshold value ΩtCount as n1, the angular velocity absolute value of tail end
Continuously less than threshold value ΩtCount as n2If, n1>n2And n1More than threshold value N, then zero section is positioned at first section, and waypoint is a+n1；
If n1<n2And n2More than threshold value N, then zero section is positioned at rear, and waypoint is b-n2；Otherwise this non-zero section interval does not exist zero section；
Add to waypoint approximate in xero-sequence, again update approximation xero-sequence；
4. removing continuous zero section of segregation section: if there being continuous print zero section, removing waypoint therebetween so that it is be combined into one section, and
Corresponding waypoint is deleted from approximation xero-sequence；
The number of the most final point in approximation xero-sequence subtracts 1, is T2；
Described T3Computational methods and T2Identical.
The invention have the benefit that
1, utilize the inventive method, it is only necessary at weared on wrist gyroscope, have wearable effective, low cost,
Sensor, advantage low in energy consumption without other.
2, further, carry out gesture identification by the method for decision tree, first set up decision method according to statistical analysis,
Then carry out fuzzy diagnosis according to fuzzy clustering method, select corresponding decision method to be identified, such that it is able to as far as possible
Use the shortest method that numeral gesture is identified, the beneficially raising of real-time, solve algorithm complexity and cause
Poor real and the serious problem of computer dependency.
Accompanying drawing explanation
Fig. 1 is the operation overview flow chart of one embodiment of the invention.
Fig. 2 is the schematic diagram of effective gesture intercepting of one embodiment of the invention.
Fig. 3 is the program flow diagram of effective gesture intercepting of one embodiment of the invention.
Fig. 4 is the classification schematic diagram of effective gesture section of one embodiment of the invention.
Fig. 5 is the feature extraction algorithm schematic diagram of one embodiment of the invention.
Fig. 6 is feature T of one embodiment of the invention2With T3Program flow diagram.
Detailed description of the invention
Below in conjunction with instantiation and accompanying drawing, the present invention will be further described.
The present invention improves a kind of Handwritten Digit Recognition method based on gyroscope, as it is shown in figure 1, comprise the following steps:
S1, by the gyroscope acquisition angle speed being worn in wrist, the corresponding one group of angular velocity group of the most each sampled point,
Often group angular velocity group includes x-axis angular velocity, y-axis angular velocity and z-axis angular velocity, and x-axis is the horizontal right direction of human body, and z-axis is
Gravity direction, the horizontal forward direction of y-axis human body；Calculate the accumulated angle speed of each sampled point respectively.
S2, the intercepting of effective gesture data: use the method searched for the most backward of three threshold values, the value to accumulated angle speed
ω temporally intercepts, it is judged that the effectively Origin And Destination of gesture, intercepts some sections of effective gesture data sections；Every section of effective hands
Gesture data segment comprises several continuous sampling points and the angular velocity group of correspondence thereof and accumulated angle speed.
Refinement, as shown in Figures 2 and 3, in described S2: the effectively judgement of the starting point of gesture: choose 3 from small to large
Threshold valuePass throughCarry out the judgement that gesture motion starts, choose and be more thanFirst sampled point,
To sampled point Ab；Search is less than the most forwardFirst sampled point, obtain sampled point As；If AbWith AsThe two is sampled
The distance of point is less than distance threshold DA, then the As starting point as effect gesture is judged；Otherwise, illustrate when gesture starts, the shake of staff
Error is relatively big, from AbStart search forward to be less thanFirst sampled point, obtain time point Am, by AmIt is set to effective gesture
Starting point；
The effectively judgement of the terminal of gesture: choose 3 threshold values from small to largeChoose and meet with thereafter
The meansigma methods of the accumulated angle speed sum of N number of sampled point is less than threshold valueTime point Bs；Then from a BsSearch is more than relatively forward
Big threshold valueFirst sampled point, obtain sampled point Bb；If BsWith BbDistance less than distance threshold DB, then by BsIt is set to
The effectively terminal of gesture；Otherwise, at the end of gesture is described, the jitter error of staff is relatively big, from BbStart search backward to be less than
First sampled point, obtain sampled point Bm, by BmIt is set to the terminal of effective gesture.
