CN108596161A - A kind of portable photographing glasses and photographic method - Google Patents
A kind of portable photographing glasses and photographic method Download PDFInfo
- Publication number
- CN108596161A CN108596161A CN201810667833.1A CN201810667833A CN108596161A CN 108596161 A CN108596161 A CN 108596161A CN 201810667833 A CN201810667833 A CN 201810667833A CN 108596161 A CN108596161 A CN 108596161A
- Authority
- CN
- China
- Prior art keywords
- eye
- image
- eyes
- function
- motion analysis
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/18—Eye characteristics, e.g. of the iris
- G06V40/193—Preprocessing; Feature extraction
-
- G—PHYSICS
- G02—OPTICS
- G02C—SPECTACLES; SUNGLASSES OR GOGGLES INSOFAR AS THEY HAVE THE SAME FEATURES AS SPECTACLES; CONTACT LENSES
- G02C11/00—Non-optical adjuncts; Attachment thereof
- G02C11/10—Electronic devices other than hearing aids
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/18—Eye characteristics, e.g. of the iris
- G06V40/197—Matching; Classification
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- Ophthalmology & Optometry (AREA)
- General Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Acoustics & Sound (AREA)
- Otolaryngology (AREA)
- Optics & Photonics (AREA)
- Image Processing (AREA)
Abstract
The present invention relates to a kind of portable photographing glasses and photographic method, including mirror holder, eyeglass and Jing Quan, motion analysis instrument is installed on Jing Quan；Motion analysis instrument includes：Eye detection module, motion analysis module and action recognition module；Eye detection module acquires and rejects class eye information by eye areas precise extracting algorithm after eye and neighbouring photographic intelligence, obtains pure eye pictorial information, then extract to eye motion information characteristics, obtain eyes widen the view, the black white image of eye closing action message；Motion analysis module, be divided into using grader widen the view, two classes of closing one's eyes are per frame image category, judge whether eyes have blink phenomenon, to obtain blink judging result；Action recognition module carries out function command identifying processing, by the eye motion message form for matching preset function instruction, judge the representative instruction meaning of blink action, complete identification blink and act included function command information, output takes pictures, record starts or End of Tape function command.
Description
Technical field
The present invention relates to a kind of portable photographing glasses and photographic methods.
Background technology
People are frequently run onto some accidents in life, have had, and also have certainly bad, these are all not levy
Million.For example you catch sight of an a lovely cat and come to sell in front of you and sprout or a butterfly flies to you on hand, again
Or it catches sight of passerby and is plundered.But when you want to dig out under cameras record at the moment, this moment does not deposit again
It is sorry to bring you more, and kitten runs, and butterfly flies away, and the face of robber is not also remembered well.Very much
It is more because wait for for a long time why photographer, which can take a splendid moment,.
For this problem, there are the related candid photograph glasses of related technology reports in foreign countries.There are mainly two types of at present：One is
Spectacles sunglasses.Spectacles sunglasses was issued on the 23rd in September in 2016.Possess 115 ° of wide-angle lens, can clap
Take the photograph the video of the first person.Click a button being furnished on its mirror holder, so that it may to record the one section long video up to 10 seconds.Even
Under three, then no longer than 30 seconds videos can be recorded.Official indicates to say that the continuation of the journey of Spectacles sunglasses can be grown
Up to 1 day.The video of shooting directly directly can be shared Snapchat using upper by Spectacles sunglasses, and carry out the later stage
Editor, can also be by wifi or Bluetooth transmission to mobile phone.Meanwhile these videos can also be read on Snapchat after i.e.
It burns.
Advantage：Easy to operate understandable, video, which can reach, to be easy to edit and be stored in mobile phone, and 115 degree are that shooting angle more connects
It is bordering on the picture angle that people sees the world.The video effect of glasses shooting is spherical shape, can be with when being watched in Snapchat app
Viewing angle is arbitrarily replaced, under gravity sensing effect, the main perspective of viewing can remain unchanged.
Disadvantage：Video can only be recorded by pushing button, and cannot be taken pictures.Stud bump influences beautiful.
Another kind is：Blincam.One newly developed company from Japan successfully has developed a entitled " Blincam "
Wearable concept camera apparatus.Although its operation principle and Google glass are quite similar, the difference is that it can be independent
It is worn on the common spectacles of any brand, and wearer then can complete to take pictures by way of blink.According to research and development of products
The introduction of team, final volume production version will possess a mating application program, and support photo being directly transferred to hand
Machine.It is connect in addition, Blincam is also equipped with a detection sensor for having exclusive patented technology, LED light, microUSB
Mouth, " HD CMOS " imaging sensor and 32GB memory spaces and bluetooth.
Advantage：Wearer can complete to take pictures by way of blink, very easy；
Disadvantage：The technology uses blinking sensor, the muscle signal that principle generates when being using blink to judge that glasses are
No blink, the disadvantage is that eyes need Large Amplitude Motion, this is extremely unnatural when can cause to use, extremely uncomfortable, in addition, its independence
The distribution of carrier's eyewear can be made uneven with the independent design of glasses, facial structure can be adversely affected for a long time；
Glasses and instrument are not merged, it is inconvenient to wear, and shape is exaggerated very much, not hidden enough, can not achieve the functions such as recording.
Invention content
The technology of the present invention solves the problems, such as：Blink movement, action can only be identified in the prior art by eye muscle signal
Amplitude must be big, unnatural when influencing beauty, and using, a kind of portable photographing glasses of this patent offer and photographic method,
Utilize real-time eyes image treatment technology so that shooting, recording and manipulation more naturally, more easy, passes through function
Set significantly shorten the operating time, photographed fine moment at the first time, and simple and practicable.
