CN105373767A - Eye fatigue detection method for smart phones - Google Patents
Eye fatigue detection method for smart phones Download PDFInfo
- Publication number
- CN105373767A CN105373767A CN201510437767.5A CN201510437767A CN105373767A CN 105373767 A CN105373767 A CN 105373767A CN 201510437767 A CN201510437767 A CN 201510437767A CN 105373767 A CN105373767 A CN 105373767A
- Authority
- CN
- China
- Prior art keywords
- fatigue
- degree
- mobile phone
- smart mobile
- eye
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/18—Eye characteristics, e.g. of the iris
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Ophthalmology & Optometry (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
According to the invention, the human eyes are accurately positioned by a scale-invariant gradient integral projection algorithm and a Kalman filtering method, the degree of fatigue of the human eyes is detected by a calculation method based on the PERLOG feature, and different intervention measures are taken according to the degree of fatigue. The process mainly comprises the following steps: first, an image frame sequence of a user when the user uses a smart phone is collected through the front-facing camera of the smart phone, all the frames are sequentially processed by the smart phone, and the smallest rectangle containing the human eyes is positioned; second, the smallest rectangle containing the human eyes is positioned within a certain range of accuracy, the rectangle obtained each time is scaled according to the same fixed ratio, the area of the rectangle is worked out, the opening and closing condition of the eyes is judged according to an eye opening and closing threshold obtained through training, and the PERLOG feature value for judging the degree of fatigue of the human eyes is updated; and third, corresponding intervention measures are taken according to the degree of fatigue.
Description
Technical field
The present invention relates to the application layer technology of operation system of smart phone, and the image processing techniques such as recognition of face, eye recognition.
Technical background
The ease for use of smart mobile phone, playability (function such as amusement, social activity, multimedia video), brought a lot of facility, and everybody life of extreme enrichment, allows everybody create very serious dependence to smart mobile phone.Combine claiming Chinese market smart mobile phone condition survey report of release according to GOOGLE and IPSOS, the universal of Chinese city smart mobile phone rises to 47% from 33% in 2012.Wherein the user of nearly 70% can use smart mobile phone to access internet every day, and smart mobile phone has become them and to have lived inseparable pith, and most people also have the custom using smart mobile phone at night.Inevitably, the excessive use of smart mobile phone brings great impact to the physical and mental health of user, and wherein digital kopiopia problem is exactly a problem highly paid attention to.Numeral kopiopia be exactly long-time produce before digital screen uncomfortable.Some typical symptoms comprise that eye is dry, the rubescent even inflammatory eye of eyes, and eye-blurred, even more serious can cause the problems such as back, neck and shoulder aches.And the universal of smart mobile phone becomes wherein main assailant behind the scenes undoubtedly.Smart mobile phone is except causing eye focus system fatigue, and the high-energy visible ray that it discharges can bring long-term impact to vision health.Research display, over-exposure can endanger retina under high-energy visible ray, and the possibility causing the eye disease such as AMD and cataract can increase, and can become more serious subsequently.And most people do not recognize these negative influences that electronic equipments such as using smart mobile phone brings.A kind ofly can detect user in the visual fatigue problem used in smart mobile phone process so design and to remind or mandatory intervention is necessary.
Although market there are some optics eyes can protecting eyes to a certain extent; but this does not all fundamentally solve the kopiopia problem using electronic equipment; be easy on the contrary connive user procedures to use these electronics to arrange; under the prerequisite protecting eyes, cross long-time these equipment of use is still cause otherwise body-mind problem, affects normal Working Life.These effective equipment are all costly simultaneously, are also contacts, are not used in most people.In order to avoid these problems can solve digital kopiopia problem again well simultaneously, the present invention proposes the solution that a kind of kopiopia arranged based on smart mobile phone detects and prevents.This be a kind of by cell-phone camera equipment untouchable gather user's facial image, and utilize the more powerful computing power of present smart mobile phone to process the fatigue state that image detects human eye, the object of feedback result is finally also reached by the interactive device (as the equipment such as display screen, loudspeaker) of smart mobile phone and the control interface of mobile phone operating system.Whole process does not rely on other hardware device, reaches the object of controlling cost well, only needs to realize optimization process in the design of software and just can reach more satisfactory effect.