S3, the choosing of characteristic quantity:
As shown in Figure 4, in order to analytical data, the collection of gesture data is carried out.Find through analyzing, at effective gesture number
According in section, the angular velocity of each axle is formed by 3 class subsegments: positive section, negative section and zero section；The definition of positive section is: at this data segment
In, the value of data is all higher than 0, and maximum is higher than first threshold；The definition of negative section is: in this data segment, the value of data is equal
Less than 0, and minima is less than Second Threshold；The definition of zero section is: in this data segment, and the value of data is in first threshold and second
Change between threshold value.
Choose following 5 characteristic quantities: the number T of (1) effectively gesture data section sampled point1, it is from the effective gesture of this section
In data segment, the beginning sampled point of effective gesture is to the total number terminating sampled point；Owing to there is the regulation limiting gesture speed,
So the number of effective gesture data section sampled point can reflect the length of gesture to a certain extent.(2) x-axis angular velocity
ωxSegments T2, for positive section, negative section, the total number of zero section of x-axis angular velocity.(3) z-axis angular velocity omegazSegments T3, for
The positive section of z-axis angular velocity, negative section, the total number of zero section.(4) the position T of zero section of x-axis angular velocity4, according to x-axis angular velocity
Headed by zero section of position occurred, section, rear, interlude are again without zero section, are identified respectively.(5) zero section of z-axis angular velocity
Position T5, carry out respectively again without zero section according to section, rear, interlude headed by the position of zero section of appearance of z-axis angular velocity
Mark.
In the present embodiment, the position T of zero section of described x-axis angular velocity4, when zero section of x-axis angular velocity occurs in first section,
T4It is designated 1；When zero section of x-axis angular velocity occurs in rear, T4It is designated-1；When zero section of x-axis angular velocity occurs in centre
Section, T4It is designated 0；When x-axis angular velocity does not has zero section, T4It is designated 2.
The position T of zero section of described z-axis angular velocity5The identification method position T with zero section of x-axis angular velocity4Identical.
S4, the calculating of characteristic quantity:
According to the definition of each characteristic quantity, calculate respectively from effective gesture data section that S2 obtains.
Wherein, T1Calculating can be obtained by the valid data after intercepting.Note obtains its sampling number from valid data section
Purpose average length of time is t1。
As shown in Figure 5 and Figure 6, described T2Computational methods as follows:
1. approximation zero point is found: owing to sampled point is discrete, so not ensuring that there is a certain sampled point is zero point.
However, it is possible to the symbol first with zero point both sides numerical value is different, finds the sampled point interval that zero point exists, take this sampled point interval
Two interval point in the less point of absolute value as approximation zero point, all approximation zero points are sorted according to sample time order,
Obtain approximating xero-sequence.
Illustrating, approximating xero-sequence in Fig. 5 is: { 14,23,25,46,60,62}.For convenience of calculation, then increase
Gesture data head and the tail point enters approximation xero-sequence, obtains: { 1,14,23,25,46,60,62,63}.
2. head and the tail zero point is removed: during owing to intercepting effective gesture, error can be there is unavoidably, so the approximation zero point to head and the tail
Test.Head and the tail from approximation xero-sequence scan for respectively, remove at a distance of the approximation zero point less than 5, update approximation zero
Point sequence.In Figure 5, removing point 62 and point 63, the approximation xero-sequence after renewal is { 1,14,23,25,46,60}.