A kind of portable photographing glasses, including mirror holder and eyeglass, the mirror holder include for installing the Jing Quan of eyeglass, being used for
The connection left and right crossbeam of two Jing Quan, nose support, positioned at the pile crown of Jing Quan and the connecting portion of pitch-angle, for hooking the position of ear
Temple；
Its feature is：Motion analysis instrument is installed on the Jing Quan；Motion analysis instrument is located at both sides mirror circle crossbeam
On；The motion analysis instrument includes：Eye detection module, motion analysis module and action recognition module；
The technology of the present invention solution：A kind of portable photographing glasses and photographic method, eye photographing operation of blinking are realized
Mainly completed by motion analysis instrument, it according to eye detection module, by eye motion detect camera acquire eye and near
Photographic intelligence after by eye areas precise extracting algorithm reject class eye information, realize the eye areas of Primary Location, obtain
Pure eye pictorial information (skin for including eyes and ocular vicinity), the class eye information includes eyebrow and nose, then right
The pure eye pictorial information carries out local auto-adaptive binary conversion treatment, obtain eyes widen the view, eye closing action message it is black
White image；Later, by motion analysis module, the eyes that eye motion detection module exports is widened the view using grader, are closed one's eyes
The black white image of action message be divided into widen the view, two classes of closing one's eyes are per frame image category, then carry out blink action recognition processing, judge
Whether eyes have blink phenomenon, to obtain blink judging result；Finally, action recognition module is by matching preset function instruction
Eye motion message form, judge the representative instruction meaning of blink action, being finally completed identification blink action is included
Function command information, the instruction of output camera function, recording start function command or End of Tape function command.
Wherein eye detection module algorithm is realized as follows：
(1) first, the eye of M*N sizes that eye camera takes and neighbouring photographic intelligence are longitudinally considered as Y-axis
J rows are denoted as, is laterally considered as X-axis and is denoted as i row, eye and neighbouring photographic intelligence are labeled as I (i, j), according to formulaMake the discrete projection function of picture Y-direction；Wherein Y (j) is eye and neighbouring photo
Information I (i, j) Y-axis is projected after function, M, N be M*N sizes eye and neighbouring photographic intelligence I (i, j) length
And width；
(2) then, the local maximum y of the discrete projection function is calculatedmaxY at this timemaxCorresponding abscissa j ', sentences
Break and pure eye pictorial information；
(3) finally, include outside eyebrow and the pure eye pictorial information of two kinds of nose using following function formula (1) rejecting
The class eye information that may influence monitoring, realizes the eye areas of Primary Location, obtains pure eye pictorial information Ir(a,b)；
Wherein Ir(a, b) is that former eye eye and neighbouring photographic intelligence I (i, j) pass through Ir(i-2, j-j ') functional transformation
Image afterwards, a=i-2, b=j-j ' be through function (1) transformation after pure eye pictorial information length and width.
The eyes feature extraction is to utilize complete local binary method (Completed Local Binary
Pattern it) carries out.Since the threshold value that illumination condition influences image binaryzation serious, simply to give image binaryzation can shadow
Widen the view in sound image, the acquisition of eye closing action message, needs that Adaptive Thresholding is utilized to carry out binary conversion treatment；
The grey level histogram for seeking needing using image of adaptive threshold, grey level histogram reflect an image pixel point
The statistical form of cloth, abscissa represent the type of the image pixel of gray scale, and ordinate represents each color value in the picture
Sum of all pixels or account for the percentage of all pixels number；
Wherein the complete local binarization of image and the grey level histogram using image seek adaptive threshold and obtain specific mistake
Cheng Wei：
(1) first, if IrThe tonal gradation of the image grey level histogram of (a, b) image is [0, L-1], according to formula (2)
Calculate the tonal gradation probability density function of I (a, b) image：
Wherein, rkIt is kth grade gray scale；nkIt is IrTonal gradation is r in (a, b)kNumber of pixels；P and Q is image I respectivelyr
The width and height of (a, b)；
(2) after, I is calculatedrThe distribution of the grey level histogram of (a, b) image it is expected, as shown in formula (3)：
(3) then, I is estimated according to the expectation E (r) acquiredrWhether some pixel belongs to skin gray scale in (a, b) image
Range or eyes tonal range are realized by formula (4) to IrThe skin of ocular vicinity and eyes is further in (a, b) image
Enhancing classification：
Wherein, r IrThe tonal gradation of each pixel of (a, b) image is as input；S is output gray level grade；Coefficient c
It is dimension scale constant for c, finds that when c values are 1.3-1.5 be optimal value according to practical debugging, other expressions are not belonging to " r<=
Other gray values of E (r) "；
(4) then, by enhancing classify after ocular vicinity skin and eyes image grey level histogram distribution more
It polarises, it is x that being located at, which enhances the range of the corresponding abscissa of sorted eye information grey level histogram,1~x2, corresponding grey scale is straight
The maximum one piece of range of ordinate value in square figure, it is x~g (x), maximum value g (r wherein in ordinate to take grey level histogramh)
Corresponding abscissa value is rh, then with rhMidpoint determines x according to formula (5) to abscissa both sides1、x2Value, wherein x2-x1Just
It is adaptive threshold T；
Wherein i=h-1, h-2 ,~, 1；J=h+1, h+2 ,~, n；A is termination condition, command deployment number and precision；
(5) finally, binary conversion treatment is carried out to image Ir (a, b) by Adaptive Thresholding, i.e.,
It obtains, wherein x is any point Ir (a, b) in image, and 1 and 0 respectively represents black and white, and S (x) is image local binaryzation
Working function, it is compared each pixel in image Ir (a, b), and pattern colour coloured picture is become to the binaryzation of black and white format
Image.
Image by this binaryzation is black white image, widening the view containing eyes, eye closing information, as next step action point
Analyse the basis of processing.