Detect using the kopiopia of electronic equipment and prevent not obtain more concern, a small amount of inventor is had to use detection human eye to use equipment to keep suitable distance to the method reminding user of digital screen distance, cannot detect that user uses the degree of fatigue of electronic equipment, can not be the fatigue proposition prevention foundation of user rightly, this be also that a lot of solution prevents this problem asthenopic.
In order to understand in real time eyes of user use situation, we collect video sequence frame by analysis, can obtain user real-time use kopiopia degree, use different intervention means according to its degree of fatigue.Very effectively can prevent the harmful effect excessively caused with eye like this.
Summary of the invention
In order to improve everybody the eyeshield consciousness in intelligent movable electronic equipment examination use procedures such as smart mobile phones; and available protecting eye strain and the physically and mentally healthy problem that causes, the present invention proposes a kind of can effective monitor user ' eye condition and intervene the method for excessive eye.
To achieve these goals, we based on using now Smartphone device the most general as hardware platform, and do not rely on third party's hardware of other complex and expensive.
Preferably, we select comparatively suitable video image sample frequency, both can realize more satisfactory Detection results, react eye fatigue degree in certain accuracy rating, the impact on other operation task of smart mobile phone can be alleviated again well, keep the availability of smart machine.
Before human eye location, necessary pre-service is done to process image, efficiency and the accuracy of human eye location can be improved preferably.Comprise skin color segmentation and Face detection, after Face detection, wherein employ the face of support vector machines to location verify, effectively reject non-locating dislocation zone, the precise region of location can also be reduced simultaneously.In view of the difference of smart phone user use habit and reading posture, also need to carry out necessary rectification to human face region, reach the object exporting positive face image.
Preferably, use a kind of strategy of global scan and verification, namely adopt cascade structure to carry out tissue classifier, adopt Adaboost algorithm Study strategies and methods.This algorithm eliminates the impact of positioning feature point error well, and classification performance has and significantly improves, and speed does not obviously decline, and substantially reaches real-time.The memory space of this algorithm is very little, and about about 2M reaches good performance in robustness, accuracy and speed.
The conventional projecting method of human eye location has quick and efficient feature, but does not have degree of accuracy preferably in various complex environment.Present invention uses Scale invariant gradient integral projection algorithm (SGIPF) and to be used as in facial image human eye area segmentation, obtain comprise more accurately people's eyes minimum rectangle, even if under the input condition of low-quality image.
In order to raising degree improving efficiency and the real-time of detection of human eye location, use Kalman filtering to extract position and these characteristic parameters of motion state of each human eye, these characteristic parameters of association successive image frame can be followed the tracks of operational objective and predict.This tracing process comprises constantly the process of estimating-measuring repeatedly-revise.
Preferably, obtaining under the rectangular area prerequisite comprising eyes, simple computation rectangular area is relatively more direct way efficiently as judging that eyes open the foundation of closing.In view of the situation of individual human eye area discrepancy, need the reference frame minimum and maximum eyes rectangular surfaces product value of user chosen as threshold value.
Beneficial effect
In order to improve everybody the eyeshield consciousness in intelligent movable electronic equipment examination use procedures such as smart mobile phones; and available protecting eye strain and the physically and mentally healthy problem that causes, the present invention proposes a kind of can effective monitor user ' eye condition and intervene the method for excessive eye.This programme uses the kopiopia based on the improvement opportunity of ripe algorithm detecting when smart mobile phone uses, and can avoid illumination, wears glasses and in the situation such as low-quality image, testing result be had to the factor of considerable influence.Can obtain user real-time use kopiopia degree, use different intervention means according to its degree of fatigue.Very effectively can prevent the harmful effect excessively caused with eye like this.
Accompanying drawing explanation
Fig. 1 is the hardware module that system works relates to
Fig. 2 is the design of bulk treatment flow process
Fig. 3 is human eye testing process
Fig. 4 is the minimum rectangle design sketch that localization package contains human eye
Specific implementation method
For making object of the present invention, technical scheme and advantage clearly, are described in further detail embodiment of the present invention below in conjunction with accompanying drawing.