3. find non-zero section interval zero section: due to the certain point in zero section of the most above-mentioned sequence of waypoint, institute
Scan for the interval that each adjacent approximation zero point of first pairing approximation xero-sequence forms, if there is the x-axis angle of certain point
Speed absolute value is more than threshold value Ωt, then this interval is that non-zero section is interval, makes this non-zero section interval for [a, b]；The angular velocity of first section
Absolute value is continuously less than threshold value ΩtCount as n1, the angular velocity absolute value of tail end is continuously less than threshold value ΩtCount as n2If,
n1>n2And n1More than threshold value N, then zero section is positioned at first section, and waypoint is a+n1；If n1<n2And n2More than threshold value N, then zero
Section is positioned at rear, and waypoint is b-n2；Otherwise this non-zero section interval does not exist zero section；Add to waypoint approximate xero-sequence
In, again update approximation xero-sequence.
In Figure 5, N=5, Ω are takent=20, interval [25,46] are non-zero section, and the first half in this interval exists zero
Section, zero section of waypoint is 36, is added to approximate in xero-sequence, the approximation xero-sequence after again updating be 1,14,
23,25,36,46,60}。
4. continuous zero section of segregation section is removed: due to the uncertainty of zero section of symbol, so approximation zero point sequence derived above
Row may be separated continuous print zero section.If there being continuous print zero section, remove waypoint therebetween so that it is be combined into one section, and from
Approximation xero-sequence deletes corresponding waypoint.
In Figure 5, [14,23], [23,25], [25,36] are zero section, so removing waypoint 23 and 25, obtain approximating zero
Point sequence is { 1,14,36,46,60}.
The number of the most final point in approximation xero-sequence subtracts 1, is T2.In the present embodiment, T2For 5-1=4.
Described T3Computational methods and T2Identical.
Remember above-mentioned calculating T2With T3The average length of time expended altogether of each step be respectively t2And t3, t2≈t3。
The interval being made up of the adjacent sectional point 4. obtained, detects the requirement whether each interval meets zero section, the most each
Whether the absolute value of point is less than threshold value Ωt, it is possible to obtain the position of zero section, and according to the position of zero section, it is identified.Fig. 5
In, the 2nd section is zero section.Note is from T2With T3Computational methods in, obtain T4With T5The average time length that expends be respectively t4And t5, its
In, t4≈t5。
S5, the identification of handwritten numeral:
The characteristic quantity of calculating is compared with these 10 the digital gesture datas of 0-9 preset, it is judged that effectively gesture number
According to section correspondence write be which numeral.
Preferably, in order to by dynamic decision Method, skip the calculating of redundancy feature amount, make as far as possible
Being identified numeral gesture by the shortest method, described S5 specifically comprises the steps of
S5-1, the foundation of decision method:
Set up 6 kinds of decision methods, be represented sequentially as by the characteristic vector order calculated: W1=[T2,T4,T3,T5], W2=
[T3,T5,T2,T4], W3=[T2,T3,T5,T4], W4=[T2,T3,T4,T5], W5=[T3,T2,T4,T5], W6=[T3,T2,T5,
T4]。
S5-2, the dynamic selection of decision method:
These 10 digital gesture datas of default 0-9 are divided into 4 classes, respectively S1={ 1}, S2={ 0,6,7}, S3=
{ 2,3,9}, S4={ 4,5,8}；Decision method corresponding to this 4 class is respectively as follows: S1Corresponding aritrary decision method, S2Corresponding W1, S3Right
Answer W2, S4Corresponding W2；
Utilize T1Average, T1Median Q2、T1First quartile Q1、T1The 3rd quartile Q3These four genus
Property, carry out fuzzy clustering according to these four attributes, according to maximum membership grade principle, it is believed that object to be identified T1Relatively it is under the jurisdiction of S1-
S4In which kind of, select the decision method of such correspondence；
S5-3, according to selected decision method, calculate characteristic vector successively, with these 10 digital handss of default 0-9
Gesture data compare, it is judged which numeral what effectively gesture data section correspondence was write is.
Above-mentioned decision method to set up principle as follows:
Characteristic vector T=[T is constituted by above-mentioned 5 characteristic quantities1,T2,T3,T4,T5].But, statistical analysis finds, and
The digital gesture of not all needs these five characteristic quantities just can identify.Note can identify that the characteristic vector of this numeral is the soonest
Tf, required time is tf.Following table has counted this ten digital T2,T3,T4,T5,TfValue.