Wherein motion analysis module is realized as follows：
By black white image after handling adaptive threshold local binarization therein, using each frame image as a square
Battle array count the first moment of black (1) inside this image array, the accounting of white (0) two class pixel and image array as widen the view,
The action message of eye closing is widened the view, the input of the action message closed one's eyes as support vector machines (SVM) algorithm, to each frame image
Action message carry out Classification and Identification processing, sort out and widen the view or eye closing action state.If being eye closing eye shape state by laststate,
NextState is state of widening the view, then judges to be exactly the action once blinked at this time, realize the analysis of eye motion.
Wherein action recognition module carries out function command identifying processing according to the blink judging result of motion analysis module,
By matching the eye motion message form of preset function instruction, judges the representative instruction meaning of blink action, be finally completed
Identify that blink acts included function command information, the instruction of output camera function, recording start function command or End of Tape
Function command, specific workflow are as follows：
(1) it when motion analysis module captures the action of the first blink, counts once, is captured in 0.1s-0.5s
When blinking to continuous second, count second；
(2) if motion analysis instrument captures third time signal of blinking in 0.1s-0.5s, third time is counted, then is moved
The instrument that performs an analysis will action recognition be at this time camera function, count zero, output binary number 01 is instructed as camera function；
(3) if after counting second, motion analysis module does not capture blink instruction after 1s, then motion analysis instrument
It is that recording starts function command by action recognition at this time, output binary number 10 is as recording sign on, in the mistake of recording
Cheng Zhong, output remains output 10 state of binary number, and counts zero；
(4) it if output is under the status condition that 10 state of binary number is that recording starts, has again identified that out continuous
Second blink when, by action recognition at this time be End of Tape function command, output binary number be 11, count zero；
With this, fixed eye motion is determined as taking pictures, starts recording, end sound-recording function instruction.
Wherein action recognition module carries out function command identifying processing according to the blink judging result of motion analysis module,
By matching the eye motion message form of preset function instruction, judges the representative instruction meaning of blink action, be finally completed
Identify that blink acts included function command information, the instruction of output camera function, recording start function command or End of Tape
Function command, specific workflow are as follows：
(1) it when motion analysis module captures the action of the first blink, counts once, occurs capturing in 0.2s continuous
Second of blink when, count second；
(2) if motion analysis instrument captures third time signal of blinking in 0.2s, third time, then motion analysis are counted
Instrument will action recognition be at this time camera function, count zero, output binary number 01 is instructed as camera function；
(3) if after counting second, motion analysis module does not capture blink instruction after 1s, then motion analysis instrument
It is that recording starts function command by action recognition at this time, output binary number 10 is as recording sign on, in the mistake of recording
Cheng Zhong, output remains 10 state of binary number, and counts zero；
(4) it if output is under the status condition that 10 state of binary number is that recording starts, has again identified that out continuous
Second of blink when, be End of Tape function command by action recognition at this time, export as binary number 11, count zero；
With this, fixed eye motion is determined as taking pictures, starts recording, end sound-recording function instruction.
The present invention can be operated by touch key-press form, and in side, temple is equipped with membrane keyboard, glasses take pictures and
The startup of sound-recording function can be realized there is pressure sensor on membrane keyboard, produced after of short duration touch by touch key-press form
Raw pressure, changes capacitance size, then will produce a kind of camera function command signal, and triggering camera is taken pictures, and is pressed for a long time super
Another recorded audio signals are generated after spending 1 second, can trigger recording sign on or End of Tape instruction.
The temple of the present invention, side is equipped with camera of taking pictures, LED flash and phonographic recorder；The temple of the other side is then
It is data processor；When data processor receives the camera function instruction that motion analysis instrument is sent out, data processor is rapid
Start camera acquisition present image of taking pictures, and detect color information in the present image information RGB that camera captures that takes pictures,
When color information is inadequate, LED flash is opened, data processor preserves camera the image collected letter of currently taking pictures later
Breath, otherwise, data processor directly preserves current the image collected；When data processor receives what motion analysis instrument was sent out
When recording starts function command, data processor starts rapidly phonographic recorder acquisition current sound information and starts to preserve, and works as data
When processor receives the End of Tape function command that motion analysis instrument is sent out, data processor terminates rapidly the acquisition of phonographic recorder
And it completes to preserve.
The advantages of the present invention over the prior art are that：Currently, being blinked without use in the patent of domestic applications or mandate
The taking pictures of eye control, glasses of record, only the Blinacam glasses of Japan can realize blink camera function at present, but it sets
Meter and the technical solution used have too many disadvantage, and Blinacam is the technical solution adopted is that using blinking sensor, principle is
The muscle signal generated when using blink judges whether glasses blink, the disadvantage is that eyes need Large Amplitude Motion, this can cause
It is extremely unnatural when use, it is extremely uncomfortable, in addition, its independent independent design with glasses can make the distribution of carrier's eyewear uneven
It is even, facial structure can be adversely affected for a long time；The present invention in contrast, uses camera as sensor, uses
Image processing techniques, can precise positioning eye position, accurate judgement eye motion, in real time acquire eyes state, sentence in real time
It is disconnected whether to generate function command signal, user using when according to the signal of blinking of oneself as starting the signal taken pictures and recorded,
Most can naturally start take pictures, sound-recording function, in addition to this, the design of the symmetrical structure of this product can effectively eliminate unilateral
Glasses and hardware integration are carried and are more convenient nature by the asymmetric influence of weight.
Description of the drawings
Fig. 1 is the structure chart of the present invention, wherein 1 be Jing Quan, 2 being crossbeam, 3 being temple, 4 be eye motion analyzer, and 5 are
LED flash, 6 is take pictures camera, and 8 be phonographic recorder, and 7 be membrane keyboard, 9 are battery, 10 are reservoir, 11 are data processing
Device, 12 be wireless signal transmitter；
Fig. 2 is the mating mirror case figure of the present invention；
Fig. 3 is prior art Spectacles sunglasses；
Fig. 4 is prior art Blincam spectacles capable of photographing corollary equipments；
Fig. 5 is structure of the invention block diagram；
Fig. 6 is work flow diagram of the present invention.