In order to improve real-time, the accuracy of human eye fatigue detecting, embodiments provide a kind of kopiopia detection method based on cell phone platform, see Fig. 2,3.Wherein Fig. 2 has set forth the treatment scheme of entire system, and wherein human eye location and the asthenopic evaluation of people are two important steps, describe in detail below in conjunction with Fig. 3:
201 to gather a frame image data every the time of 120 milliseconds;
Wherein, when applying PERCLOS method and judging eye fatigue degree, the sampling rate of image needs to reach certain requirement.When sampling interval duration is at least less than or equal to 120ms, the value of PERCLOS is stable, so just can not affect tired result of determination.
203 use complexion model segmentation area of skin color to carry out Face datection
1) image collected is calculated gray-scale value, then it is carried out non-linear piecewise look at YCbCr color space and transform in the color space of YCb ' Cr '.
2) change point obtaining image intensity value sees whether be arranged in this ellipse, thus determines whether it is face complexion.If so, step 3 is performed); If not, then the next frame inputting present frame is detected video image, re-executes step 1), until traveled through all frames of detected video image;
Wherein, from light compensation and nonlinear color transformation, colour of skin point is gathered in an ellipse.Therefore, the area of skin color of the picture gathered will be separated with non-area of skin color by we, just needs first by image binaryzation, order
s=e
-tonly have as s < e
-t, then colour of skin point is thought.
206 carry out Face detection employing be cascade AdaBoost method and SimpleSVM checking algorithm.Wherein, utilize the step of cascade AdaBoost Face datection as follows:
1) according to training sample, a Weak Classifier is trained for each possible rectangular characteristic;
2) suitable Weak Classifier is selected, according to cascade AdaBoost algorithm, calculate the classification error rate of the classification results of each Weak Classifier, and select the Weak Classifier with minimal error rate, classification results according to this sorter upgrades sample weights, the result that weight upgrades is the weight of the sample increasing this sorter classification error, these samples of retraining to make the Weak Classifier selected below.
3) step 2 is repeated), until select a Weak Classifier;
4) T Weak Classifier selecting of cascade, forms strong classifier by following formula;
5) step 2 is repeated) to 4), structure cascade of strong classifiers;
6) for the image of input, in the cascade classifier train the input of each possible subwindow, obtain testing result, then merge some adjacent subwindows and obtain final Face datection result.In order to obtain more through accurate Output rusults, also need to use SVM classifier to do further scanning.Output rusults after these two kinds of sorter process is just as the final face exported.
207 faces exported for detection-phase, although be that front is rectified haply, for the small scale face under different attitude and under low resolution, for reaching accuracy and rapidity, also available SimpleDAM corrects further.The concrete steps using SimpleDAM face to correct are as follows:
1) initialization current texture is face texture t ← t that testing result is confined
0,
2) according to current texture, by the linear relationship formula s=R*t+ ε between strip and texture, the position of three unique points is obtained.If the position of three unique points and mean place are very close, then terminate.
Wherein, wherein t is through the projection of certain face texture corrected at its principle components space (PCA), and s is the projection of shape at its principle components space.In our method, consider the simplest situation, only need three pairs of corresponding point, the face just non-frontal can rectified, is corrected to the attitude that front is rectified.Our hypothesis, Face datection exports the face texture confined, and these three unique points on the face " between the vector that eyes and face " center " form, exist simple linear relationship s=R*t+ ε.
3) according to the position of three unique points, whole picture (or comprise face and around an image window on) on apply affined transformation, will tilt face normalization; Make current texture be correct after face texture; Forward 2 to).
208 adopt the innovatory algorithm based on projection algorithm to carry out the location of human eye, and common projection algorithm can not locate human eye badly from facial image.And extraordinary effect can be obtained based on Scale invariant gradient integral projection algorithm (SGIPF) human eye location.Its step can be described as:
1. calculate the SGIPF of facial image to be detected, using the maximal value place of SGIPF curve as eyes at face
Position in image in vertical direction, is designated as P
v;
2. then upper and lower for the curve near maximal value two parts are fitted to the curve of monotonic decreasing by normalization SGIPF curve respectively;
3. according to formula 2.14, on the curve after normalization and dull matching, from position P
vset out respectively to two-sided search in place, until stop when the value of curve drops to α=0.5, the position of stopping place, respectively as the up-and-down boundary of eyes, remembers that the distance now between two borders is H;
4. if H ∈ [0.05,0.2], then think that the border that previous step draws is correct, return results; Otherwise, continue next step;
5. repeat the 3rd step, according to formula
P(0.5)＞P(0.4)＞P(0.6)＞P(0.3)＞P(0.7)＞P(0.2)
In order, select α=0.5 successively, 0.4,0.6,0.3,0.7,0.2 to draw as stop condition and two catches up with lower boundary, and calculates the spacing H of up-and-down boundary, until meet H ∈ [0.05,0.2], if do not met, then using distance up and down each 0.1*m place as eyes up-and-down boundary (m is the height of facial image).