Due to T2With T4、T3With T5Dependent restriction, carry out Handwritten Digit Recognition and be divided into six kinds of decision methods, be designated as Wi
(i=1～6).Respectively with calculate characteristic vector order be expressed as [T2,T4,T3,T5],[T3,T5,T2,T4],[T2,T3,T5,
T4], [T2,T3,T4,T5],[T3,T2,T4,T5],[T3,T2,T5,T4].When note uses the cost of every kind of method each numeral of identification
Between be twi(i=1～6), remembers twiWith tfDifference be Δ twi(i=1～6).
Every kind of method has some numerals during identifying, calculate unnecessary characteristic quantity, cause real-time poor
And energy consumption increases.So the method that dynamically selection can be used, in advance gesture data is differentiated, obtain approximate range
After, then determine the characteristic quantity of this calculating.
Utilize T1Average, median Q2, first quartile Q1, the 3rd quartile Q3These four attributes, according to this four
Individual attribute carries out fuzzy clustering.The gesture data that 10 are digital is divided into 4 classes, be respectively 1}, 0,6,7}, and 2,3,9}, 4,
5,8} uses S respectively1～S4Represent.Can be by T1By calculating affiliated all kinds of degree of membership, judge the class of done gesture faintly
Not.According to maximum membership grade principle, it is believed that object to be identified T1Relatively it is under the jurisdiction of class Sk, select such decision method Wk.So
After, determine the decision method of every class.Δ t with every apoplexy due to endogenous wind numeralw1With minimum principle, S1Can be selected for any one, S2Select
W1, S3Select W2,S4Select W2.By being dynamically selected decision method, the shortest method logarithm can be used as much as possible
Word gesture is identified, beneficially the raising of real-time.
Concrete, Fig. 5 is the data and curves figure of gesture " 4 ".The data length T of gesture1It is 64, former according to maximum membership degree
Then, this gesture is assigned to S4Class, then by S4Characteristic quantity computation sequence W of class2=[T3,T5,T2,T4] calculate.According to this
Method, identifies gesture " 4 ", it is only necessary to calculate characteristic quantity T3And T5, save calculating T2And T4Elapsed time.
Above example is merely to illustrate design philosophy and the feature of the present invention, its object is to make the technology in this area
Personnel will appreciate that present disclosure and implement according to this, and protection scope of the present invention is not limited to above-described embodiment.So, all depend on
The equivalent variations made according to disclosed principle, mentality of designing or modification, all within protection scope of the present invention.
Claims (5)
1. a Handwritten Digit Recognition method based on gyroscope, it is characterised in that: it comprises the following steps:
S1, by the gyroscope acquisition angle speed being worn in wrist, the most each sampled point corresponding one group of angular velocity group, often group
Angular velocity group includes x-axis angular velocity, y-axis angular velocity and z-axis angular velocity, and x-axis is the horizontal right direction of human body, and z-axis is gravity
Direction, the horizontal forward direction of y-axis human body；Calculate the accumulated angle speed of each sampled point respectively；
S2, the intercepting of effective gesture data: use the method searched for the most backward of three threshold values, value ω of accumulated angle speed is pressed
Time intercepts, it is judged that the effectively Origin And Destination of gesture, intercepts some sections of effective gesture data sections；Every section of effective gesture number
Several continuous sampling points and the angular velocity group of correspondence thereof and accumulated angle speed is comprised according to section；
S3, the choosing of characteristic quantity:
In effective gesture data section, the angular velocity of each axle is formed by 3 class subsegments: positive section, negative section and zero section；Determining of positive section
Justice is: in this data segment, and the value of data is all higher than 0, and maximum is higher than first threshold；The definition of negative section is: in these data
Duan Zhong, the value of data is respectively less than 0, and minima is less than Second Threshold；The definition of zero section is: in this data segment, the value of data
Change between first threshold and Second Threshold；
Choose following 5 characteristic quantities: the number T of (1) effectively gesture data section sampled point1, it is from this section effective gesture data section
In, the beginning sampled point of effective gesture is to the total number terminating sampled point；(2) x-axis angular velocity omegaxSegments T2, for x-axis angle
The positive section of speed, negative section, the total number of zero section；(3) z-axis angular velocity omegazSegments T3, for the positive section of z-axis angular velocity, negative section,
The total number of zero section；(4) the position T of zero section of x-axis angular velocity4, according to section headed by the position of zero section of x-axis angular velocity appearance,
Rear, interlude, again without zero section, are identified respectively；(5) the position T of zero section of z-axis angular velocity5, according to z-axis angular velocity
Zero section appearance position headed by section, rear, interlude again without zero section, be identified respectively；
S4, the calculating of characteristic quantity:
According to the definition of each characteristic quantity, calculate respectively from effective gesture data section that S2 obtains；
S5, the identification of handwritten numeral:
The characteristic quantity of calculating is compared with these 10 the digital gesture datas of 0-9 preset, it is judged that effectively gesture data section
Which numeral what correspondence was write is.