Specific implementation mode
The following describes the present invention in detail with reference to the accompanying drawings and embodiments.
As shown in Figure 1, the present invention includes mirror holder and eyeglass, mirror holder includes mirror circle 1, crossbeam 2, temple 3, at bridge of the nose position
There is eye motion analyzer 4 in bridge, forward left side temple is LED flash 5 and camera 6 and phonographic recorder 8, is tactile rearward
Touch key 7, touch to take pictures, long touch starts for recording, then vice-minister touches as End of Tape, the temple of right be then the battery 9 of glasses,
Reservoir 10, data processor 11 and wireless signal transmitter 12.
As shown in Fig. 2, corollary equipment of the spectacle case as glasses, the spectacle case top and the bottom of this glasses conceal a shifting
Dynamic power supply, and the corresponding position design plug of glasses is placed on top, it can be connected with the microUSB interfaces of temple.In addition also
There are the mating sunglasses eyeglass that may be mounted at crossbeam upper groove, 3D eyeglasses etc..Needs based on mobile will have on mobile phone
Mating APP can manipulate glasses at any time by WIFI or bluetooth connection on APP.
As shown in figure 3, being Spectacles sunglasses, which may be implemented, by key control camera function, to utilize
Eye motion realizes the manipulation of function.
As shown in figure 4, being Blincam spectacles capable of photographing corollary equipments, the equipment and technology scheme eye movement sensor, it is necessary to
It could be realized and be taken pictures with non-constance force, image processing techniques is not utilized to realize that blink operation, eyewear concentrate on glasses
On one side, it is unfavorable for carrying for a long time, needs independent glasses, be not merged glasses and instrument, it is inconvenient to wear, outside
Shape is exaggerated very much, not hidden enough, can not achieve the functions such as recording.
As shown in figure 5, motion analysis instrument of the present invention first it according to the eye motion of eye detection module detect camera
Class eye information is rejected by eye areas precise extracting algorithm after acquisition eye and neighbouring photographic intelligence, realizes Primary Location
Eye areas obtains pure eye pictorial information (skin for including eyes and ocular vicinity), and the class eye information includes eyebrow
Hair and nose, then local auto-adaptive binary conversion treatment is carried out to the pure eye pictorial information, it obtains eyes and widens the view, closes one's eyes
The black white image of action message；Later, by motion analysis module, the eye for being exported eye motion detection module using grader
Eyeball is widened the view, the black white image of eye closing action message be divided into widen the view, two classes of closing one's eyes are per frame image category, then carry out blink action recognition
Processing, judges whether eyes have blink phenomenon, to obtain blink judging result；Finally, action recognition module passes through matching
The eye motion message form of preset function instruction judges the representative instruction meaning of blink action, is finally completed identification blink
The included function command information of action, the instruction of output camera function, recording start function command or End of Tape function command.
Still further aspect, push button signalling is also as the control function identification signal of data processor.
As shown in fig. 6, for the work flow diagram of the present invention, wherein eye motion analyzer mainly passes through following steps
It carries out：
(1) first, the eye of M*N sizes that eye camera takes and neighbouring photographic intelligence are longitudinally considered as Y-axis
J rows are denoted as, is laterally considered as X-axis and is denoted as i row, eye and neighbouring photographic intelligence are labeled as I (i, j), according to formulaMake the discrete projection function of picture Y-direction；Wherein Y (j) is that eye and neighbouring photo are believed
Breath I (i, j) Y-axis is projected after function, M, N be M*N sizes eye and neighbouring photographic intelligence I (i, j) length and
It is wide；
(2) then, the local maximum y of the discrete projection function is calculatedmaxY at this timemaxCorresponding abscissa j ', sentences
Break and pure eye pictorial information；
(3) finally, include outside eyebrow and the pure eye pictorial information of two kinds of nose using following function formula (6) rejecting
The class eye information that may influence monitoring, realizes the eye areas of Primary Location, obtains pure eye pictorial information Ir(a,b)；
Wherein Ir(a, b) is that former eye eye and neighbouring photographic intelligence I (i, j) pass through Ir(i-2, j-j ') functional transformation
Image afterwards, a=i-2, b=j-j ' be through function formula (6) transformation after pure eye pictorial information length and width.