In order to improve the efficiency of detection, whole processing procedure combines kalman filter method and optimizes further, all can do a prediction to next frame image position of human eye after every two field picture process.Kalman filtering state-space method descriptive system, determines the state variable of description eye state, measurement equation and initialization condition below.
1) because the relative mobile phone screen of eyes of user constantly moves and position is not fixed, the state variable therefore describing eye state can be designated as: x
k-1=[m
k-1, n
k-1, u
x-1, v
k-1].Wherein m, n are the position on horizontal and vertical direction, and u, v are the speed in horizontal and vertical direction.
2) State Forecasting Model is set up: owing to not having the prediction input quantity of eye state, its state equation can be expressed as
wherein A is state-transition matrix, and its value rises certainly in the time interval of adjacent two frames; P (w) the i.e. probability density function clothes of error, from Gaussian distribution.
3) state updating model is set up: in note process flow diagram, previous step is x by the eye position that human eye detection obtains
k=(m
k, n
k), then the renewal equation that can obtain its quantity of state is
be easy to get
V
kthe measuring error of previous step human eye, Gaussian distributed.
4) starting condition is determined: Kalman filtering method be utilized to predict the position of human eye, following starting condition must be determined, human eye original state x
0, x
0error co-variance matrix p
0, the initial value of the error co-variance matrix Q of status predication and the error co-variance matrix R of state updating.Appoint and get the image that two continuous frames accurately detects position of human eye, its position of human eye is designated as (m
t, n
t) and (m
t+1, n
t+1), then x
0=[m
t+1, n
t+1, (m
t+1-m
t)/A, (n
t+1-n
t)/A], the error of covariance matrix p of state variable
0, the error co-variance matrix Q of status predication and the error co-variance matrix R of state updating depends on the error of eye position pixel and the error of prediction of speed, can according to actual conditions sets itself.
Utilize above-mentioned measurement equation and initialization condition, just can be gone out the position of next frame human eye with Kalman filtering method by iteration Accurate Prediction.
210 use and carry out appraiser's kopiopia state based on the computing method of PERLOG feature, time scale shared when referring to eyes closed within the regular hour.Research shows according to many data: if the ratio of eyes closed reaches more than 70% within the unit interval, meet formula below, then can think user's kopiopia.
As can be seen from the definition of PERLOG, the key of these computing method is the closed degree calculating eyes.Can be calculated the area of two by the eye image of the prediction of complexion model gained above, thus the degree of opening characterizing eyes is extraordinary way.Its specific practice is divided into three steps below:
1) human eye opens area a maximal value and minimum value, is designated as M
left, M
rightand m
left, m
right; Along with the collection of data, two areas function over time can be obtained, be designated as s (t)
leftwith s (t)
right.Wherein M
left, M
rightand m
left, m
right(note: M is determined according to the data collected before each user
leftand M
rightminimum value for non-zero).
2) at detection-phase, if s (t)
leftwith s (t)
rightoccur comparing M
left, M
rightlarge value or occurred comparing M
leftand M
rightlittle value, these values are designated as M
left, M
rightand m
left, m
right, at next frame, adjustment M
left, M
rightand m
left, m
rightvalue.
3) degree of opening of note right and left eyes is p (t)
leftwith p (t)
right:
P (t)
left=(s (t)
left-m
left)/(M
left-m
left)
P (t)
right=(s (t)
right-m
right)/(M
right-m
right)
P (t)
leftwith p (t)
rightbe the value after normalization, eyes are opened degree p (t) and PERLOG and are:
P (t)=(p (t)
left+ p (t)
right)/2
PERLOG＝∑(p(t)≤0.3)
Finally, in forecast model, kopiopia is divided into three ranks in various degree: slight kopiopia, moderate kopiopia and severe kopiopia.According to the judgement of forecast model, respectively facing to three kinds in various degree kopiopia provide different interventions:
1) if slightly tired situation, use voice reminder, and use has to the full frame prompting rest of the friendly UI of eye visual comparison and provides eye protection suggestion;
2) if moderate fatigue state, use voice reminder, and it is standby directly to trigger mobile phone blank screen;
3) if sever fatigue state, use voice reminder, and trigger shutdown command after countdown blink.