Handwritten Digit Recognition method based on gyroscope the most according to claim 1, it is characterised in that: described S5 is concrete
Comprise the steps of
S5-1, the foundation of decision method:
Set up 6 kinds of decision methods, be represented sequentially as by the characteristic vector order calculated: W1=[T2,T4,T3,T5], W2=[T3,T5,
T2,T4], W3=[T2,T3,T5,T4], W4=[T2,T3,T4,T5], W5=[T3,T2,T4,T5], W6=[T3,T2,T5,T4]；
S5-2, the dynamic selection of decision method:
These 10 digital gesture datas of default 0-9 are divided into 4 classes, respectively S1={ 1}, S2={ 0,6,7}, S3=2,3,
9}, S4={ 4,5,8}；Decision method corresponding to this 4 class is respectively as follows: S1Corresponding aritrary decision method, S2Corresponding W1, S3Corresponding W2,
S4Corresponding W2；
Utilize T1Average, T1Median Q2、T1First quartile Q1、T1The 3rd quartile Q3These four attributes, depend on
Fuzzy clustering is carried out, according to maximum membership grade principle, it is believed that object to be identified T according to these four attributes1Relatively it is under the jurisdiction of S1-S4In
Which kind of, select the decision method of such correspondence；
S5-3, according to selected decision method, calculate characteristic vector successively, with these 10 digital gesture numbers of default 0-9
According to comparing, it is judged which numeral what effectively gesture data section correspondence was write is.
Handwritten Digit Recognition method based on gyroscope the most according to claim 1 and 2, it is characterised in that: described S2
In: the effectively judgement of the starting point of gesture: choose 3 threshold values from small to largePass throughCarry out gesture fortune
The dynamic judgement started, chooses and is more thanFirst sampled point, obtain sampled point Ab；Search is less than the most forwardFirst
Individual sampled point, obtains sampled point As；If AbWith AsThe distance of the two sampled point is less than distance threshold DA, then A is judgedsFor effect hands
The starting point of gesture；Otherwise, from AbStart search forward to be less thanFirst sampled point, obtain time point Am, by AmIt is set to effectively
The starting point of gesture；
The effectively judgement of the terminal of gesture: choose 3 threshold values from small to largeChoose and meet with the most N number of
The meansigma methods of the accumulated angle speed sum of sampled point is less than threshold valueTime point Bs；Then from a BsSearch is more than bigger threshold forward
ValueFirst sampled point, obtain sampled point Bb；If BsWith BbDistance less than distance threshold DB, then by BsIt is set to effectively
The terminal of gesture；Otherwise, from BbStart search backward to be less thanFirst sampled point, obtain sampled point Bm, by BmIt is set to have
The terminal of effect gesture.