The eyes feature extraction is to utilize complete local binary method (Completed Local Binary
Pattern it) carries out.Since the threshold value that illumination condition influences image binaryzation serious, simply to give image binaryzation can shadow
Widen the view in sound image, the acquisition of eye closing action message, needs that Adaptive Thresholding is utilized to carry out binary conversion treatment；
The grey level histogram for seeking needing using image of adaptive threshold, grey level histogram reflect an image pixel point
The statistical form of cloth, abscissa represent the type of the image pixel of gray scale, and ordinate represents each color value in the picture
Sum of all pixels or account for the percentage of all pixels number；
Wherein the complete local binarization of image and the grey level histogram using image seek adaptive threshold and obtain specific mistake
Cheng Wei：
(1) first, if IrThe tonal gradation of the image grey level histogram of (a, b) image is [0, L-1], according to formula (7)
Calculate the tonal gradation probability density function of I (a, b) image：
Wherein, rkIt is kth grade gray scale；nkIt is IrTonal gradation is r in (a, b)kNumber of pixels；P and Q is image I respectivelyr
The width and height of (a, b)；
(2) after, I is calculatedrThe distribution of the grey level histogram of (a, b) image it is expected, as shown in formula (8)：
(3) then, I is estimated according to the expectation E (r) acquiredrWhether some pixel belongs to skin gray scale in (a, b) image
Range or eyes tonal range are realized by formula (9) to IrThe skin of ocular vicinity and eyes is further in (a, b) image
Enhancing classification：
Wherein, r IrThe tonal gradation of each pixel of (a, b) image is as input；S is output gray level grade；Coefficient c
It is dimension scale constant for c, finds that when c values are 1.3 be optimal value according to practical debugging, other expressions are not belonging to " r<=E
(r) " other gray values；
(4) then, by enhancing classify after ocular vicinity skin and eyes image grey level histogram distribution more
It polarises, it is x that being located at, which enhances the range of the corresponding abscissa of sorted eye information grey level histogram,1~x2, corresponding grey scale is straight
The maximum one piece of range of ordinate value in square figure, it is x~g (x), maximum value g (r wherein in ordinate to take grey level histogramh)
Corresponding abscissa value is rh, then with rhMidpoint determines x according to formula (10) to abscissa both sides1、x2Value, wherein x2-x1Just
It is adaptive threshold T；
Wherein i=h-1, h-2 ,~, 1；J=h+1, h+2 ,~, n；A is termination condition, command deployment number and precision；
(5) finally, binary conversion treatment is carried out to image Ir (a, b) by Adaptive Thresholding, i.e.,
It obtains, wherein x is any point Ir (a, b) in image, and 1 and 0 respectively represents black and white, and S (x) is image local binaryzation
Working function, it is compared each pixel in image Ir (a, b), and pattern colour coloured picture is become to the binaryzation of black and white format
Image.
Image by this binaryzation is black white image, widening the view containing eyes, eye closing information, as next step action point
Analyse the basis of processing.
Wherein motion analysis module is realized as follows：
By black white image after handling adaptive threshold local binarization therein, using each frame image as a square
Battle array counts the accounting of black, the white two class pixel inside this image array and the first moment of image array as widening the view, close one's eyes
Action message is widened the view, the input of the action message closed one's eyes as support vector machines (SVM) algorithm, the action to each frame image
Information carries out Classification and Identification processing, sorts out and widens the view or eye closing action state.If being eye closing eye shape state, Xia Yizhuan by laststate
State is state of widening the view, then judges to be exactly the action once blinked at this time, realize the analysis of eye motion.
Wherein action recognition module carries out function command identifying processing according to the blink judging result of motion analysis module,
By matching the eye motion message form of preset function instruction, judges the representative instruction meaning of blink action, be finally completed
Identify that blink acts included function command information, the instruction of output camera function, recording start function command or End of Tape
Function command, specific workflow are as follows：
(1) it when motion analysis module captures the action of the first blink, counts once, occurs capturing in 0.2s continuous
Second of blink when, count second；
(2) if motion analysis instrument captures third time signal of blinking in 0.2s, third time, then motion analysis are counted
Instrument will action recognition be at this time camera function, count zero, output 01 is instructed as camera function；
(3) if after counting second, motion analysis module does not capture blink instruction after 1s, then motion analysis instrument
It is that recording starts function command by action recognition at this time, output 10 is as recording sign on, during recording, output
It remains 10 states, and counts zero；
(4) it if output is under the status condition that 10 states are that recording starts, has again identified that out continuous second
It is End of Tape function command by action recognition at this time when blink, it is 11 to export, and counts zero；With this, by fixed eye
Action is determined as taking pictures, starts recording, terminates sound-recording function instruction.
Wherein action recognition module carries out function command identifying processing according to the blink judging result of motion analysis module,
By matching the eye motion message form of preset function instruction, judges the representative instruction meaning of blink action, be finally completed
Identify that blink acts included function command information, the instruction of output camera function, recording start function command or End of Tape
Function command, specific workflow are as follows：
(1) it when motion analysis module captures the action of the first blink, counts once, occurs capturing in 0.2s continuous
Second of blink when, count second；
(2) if motion analysis instrument captures third time signal of blinking in 0.2s, third time, then motion analysis are counted
Instrument will action recognition be at this time camera function, count zero, output binary number 01 is instructed as camera function；
(3) if after counting second, motion analysis module does not capture blink instruction after 1s, then motion analysis instrument
It is that recording starts function command by action recognition at this time, output binary number 10 is as recording sign on, in the mistake of recording
Cheng Zhong, output remains 10 state of binary number, and counts zero；
(4) it if output is under the status condition that 10 state of binary number is that recording starts, has again identified that out continuous
Second of blink when, be End of Tape function command by action recognition at this time, export as binary number 11, count zero；
With this, fixed eye motion is determined as taking pictures, starts recording, end sound-recording function instruction.
The present invention can be operated by touch key-press form, and in side, temple is equipped with membrane keyboard, glasses take pictures and
The startup of sound-recording function can be realized there is pressure sensor on membrane keyboard, produced after of short duration touch by touch key-press form
Raw pressure, changes capacitance size, then will produce a kind of camera function command signal, and triggering camera is taken pictures, and is pressed for a long time super
Another recorded audio signals are generated after spending 1 second, can trigger recording sign on or End of Tape instruction.
The temple of the present invention, side is equipped with camera of taking pictures, LED flash and phonographic recorder；The temple of the other side is then
It is data processor；When data processor receives the camera function instruction that motion analysis instrument is sent out, data processor is rapid
Start camera acquisition present image of taking pictures, and detect color information in the present image information RGB that camera captures that takes pictures,
When color information is inadequate, LED flash is opened, data processor preserves camera the image collected letter of currently taking pictures later
Breath, otherwise, data processor directly preserves current the image collected；When data processor receives what motion analysis instrument was sent out
When recording starts function command, data processor starts rapidly phonographic recorder acquisition current sound information and starts to preserve, and works as data
When processor receives the End of Tape function command that motion analysis instrument is sent out, data processor terminates rapidly the acquisition of phonographic recorder
And it completes to preserve.