Claims (6)
1. an eye fatigue detection system for smart mobile phone use, is characterized in that:
(1) by the preposition picture pick-up device of smart mobile phone, user is gathered at the sequence of image frames using smart mobile phone process;
(2) smart mobile phone is used to process each frame successively, by the minimum rectangle of the accurate localization package of a series of technology containing human eye;
(3) in certain accuracy rating, the minimum rectangle comprising human eye is navigated to.The rectangular zoom at every turn obtained calculates rectangular area in identical fixed proportion, then according to train obtain the threshold determination eyes that open and close eyes open the situation of closing;
(4) the PERCLOS eigenwert judging human eye degree of fatigue is upgraded;
(5) if do not have fatigue that cycle detection will be continued, step (1) is returned; If there is kopiopia situation in various degree, (6) step will be performed;
(6) different intervening measures is made according to degree of fatigue.
2. the eye fatigue detection method of smart mobile phone use according to claim 1, it is characterized in that, although smart mobile phone major part has higher hardware configuration, but too high-frequency data acquisition and calculation task can affect the smooth degree that user uses other task of smart mobile phone process to a certain extent, can obtain good Detection results basis is exhausted and may be necessary by little picture-taken frequency.Experimental result shows to choose that to get a picture frame sample every 120ms be than better suited.
3. accurate localization package according to claim 1 is containing the technology of human eye minimum rectangle, it is characterized in that, use on improvement opportunity based on ripe algorithm, illumination can be avoided, wear glasses and in the situation such as low-quality image, testing result had to the factor of considerable influence.Main step is as follows:
(1) complexion model segmentation area of skin color is used to do necessary pre-service;
(2) human eye predicted position is evaluated.If prediction is correct, directly terminate the human eye location of this two field picture; If the mistake of predicting the outcome, continue to perform step (3);
(3) use the CascadeAdaboost trained to position face, and use the method for support vector machines to verify positioning result;
(4) correct face direction, what (5) were inputted is all positive face image;
(5) use the partitioning algorithm of Scale invariant gradient integral projection algorithm (Scale-invariantGradientIntegralProjectionFunction is called for short SGIPF), isolate the minimum rectangle comprising human eye;
(6) use Kalman filtering algorithm to predict position of human eye next time, correct predicting the outcome can improve the detection efficiency of system greatly.
4. human eye degree of fatigue assessment technique according to claim 1, it is characterized in that, PERCLOS (PercentageofEyelidClosureOverTime) is generally acknowledged fatigue state judgment criteria, has now been considered to judge the most effective evaluate parameter of driving fatigue.The present invention uses the value of PERCLOS to use kopiopia degree to assess user using in Smartphone device fatigue detecting.According to the tired feature that smart mobile phone uses, threshold value needs to do suitable adjustment, and then as the evaluation method of evaluation degree of fatigue, degree of fatigue is divided into slightly, moderate and severe three ranks.
5. the eye fatigue detection method of smart mobile phone use according to claim 1, it is characterized in that, the intervening measure after kopiopia degree being detected is divided into:
(1) if slightly tired situation, use voice reminder, and use has to the full frame prompting rest of the friendly UI of eye visual comparison and provides eye protection suggestion;
(2) if moderate fatigue state, use voice reminder, and it is standby directly to trigger mobile phone blank screen;
(3) if sever fatigue state, use voice reminder, and trigger shutdown command after countdown blink.