Handwritten Digit Recognition method based on gyroscope the most according to claim 1 and 2, it is characterised in that: described x-axis
The position T of zero section of angular velocity4, when zero section of x-axis angular velocity occurs in first section, T4It is designated 1；When zero section of x-axis angular velocity goes out
Rear, T now4It is designated-1；When zero section of x-axis angular velocity occurs in interlude, T4It is designated 0；When x-axis angular velocity does not has zero
Section, T4It is designated 2；
The position T of zero section of described z-axis angular velocity5The identification method position T with zero section of x-axis angular velocity4Identical.
Handwritten Digit Recognition method based on gyroscope the most according to claim 1 and 2, it is characterised in that: described T2's
Computational methods are as follows:
1. finding approximation zero point: the symbol first with zero point both sides numerical value is different, the sampled point finding zero point existence is interval, takes this
The point that in two interval point that sampled point is interval, absolute value is less is as approximation zero point, by all approximation zero points according to the sampling time
Order sequence, obtains approximating xero-sequence；
2. remove head and the tail zero point: the head and the tail from approximation xero-sequence scan for respectively, remove at a distance of the approximation zero point less than 5,
Update approximation xero-sequence；
3. find non-zero section interval zero section: the first interval of each adjacent approximation zero point composition of pairing approximation xero-sequence is carried out
, if there is the x-axis angular velocity absolute value of certain point more than threshold value Ω in searcht, then this interval is that non-zero section is interval, makes this non-zero section
Interval is [a, b]；The angular velocity absolute value of first section is continuously less than threshold value ΩtCount as n1, the angular velocity absolute value of tail end is continuous
Less than threshold value ΩtCount as n2If, n1>n2And n1More than threshold value N, then zero section is positioned at first section, and waypoint is a+n1；If n1
<n2And n2More than threshold value N, then zero section is positioned at rear, and waypoint is b-n2；Otherwise this non-zero section interval does not exist zero section；To divide
Section point adds to approximate in xero-sequence, again updates approximation xero-sequence；
4. removing continuous zero section of segregation section: if there being continuous print zero section, removing waypoint therebetween so that it is be combined into one section, and from closely
Like xero-sequence is deleted corresponding waypoint；
The number of the most final point in approximation xero-sequence subtracts 1, is T2；
Described T3Computational methods and T2Identical.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610524482.XA CN106127188B (en) | 2016-07-05 | 2016-07-05 | A kind of Handwritten Digit Recognition method based on gyroscope |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610524482.XA CN106127188B (en) | 2016-07-05 | 2016-07-05 | A kind of Handwritten Digit Recognition method based on gyroscope |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106127188A true CN106127188A (en) | 2016-11-16 |
| CN106127188B CN106127188B (en) | 2019-05-17 |
Family
ID=57469500
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610524482.XA Active CN106127188B (en) | 2016-07-05 | 2016-07-05 | A kind of Handwritten Digit Recognition method based on gyroscope |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106127188B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106950927A (en) * | 2017-02-17 | 2017-07-14 | 深圳大学 | A kind of method and Intelligent worn device of control smart home |
| CN107153780A (en) * | 2017-05-05 | 2017-09-12 | 西安交通大学苏州研究院 | The writing behavioural characteristic authentication method of electronic equipment is dressed based on wrist |
| CN110569800A (en) * | 2019-09-10 | 2019-12-13 | 武汉大学 | detection method of handwriting signal |
| CN117075742A (en) * | 2023-10-17 | 2023-11-17 | 深圳市魔样科技有限公司 | Intelligent finger ring control method based on gesture recognition |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102945362A (en) * | 2012-10-18 | 2013-02-27 | 中国科学院计算技术研究所 | Isomerous data fusion based coordinated gesture recognition method and system of sensor |
| CN103513788A (en) * | 2013-09-13 | 2014-01-15 | 广东欧珀移动通信有限公司 | Gesture recognition method and system based on gyroscope sensor and mobile terminal |