The workflow of taking pictures or record of the present invention is simply schematically as follows shown:
(1) the motion analysis instrument in glasses quickly positions eye locations according to eye detection module in real time first, obtains pure
Pure eye pictorial information Ir (a, b)；
(2) the eyes feature extraction of motion analysis instrument is mainly exported according to step (1) eye detection module in glasses
Pure eye pictorial information, according to adaptive local binarization method, by image binaryzation be processed into widening the view containing eyes,
The black and white picture of eye closing information exports；
(3) in glasses motion analysis instrument motion analysis module, mainly according to step (2) eyes characteristic extracting module
The widening the view containing eyes of output, the black and white pictorial information of eye closing information, by count inside black and white picture matrix it is black,
The input of the accounting of white two classes pixel and the first moment of image array as SVM algorithm differentiates the action widened the view and closed one's eyes, in turn
Judge the action of output blink；
(4) in glasses motion analysis instrument action recognition module, mainly according to step (3) motion analysis module output blinking
The action judging result of eye carries out the identifying processing of function command by matching fixed different type of action, identifies and take pictures
Function command, recording start function command and End of Tape function command as output；
(5) membrane keyboard in glasses is imitated according to power when touching long, in short-term, identification output camera function instruction and recording
Start function command and End of Tape function command；
(6) camera function instruction, the recording that the data processor in glasses is exported according to step (4) and step (5) start
Function command and End of Tape function command control the recording operation that camera of taking pictures carries out photographing operation and phonographic recorder.
The supporting structure of the present invention includes a hiding mobile power of spectacle case top and the bottom, and glasses are placed on top
Plug is designed in corresponding position, can be connected with the microUSB interfaces of temple.
Above example is provided just for the sake of the description purpose of the present invention, and is not intended to limit the scope of the present invention.This
The range of invention is defined by the following claims.It does not depart from spirit and principles of the present invention and the various equivalent replacements made and repaiies
Change, should all cover within the scope of the present invention.
Claims (16)
1. a kind of portable photographing glasses, including mirror holder and eyeglass, mirror holder includes Jing Quan, crossbeam and temple；It is characterized in that：Institute
It states and motion analysis instrument is installed on Jing Quan；Motion analysis instrument is located on the mirror circle crossbeam of both sides；
The motion analysis instrument includes：Eye detection module, motion analysis module and action recognition module；
Eye detection module contains eye motion detection camera as eye motion sensor, is taken the photograph by eye motion detection
As rejecting class eye information by eye areas precise extracting algorithm after head acquisition eye and neighbouring photographic intelligence, it is preliminary fixed to realize
The eye areas of position, obtains pure eye pictorial information, the pure eye pictorial information includes eyes and ocular vicinity
Skin, the class eye information includes eyebrow and nose；Eye motion information is carried out to the pure eye pictorial information again
Feature extraction, obtain eyes widen the view, the black white image of eye closing action message；
Motion analysis module, the widening the view of the eyes obtained according to eye detection module, eye closing action message is divided into using grader
Widen the view, the every frame image category of two classes of closing one's eyes, then carry out blink action recognition processing, judges whether eyes have blink phenomenon, from
And obtain blink judging result；
Action recognition module carries out function command identifying processing, passes through matching according to the blink judging result of motion analysis module
The eye motion message form of preset function instruction judges the representative instruction meaning of blink action, is finally completed identification blink
The included function command information of action, the instruction of output camera function, recording start function command or End of Tape function command.
2. a kind of portable photographing glasses according to claim 1, it is characterised in that：The eye detection module algorithm is real
It is now as follows：
(1) first, the eye of M*N sizes that eye camera takes and neighbouring photographic intelligence are longitudinally considered as Y-axis and are denoted as j
Row is laterally considered as X-axis and is denoted as i row, and eye and neighbouring photographic intelligence are labeled as I (i, j), according to formulaMake the discrete projection function of picture Y-direction；Wherein Y (j) is that eye and neighbouring photo are believed
Breath I (i, j) Y-axis is projected after function, M, N be M*N sizes eye and neighbouring photographic intelligence I (i, j) length and
It is wide；
(2) then, the local maximum y of the discrete projection function is calculatedmaxY at this timemaxCorresponding abscissa j ', judges
Pure eye pictorial information；
(3) finally, the class eye information of monitoring may be influenced by being rejected using following function formula (1), realize the eyes of Primary Location
Region obtains pure eye pictorial information Ir(a,b)；
Wherein Ir(a, b) is that former eye eye and neighbouring photographic intelligence I (i, j) pass through IrAfter (i-2, j-j ') functional transformation
Image, a=i-2, b=j-j ' be through function (1) transformation after pure eye pictorial information length and width.
3. a kind of portable photographing glasses according to claim 1, it is characterised in that：In the eye detection module, eye
The extraction of eyeball genius loci is carried out using complete local binary method (Completed Local Binary Pattern), due to
The threshold value that illumination condition influences image binaryzation serious, simply to give image binaryzation can influence to widen the view in image, close one's eyes
The acquisition of action message carries out binary conversion treatment using Adaptive Thresholding；
The complete local binarization of described image and using image grey level histogram seek adaptive threshold obtain detailed process be：
(1) I is setrThe tonal gradation of the image grey level histogram of (a, b) image is [0, L-1], and I (a, b) is calculated according to formula (2)
The tonal gradation probability density function of image：
Wherein, rkIt is kth grade gray scale；nkIt is IrTonal gradation is r in (a, b)kNumber of pixels；P and Q is image I respectivelyr(a,
B) width and height；
(2) according to P (rk), I is calculatedrThe distribution of the grey level histogram of (a, b) image it is expected, as shown in formula (3)：
(3) I is estimated according to the expectation E (r) acquiredrWhether some pixel belongs to skin tonal range or eyes in (a, b) image
Tonal range is realized by formula (4) to IrThe skin of ocular vicinity and eyes further enhances classification in (a, b) image：
Wherein, r IrThe tonal gradation of each pixel of (a, b) image is as input；S is output gray level grade；Coefficient c is ruler
Proportionality constant is spent, other expressions are not belonging to " r<Other gray values of=E (r) "；
(4) skin of ocular vicinity and eyes more polarise in the distribution of image grey level histogram after classifying using enhancing,
It is x that being located at, which enhances the range of the corresponding abscissa of sorted eye information grey level histogram,1~x2, in corresponding grey scale histogram
The maximum one piece of range of ordinate value, it is x~g (x), maximum value g (r wherein in ordinate to take grey level histogramh) corresponding
Abscissa value is rh, then with rhMidpoint determines x according to formula (5) to abscissa both sides1、x2Value, wherein x2-x1It is as adaptive
Answer threshold value T；
Wherein i=h-1, h-2 ,~, 1；J=h+1, h+2 ,~, n；A is termination condition, command deployment number and precision；
(5) by Adaptive Thresholding to image Ir(a, b) carries out binary conversion treatment, i.e.,It obtains, wherein
X is any point Ir (a, b) in image, and 1,0 respectively represents black and white, and S (x) is image local binaryzation working function,
It is compared each pixel in image Ir (a, b), and pattern colour coloured picture is become to the binary image of black and white format；By
The image of this binaryzation is black white image, widening the view containing eyes, eye closing information, the base as next step motion analysis processing
Plinth.
4. a kind of portable photographing glasses according to claim 1, it is characterised in that：The coefficient c values are for 1.3-1.5
Optimal value.
5. a kind of portable photographing glasses according to claim 1, it is characterised in that：The motion analysis module, identification
Two classes act：It closes one's eyes and widens the view, by black white image after handling the adaptive threshold local binarization, by each frame figure
As the first moment conduct of the accounting and image array that count black, the white two class pixel inside this image array as a matrix
The action message widen the view, closed one's eyes is widened the view, the input of the action message closed one's eyes as support vector machines (SVM) algorithm, to each frame
The action message of image carries out Classification and Identification processing, sorts out and widens the view or eye closing action state；If being eye closing eye by laststate
State, NextState are state of widening the view, then judge to be exactly the action once blinked at this time, realize the analysis of eye motion.
6. a kind of portable photographing glasses according to claim 1, it is characterised in that：The action recognition module, according to
The blink judging result of motion analysis module carries out function command identifying processing, dynamic by the eye for matching preset function instruction
Make message form, judge the representative instruction meaning of blink action, is finally completed identification blink and acts included function command
Information, the instruction of output camera function, recording start function command or End of Tape function command.
7. a kind of portable photographing glasses according to claim 6, it is characterised in that：The specific work of action recognition module
It is as follows to make flow：
(1) it when motion analysis module captures the action of the first blink, counts once, occurs the company of capturing in 0.1s-0.5s
When continuous second is blinked, count second；
(2) if motion analysis instrument captures third time signal of blinking in 0.1s-0.5s, third time is counted, then action point
Analyzer will action recognition be at this time camera function, count zero, output binary number 01 is instructed as camera function；
(3) if after counting second, motion analysis module does not capture blink instruction after 0.1s-0.5s, then action point
Action recognition at this time is that recording starts function command by analyzer, and output binary number 10 is being recorded as recording sign on
During, output remains 10 states, and counts zero；
(4) if 10 state of binary number of output is to record under the status condition started, again identified that out continuous the
It is End of Tape function command by action recognition at this time when secondary blink, output binary number is 11, counts zero；With this,
Fixed eye motion is determined as taking pictures, starts recording, end sound-recording function instruction.
8. a kind of portable photographing glasses according to claim 1, it is characterised in that：The temple, the installation of side temple
It takes pictures camera, LED flash and phonographic recorder；The temple of the other side is then data processor；When data processor receives
When the camera function sent out to motion analysis instrument instructs, data processor starts rapidly camera acquisition present image of taking pictures, and
Color information in the present image information RGB that camera captures that takes pictures is detected, when color information is inadequate, opens LED flashes of light
Lamp, the current camera the image collected information of taking pictures of data processor preservation later, otherwise, data processor, which directly preserves, works as
Preceding the image collected；
When data processor, which receives the recording that motion analysis instrument is sent out, starts function command, data processor starts rapidly record
Sound device acquires current sound information and starts to preserve, when data processor receives the End of Tape function that motion analysis instrument is sent out
When instruction, data processor terminates rapidly the acquisition of phonographic recorder and completes to preserve.
9. according to claim 1-8 it is one of arbitrary described in a kind of portable photographing glasses, it is characterised in that：Further include matching noose
Structure, the supporting structure：Needs based on charging, spectacle case top and the bottom conceal a mobile power, and place eye on top
Plug is designed in the corresponding position of mirror, is connected with the microUSB interfaces of temple.
10. a kind of portable photographing glasses according to claim 8, it is characterised in that：The data processing structure includes
Processor, reservoir, wireless signal transmitter and battery.
11. a kind of portable photographing glasses according to claim 8, it is characterised in that：The temple tail portion then there are one
USB interface is for charging.
12. a kind of portable photographing glasses according to claim 1, it is characterised in that：Groove is provided on the crossbeam, it will
Mating mating associated lens, including the installation of sunglasses eyeglass, 3D eyeglasses is in a groove.
13. a kind of portable photographing glasses according to claim 8, it is characterised in that：It is used in the camera of taking pictures
Sensor be cmos sensor.
14. a kind of portable photographing glasses according to claim 8, it is characterised in that：Containing tactile in the left part temple
Button is touched, the startup that glasses are taken pictures with sound-recording function is realized there is pressure sensor on membrane keyboard by touch key-press form,
Pressure is generated after of short duration touch, changes capacitance size, then generates a kind of camera function command signal, triggering camera is taken pictures, long
Another recorded audio signals are generated after time pressing, can trigger recording sign on or End of Tape instruction.
15. a kind of portable photographing glasses according to claim 1, it is characterised in that：Needs based on mobile, it is described
There is the APP of mating self-defined eye motion setting to be made by oneself at any time on APP by WIFI or bluetooth connection mobile phone in temple
Justice setting take pictures start when eye motion and recording start, at the end of eye motion.
16. a kind of method taken pictures or recorded using one of the arbitrary portable photographing glasses of claim 1-15,
It is characterized in that, steps are as follows：
(1) the motion analysis instrument in glasses quickly positions eye locations according to eye detection module in real time first, obtains pure
Eye pictorial information Ir (a, b)；
(2) in glasses the eyes feature extraction of motion analysis instrument mainly according to step (1) eye detection module output it is pure
Image binaryzation is processed into widening the view containing eyes, closed one's eyes according to adaptive local binarization method by pure eye pictorial information
The black and white picture of information exports；
(3) in glasses motion analysis instrument motion analysis module, according to step (2) eyes characteristic extracting module output contain
There are the widening the view of eyes, the black and white pictorial information of eye closing information, passes through black, the white two class picture counted inside black and white picture matrix
The input of the accounting of element and the first moment of image array as SVM algorithm differentiates the action widened the view and closed one's eyes, and then judges output
The action of blink；
(4) in glasses motion analysis instrument action recognition module, mainly according to the blink of step (3) motion analysis module output
Judging result is acted, the identifying processing of function command is carried out by matching fixed different type of action, identifies camera function
Instruction, recording start function command and End of Tape function command as output；
(5) membrane keyboard in glasses is imitated according to power when touching long, in short-term, and identification output camera function instruction and recording start
Function command and End of Tape function command；
(6) camera function instruction, the recording that the data processor in glasses is exported according to step (4) and step (5) start function
Instruction and End of Tape function command control the recording operation that camera of taking pictures carries out photographing operation and phonographic recorder.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810667833.1A CN108596161A (en) | 2018-06-26 | 2018-06-26 | A kind of portable photographing glasses and photographic method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810667833.1A CN108596161A (en) | 2018-06-26 | 2018-06-26 | A kind of portable photographing glasses and photographic method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108596161A true CN108596161A (en) | 2018-09-28 |
Family
ID=63634297
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810667833.1A Withdrawn CN108596161A (en) | 2018-06-26 | 2018-06-26 | A kind of portable photographing glasses and photographic method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108596161A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112052779A (en) * | 2020-09-01 | 2020-12-08 | 深圳市研盛芯控电子技术有限公司 | Access control system who possesses palm vein discernment and iris discernment |
| CN115866388A (en) * | 2022-11-24 | 2023-03-28 | 广州新城建筑设计院有限公司 | Intelligent glasses shooting control method and device, storage medium and electronic equipment |
-
2018
- 2018-06-26 CN CN201810667833.1A patent/CN108596161A/en not_active Withdrawn
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112052779A (en) * | 2020-09-01 | 2020-12-08 | 深圳市研盛芯控电子技术有限公司 | Access control system who possesses palm vein discernment and iris discernment |
| CN112052779B (en) * | 2020-09-01 | 2024-01-16 | 深圳市研盛芯控电子技术有限公司 | Access control system with palm vein recognition and iris recognition |
| CN115866388A (en) * | 2022-11-24 | 2023-03-28 | 广州新城建筑设计院有限公司 | Intelligent glasses shooting control method and device, storage medium and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Hernandez-Ortega et al. | Deepfakeson-phys: Deepfakes detection based on heart rate estimation | |
| CN106056064B (en) | A kind of face identification method and face identification device | |
| CN103024338B (en) | There is the display device of image capture and analysis module | |
| US20230360254A1 (en) | Pose estimation method and related apparatus | |
| WO2019128507A1 (en) | Image processing method and apparatus, storage medium and electronic device | |
| CN108470169A (en) | Face identification system and method | |
| CN105245786B (en) | A kind of self-timer method based on intelligent testing light, self-heterodyne system and camera terminal | |
| US20130076943A1 (en) | Apparatus and method for image recognition of facial areas in photographic images from a digital camera | |
| CN106203391A (en) | Face identification method based on intelligent glasses | |
| CN109359647A (en) | Identify the method, equipment and computer readable storage medium of a variety of certificates | |
| CN105874473A (en) | Apparatus and method for acquiring image for iris recognition using distance of facial feature | |
| CN106471521A (en) | The identification of self adaptation eyes artifact and correction system | |
| CN104143086A (en) | Application technology of portrait comparison to mobile terminal operating system | |
| WO2020259127A1 (en) | Registration method, registration device, headset equipment and storage medium | |
| CN111008971B (en) | Aesthetic quality evaluation method of group photo image and real-time shooting guidance system | |
| CN109977846B (en) | Living body detection method and system based on near-infrared monocular photography | |
| CN208351494U (en) | Face identification system | |
| CN109583375A (en) | A kind of the facial image illumination recognition methods and system of multiple features fusion | |
| CN109657627A (en) | Auth method, device and electronic equipment | |
| CN111860394A (en) | Gesture estimation and gesture detection-based action living body recognition method | |
| CN113887386A (en) | Fatigue detection method based on multi-feature fusion of deep learning and machine learning | |
| WO2023160075A1 (en) | Image inpainting method and apparatus, and device and medium | |
| CN108596161A (en) | A kind of portable photographing glasses and photographic method | |
| CN111259757B (en) | Living body identification method, device and equipment based on image | |
| CN103856708A (en) | Automatic focusing method, photographic device and computer readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20180928 |
|
| WW01 | Invention patent application withdrawn after publication |