6. the realization of this method is based on smart mobile phone, but flat board waits mobile electronic equipment principle to be in realization similar, and the identical actualizing technology based on these equipment is also belong to this claim.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510437767.5A CN105373767A (en) | 2015-07-23 | 2015-07-23 | Eye fatigue detection method for smart phones |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510437767.5A CN105373767A (en) | 2015-07-23 | 2015-07-23 | Eye fatigue detection method for smart phones |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105373767A true CN105373767A (en) | 2016-03-02 |
Family
ID=55375953
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510437767.5A Pending CN105373767A (en) | 2015-07-23 | 2015-07-23 | Eye fatigue detection method for smart phones |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105373767A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105512613A (en) * | 2015-11-26 | 2016-04-20 | 中山大学 | Smartphone-based eye fatigue detection method |
| CN105853160A (en) * | 2016-03-28 | 2016-08-17 | 天脉聚源(北京)传媒科技有限公司 | Eyesight fatigue preventing method and device |
| CN106200939A (en) * | 2016-06-29 | 2016-12-07 | 广东欧珀移动通信有限公司 | Sight protectio method based on terminal unit, device and terminal unit |
| CN106503614A (en) * | 2016-09-14 | 2017-03-15 | 厦门幻世网络科技有限公司 | A kind of photo acquisition methods and device |
| CN106803065A (en) * | 2016-12-27 | 2017-06-06 | 广州帕克西软件开发有限公司 | A kind of interpupillary distance measuring method and system based on depth information |
| CN107085715A (en) * | 2017-05-19 | 2017-08-22 | 武汉理工大学 | A kind of television set intelligently detects the dormant system and method for user |
| CN107132675A (en) * | 2017-05-15 | 2017-09-05 | 盐城华星光电技术有限公司 | A kind of LCD MODULE（LCM）And viewing ratio determines method |
| CN107145864A (en) * | 2017-05-08 | 2017-09-08 | 湖南科乐坊教育科技股份有限公司 | A kind of concentration appraisal procedure and system |
| WO2018233072A1 (en) * | 2017-06-21 | 2018-12-27 | 深圳市盛路物联通讯技术有限公司 | Weather information combined vehicle parking guidance system and method |
| CN110825220A (en) * | 2019-09-29 | 2020-02-21 | 深圳市火乐科技发展有限公司 | Eyeball tracking control method and device, intelligent projector and storage medium |
| CN111399733A (en) * | 2020-03-23 | 2020-07-10 | 郭继胜 | Method, device, control equipment and system for solving addiction of electronic equipment |
| CN113591666A (en) * | 2021-07-26 | 2021-11-02 | 深圳创联时代电子商务有限公司 | Control method and device applied to mobile phone, computer readable medium and mobile phone |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104013414A (en) * | 2014-04-30 | 2014-09-03 | 南京车锐信息科技有限公司 | Driver fatigue detecting system based on smart mobile phone |
| CN104574817A (en) * | 2014-12-25 | 2015-04-29 | 清华大学苏州汽车研究院（吴江） | Machine vision-based fatigue driving pre-warning system suitable for smart phone |
-
2015
- 2015-07-23 CN CN201510437767.5A patent/CN105373767A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104013414A (en) * | 2014-04-30 | 2014-09-03 | 南京车锐信息科技有限公司 | Driver fatigue detecting system based on smart mobile phone |
| CN104574817A (en) * | 2014-12-25 | 2015-04-29 | 清华大学苏州汽车研究院（吴江） | Machine vision-based fatigue driving pre-warning system suitable for smart phone |
Non-Patent Citations (3)
| Title |
|---|
| H.SCHNEIDERMAN,T.KANADE: "A statistical method for 3D object detection applied to faces and cars", 《PROCEEDINGS IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION》 * |
| 孟春宁等: "基于梯度积分投影和最大期望算法的人眼定位", 《光电子·激光》 * |
| 艾海舟等: "人脸检测与检索", 《计算机学报》 * |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105512613A (en) * | 2015-11-26 | 2016-04-20 | 中山大学 | Smartphone-based eye fatigue detection method |
| CN105853160B (en) * | 2016-03-28 | 2018-03-06 | 天脉聚源(北京)传媒科技有限公司 | A kind of method and device of anti-asthenopia |
| CN105853160A (en) * | 2016-03-28 | 2016-08-17 | 天脉聚源(北京)传媒科技有限公司 | Eyesight fatigue preventing method and device |
| CN106200939A (en) * | 2016-06-29 | 2016-12-07 | 广东欧珀移动通信有限公司 | Sight protectio method based on terminal unit, device and terminal unit |
| CN106200939B (en) * | 2016-06-29 | 2019-04-19 | Oppo广东移动通信有限公司 | Sight protectio method, apparatus and terminal device based on terminal device |
| CN106503614A (en) * | 2016-09-14 | 2017-03-15 | 厦门幻世网络科技有限公司 | A kind of photo acquisition methods and device |
| CN106503614B (en) * | 2016-09-14 | 2020-01-17 | 厦门黑镜科技有限公司 | Photo obtaining method and device |
| CN106803065A (en) * | 2016-12-27 | 2017-06-06 | 广州帕克西软件开发有限公司 | A kind of interpupillary distance measuring method and system based on depth information |
| CN107145864A (en) * | 2017-05-08 | 2017-09-08 | 湖南科乐坊教育科技股份有限公司 | A kind of concentration appraisal procedure and system |
| CN107132675A (en) * | 2017-05-15 | 2017-09-05 | 盐城华星光电技术有限公司 | A kind of LCD MODULE（LCM）And viewing ratio determines method |
| CN107085715A (en) * | 2017-05-19 | 2017-08-22 | 武汉理工大学 | A kind of television set intelligently detects the dormant system and method for user |
| WO2018233072A1 (en) * | 2017-06-21 | 2018-12-27 | 深圳市盛路物联通讯技术有限公司 | Weather information combined vehicle parking guidance system and method |
| CN110825220A (en) * | 2019-09-29 | 2020-02-21 | 深圳市火乐科技发展有限公司 | Eyeball tracking control method and device, intelligent projector and storage medium |
| CN110825220B (en) * | 2019-09-29 | 2023-12-08 | 深圳市火乐科技发展有限公司 | Eyeball tracking control method, device, intelligent projector and storage medium |
| CN111399733A (en) * | 2020-03-23 | 2020-07-10 | 郭继胜 | Method, device, control equipment and system for solving addiction of electronic equipment |
| CN113591666A (en) * | 2021-07-26 | 2021-11-02 | 深圳创联时代电子商务有限公司 | Control method and device applied to mobile phone, computer readable medium and mobile phone |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105373767A (en) | Eye fatigue detection method for smart phones | |
| US11393206B2 (en) | Image recognition method and apparatus, terminal, and storage medium | |
| Wang et al. | Blink detection using Adaboost and contour circle for fatigue recognition | |
| CN102324025B (en) | Human face detection and tracking method based on Gaussian skin color model and feature analysis | |
| CN102831439B (en) | Gesture tracking method and system | |
| US11715231B2 (en) | Head pose estimation from local eye region | |
| CN107085715A (en) | A kind of television set intelligently detects the dormant system and method for user | |
| KR20210100602A (en) | Face image-based risk recognition method, apparatus, computer device and storage medium | |
| CN106257489A (en) | Expression recognition method and system | |
| US9892315B2 (en) | Systems and methods for detection of behavior correlated with outside distractions in examinations | |
| CN105354986A (en) | Driving state monitoring system and method for automobile driver | |
| CN103150870B (en) | Train motorman fatigue detecting method based on videos | |
| US20140016860A1 (en) | Facial analysis to detect asymmetric expressions | |
| KR20200130440A (en) | A method for identifying an object in an image and a mobile device for executing the method (METHOD FOR IDENTIFYING AN OBJECT WITHIN AN IMAGE AND MOBILE DEVICE FOR EXECUTING THE METHOD) | |
| CN108596087B (en) | Driving fatigue degree detection regression model based on double-network result | |
| CN105224285A (en) | Eyes open and-shut mode pick-up unit and method | |
| KR20200012355A (en) | Online lecture monitoring method using constrained local model and Gabor wavelets-based face verification process | |
| CN105912126B (en) | A kind of gesture motion is mapped to the adaptive adjusting gain method at interface | |
| CN110705454A (en) | Face recognition method with living body detection function | |
| CN105844727A (en) | Intelligent dynamic human face recognition attendance checking record management system | |
| Pandey et al. | Dumodds: Dual modeling approach for drowsiness detection based on spatial and spatio-temporal features | |
| US20160140395A1 (en) | Adaptive sampling for efficient analysis of ego-centric videos | |
| CN109657550B (en) | Fatigue degree detection method and device | |
| AlKishri et al. | Enhanced image processing and fuzzy logic approach for optimizing driver drowsiness detection | |
| Poon et al. | Driver distracted behavior detection technology with YOLO-based deep learning networks |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20160302 |
|
| WD01 | Invention patent application deemed withdrawn after publication |