| US9129400B1 (en) * | 2011-09-23 | 2015-09-08 | Amazon Technologies, Inc. | Movement prediction for image capture |
| CN105138133A (en) * | 2015-09-14 | 2015-12-09 | 李玮琛 | Biological signal gesture recognition device and method |
-
2016
- 2016-07-05 CN CN201610524482.XA patent/CN106127188B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9129400B1 (en) * | 2011-09-23 | 2015-09-08 | Amazon Technologies, Inc. | Movement prediction for image capture |
| CN102945362A (en) * | 2012-10-18 | 2013-02-27 | 中国科学院计算技术研究所 | Isomerous data fusion based coordinated gesture recognition method and system of sensor |
| CN103513788A (en) * | 2013-09-13 | 2014-01-15 | 广东欧珀移动通信有限公司 | Gesture recognition method and system based on gyroscope sensor and mobile terminal |
| CN105138133A (en) * | 2015-09-14 | 2015-12-09 | 李玮琛 | Biological signal gesture recognition device and method |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106950927A (en) * | 2017-02-17 | 2017-07-14 | 深圳大学 | A kind of method and Intelligent worn device of control smart home |
| CN106950927B (en) * | 2017-02-17 | 2019-05-17 | 深圳大学 | A kind of method and intelligent wearable device controlling smart home |
| CN107153780A (en) * | 2017-05-05 | 2017-09-12 | 西安交通大学苏州研究院 | The writing behavioural characteristic authentication method of electronic equipment is dressed based on wrist |
| CN110569800A (en) * | 2019-09-10 | 2019-12-13 | 武汉大学 | detection method of handwriting signal |
| CN110569800B (en) * | 2019-09-10 | 2022-04-12 | 武汉大学 | Detection method of handwriting signal |
| CN117075742A (en) * | 2023-10-17 | 2023-11-17 | 深圳市魔样科技有限公司 | Intelligent finger ring control method based on gesture recognition |
| CN117075742B (en) * | 2023-10-17 | 2024-01-30 | 深圳市魔样科技有限公司 | Intelligent finger ring control method based on gesture recognition |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106127188B (en) | 2019-05-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhu et al. | Efficient human activity recognition solving the confusing activities via deep ensemble learning | |
| CN106127188B (en) | A kind of Handwritten Digit Recognition method based on gyroscope | |
| CN106096662B (en) | Human motion state identification based on acceleration transducer | |
| CN102640085B (en) | For identifying the system and method for gesture | |
| CN105005769B (en) | A kind of sign Language Recognition Method based on depth information | |
| CN103824051B (en) | Local region matching-based face search method | |
| CN110287825B (en) | Tumble action detection method based on key skeleton point trajectory analysis | |
| CN108647663B (en) | Human body posture estimation method based on deep learning and multi-level graph structure model | |
| CN105718868A (en) | Face detection system and method for multi-pose faces | |
| CN106407874A (en) | Handwriting recognition method based on handwriting coordinate sequence | |
| Jensen et al. | Classification of kinematic swimming data with emphasis on resource consumption | |
| CN103246891A (en) | Chinese sign language recognition method based on kinect | |
| CN105877757A (en) | Multi-sensor integrated human motion posture capturing and recognizing device | |
| CN111709310A (en) | Gesture tracking and recognition method based on deep learning | |
| CN110110650A (en) | Face identification method in pedestrian | |
| CN103345626A (en) | Intelligent wheelchair static gesture identification method | |
| CN105335711B (en) | Fingertip Detection under a kind of complex environment | |
| Shi et al. | Moving target detection algorithm in image sequences based on edge detection and frame difference | |
| CN111274998A (en) | Parkinson's disease finger knocking action identification method and system, storage medium and terminal | |
| CN110472506A (en) | A kind of gesture identification method based on support vector machines and Neural Network Optimization | |
| CN108921011A (en) | A kind of dynamic hand gesture recognition system and method based on hidden Markov model | |
| CN116226691A (en) | Intelligent finger ring data processing method for gesture sensing | |
| CN104586402B (en) | A kind of feature extracting method of physical activity | |
| CN108717548B (en) | Behavior recognition model updating method and system for dynamic increase of sensors | |
| Mekruksavanich et al. | Sensor-based complex human activity recognition from smartwatch data using hybrid deep learning network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |