CN110276276A - The determination method and system of examinee's face direction of visual lines in a kind of Driving Test - Google Patents
The determination method and system of examinee's face direction of visual lines in a kind of Driving Test Download PDFInfo
- Publication number
- CN110276276A CN110276276A CN201910475340.2A CN201910475340A CN110276276A CN 110276276 A CN110276276 A CN 110276276A CN 201910475340 A CN201910475340 A CN 201910475340A CN 110276276 A CN110276276 A CN 110276276A
- Authority
- CN
- China
- Prior art keywords
- visual lines
- face
- face direction
- facial image
- determination method
- 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
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/59—Context or environment of the image inside of a vehicle, e.g. relating to seat occupancy, driver state or inner lighting conditions
- G06V20/597—Recognising the driver's state or behaviour, e.g. attention or drowsiness
-
- 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
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
-
- 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
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
-
- G—PHYSICS
- G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
- G09B—EDUCATIONAL OR DEMONSTRATION APPLIANCES; APPLIANCES FOR TEACHING, OR COMMUNICATING WITH, THE BLIND, DEAF OR MUTE; MODELS; PLANETARIA; GLOBES; MAPS; DIAGRAMS
- G09B19/00—Teaching not covered by other main groups of this subclass
- G09B19/16—Control of vehicles or other craft
- G09B19/167—Control of land vehicles
Landscapes
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Educational Technology (AREA)
- Educational Administration (AREA)
- Entrepreneurship & Innovation (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses the determination method and system of examinee's face direction of visual lines in a kind of Driving Test, method includes: the interior image for obtaining driver's seat region;Interior image is detected using human face target detection model, determine whether there is facial image, if it exists, then facial image is input in face direction of visual lines disaggregated model, to calculate confidence level of the facial image respectively on preset N class face direction of visual lines;Facial image is ranged into the maximum face direction of visual lines of confidence value.Its system includes: the memory for collecting computer program instructions for storing data, establishes the processor communicated with memory, the processor executes the step of computer program instructions are to execute the determination method.It is primarily adapted for use in Driving Test Subject Three using this programme, direction can be paid close attention to the sight of driver and quickly accurately determined, and will determine that result is uploaded to points-scoring system, auxiliary examiner makes just marking.
Description
Technical field
The present invention relates to the artificial intelligence judgment technology field of motor vehicle driving license examination, in particular to examined in a kind of Driving Test
The determination method and system of stranger's face direction of visual lines.
Background technique
Constantly improve with living standards of the people with the continuous social and economic development, Urban vehicles poputation rapidly increases
Long, the number that driving license test is participated in annual application is also increasing.The workload of driving license test also increases with it.
Traditional driving license test is judged mainly by artificial cognition, and examiner is sitting in the juxtaposed position of driver, is difficult standard
True making to the sight concern direction of examination personnel accurately checks and judges.
Currently, can be used in the prior art check pilot's line of vision direction technological means mainly pass through it is wearable
Sensor is realized.However, these equipment are not only at high price, but also inconvenient for use, and error is larger in detection, can not be by it
Suitable for existing examination of driver system.
Summary of the invention
For the technical problems existing in the prior art mentioned above, object of the present invention is to: it proposes to examine in a kind of Driving Test
The determination method and system of stranger's face direction of visual lines, by utilizing visual pattern processing technique, to driver in Driving Test process
In sight concern direction detected automatically, assist examiner judge driver whether there is unlawful practice.
The technical solution adopted by the present invention to solve the technical problems is:
The determination method of examinee's face direction of visual lines, includes the following steps: in a kind of Driving Test
S1, the interior image for obtaining driver's seat region;
S2, the interior image got is carried out using the preset human face target detection model based on deep learning
Detection, and obtain testing result;
S3, facial image is determined whether there is according to testing result, if it exists facial image, then obtain the face figure
Picture;
S4, the facial image is input in the preset face direction of visual lines disaggregated model based on deep learning, with
Calculate confidence level of the facial image respectively on preset N class face direction of visual lines;
S5, the facial image is ranged to the maximum face direction of visual lines of confidence value, and then obtains facial image
Face direction of visual lines categorizing information.
Advanced optimize technical solution, the determination method further include:
The human face target detection model of A1, prebuild based on deep learning;
The face direction of visual lines disaggregated model of A2, prebuild based on deep learning.
Advanced optimize technical solution, the prebuild based on deep learning human face target detection model the step of wrap
It includes:
A11, human face target detection infrastructure network is built;
A12, the interior image pattern for obtaining different angle, illumination and picture quality;
A13, the facial image in the interior image pattern is marked using rectangle frame;
A14, the location information [x ', y ', width ', height '] for recording the rectangle frame;Wherein, x ', y ' indicate rectangle
Frame top left co-ordinate point, width ' indicate that rectangle width, height ' indicate rectangular elevation;
It A15, take the location information of the rectangle frame in every interior image pattern and the car image pattern as one group of training
Data form target detection training dataset;
A16, infrastructure network is detected using the target detection training dataset training human face target, parameter is more
The human face target detection model based on deep learning is obtained after new.
Technical solution is advanced optimized, the human face target detection infrastructure network includes: by 8 convolutional layers, 8
ReLU active coating and 3 pond layers composition face characteristic extractor, and connect with the face characteristic extractor first
Convolutional layer and the second convolutional layer.
Technical solution is advanced optimized, the facial image detection model using preset based on deep learning is to acquisition
To the interior image detected, and obtain testing result and include the following steps:
S21, the interior image that will acquire, which are input in the face characteristic extractor, obtains face characteristic V;
S22, the face characteristic V is inputted into calculating processing in first convolutional layer, obtaining N number of size is 2*13*13
Array;Wherein, each array represents original image is projected in the region of 13*13 size after each position there are face and
There is no the confidence levels of face, can detecte the various sizes of face of N class altogether;
S23, the face characteristic V is inputted into calculating processing in second convolutional layer, obtaining N number of size is 4*13*13
Array, wherein after the representative of each array projects to original image in the region of 13*13 size, a left side for face in the area
The offset at upper angle and two, lower right corner point and the regional center point;
S24, by the parsing to first convolutional layer and the calculated result of the second convolutional layer, finally obtain N number of one-dimensional
Array [score, x, y, width, height];
Wherein, x, y indicate that rectangle frame top left co-ordinate point, width indicate that rectangle width, height indicate rectangular elevation,
Score indicates that there are the probability of face in the region for the rectangle frame that [x, y, width, height] is represented.
Technical solution is advanced optimized, the facial image that determines whether there is according to testing result includes the following steps:
Size between S31, comparison score value and the threshold value of setting;
If S32, score value are greater than or equal to threshold value, it is determined that the location information institute in array corresponding to the score value
There are facial images in the rectangle frame of mark;
If S33, score value are less than threshold value, it is determined that the location information in array corresponding to the score value was identified
Facial image is not present in rectangle frame.
Advanced optimize technical solution, the prebuild based on deep learning face direction of visual lines disaggregated model the step of
Include:
A21, face direction of visual lines basis of classification network structure is built;
A22, the facial image marked in every interior image pattern by rectangle frame is obtained;
A23, N number of face direction of visual lines label is pre-seted, each face direction of visual lines label represents a kind of face sight side
To;
A24, the facial image is marked using face direction of visual lines label;
A25, using each facial image and corresponding face direction of visual lines label as one group of training data, shape
At face direction of visual lines classification based training data set;
A26, face direction of visual lines classification based training data set training face direction of visual lines basis of classification network knot is used
Structure learns facial image Pixel Information to the mapping relations of face direction of visual lines information, updates and save in deep neural network
Each network layer parameter, obtain the face direction of visual lines disaggregated model based on deep learning.
Technical solution is advanced optimized, the face direction of visual lines includes 8 classes, is respectively as follows: observation left side car door direction, sees
Examine left-hand mirror direction, observation inside rear-view mirror direction, scope dial side, observation right side car door direction, observation right rear view mirror
Direction, observation front and observation gear direction;
Technical solution is advanced optimized, the face direction of visual lines basis of classification network structure is by 8 convolutional layers, 7
The network structure of PRelu active coating, 3 pond layers and 1 Dropout layers of composition.
Advanced optimize technical solution, the determination method further include: be uploaded to face direction of visual lines categorizing information
Layer points-scoring system, to assist examiner to make scoring judgement.In addition, if facial image is not detected in image in the car, by nothing
Face information directly feeds back to upper layer points-scoring system.
The decision-making system of examinee's face direction of visual lines in a kind of Driving Test, comprising: memory is configured as storing data and meter
Calculation machine program instruction;
The processor communicated is established with memory, the processor executes the computer program instructions to execute above-mentioned
The step of one determination method.
The beneficial effects of the present invention are: being primarily adapted for use in Driving Test Subject Three using this programme, face mesh is realized
Mark detection and face direction of visual lines determine, can pay close attention to direction to the sight of driver using this programme and quickly accurately make
Judgement, and judging result is uploaded to points-scoring system, auxiliary examiner makes just marking, this has not only saved manpower, simultaneously
In turn ensure the just of work about test.
Detailed description of the invention
Fig. 1 is the structural block diagram of involved processing module in this programme.
Fig. 2 is the structural block diagram of image capture module.
Fig. 3 is the structural block diagram of facial image detection module.
Fig. 4 is the structural block diagram of direction of visual lines determination module.
The flow chart of determination method in Fig. 5 this programme.
In figure: 1- image capture module, 2- facial image detection module, 3- direction of visual lines determination module, the transmission of 4- signal
Module, 5- points-scoring system;11- acquisition unit, 12- data processing unit;21- Face datection unit, 22- face location region mark
Remember unit;31- face extraction unit, 32- direction of visual lines sort out unit.
Specific embodiment
Below in conjunction with attached drawing, the present invention will be further described.
Processing module involved in this programme is as shown in Figure 1, include image capture module 1, facial image detection module
2, direction of visual lines determination module 3 and signal transmission module 4.
Image acquisition units 1 acquisition unit 11 and data processing unit 12 as shown in Fig. 2, be made of.Wherein, acquisition unit
11 pass through transmission of network to data for image in collecting vehicle, and by collected interior image using monocular or binocular camera
Then processing unit 12 carries out preliminary treatment to image.
The hardware of acquisition unit 11 needs for camera to be mounted on during installation, in the case of monocular in front of steering wheel, guarantees
It can completely include the facial information of driver to driver's camera picture of different building shape;Two are imaged under biocular case
Head is separately mounted to driver's left and right side, and guarantee can completely include to two camera pictures of driver of different building shape
The facial information of driver;
Facial image detection module 2 is as shown in figure 3, by 22 structure of Face datection unit 21 and face location zone marker unit
At.Wherein, Face datection unit 21 is detected with the presence or absence of face in interior image, and the face location information transmission that will test
To face location zone marker unit 22, and face location region is marked using rectangle frame;If face position is not detected
Confidence breath, then directly fed back in signal transmission module 4, and be uploaded in points-scoring system 5.
Direction of visual lines determination module 3 is constituted as shown in figure 4, sorting out unit 32 by face extraction unit 31 and direction of visual lines.
Wherein, face extraction unit 31 extracts face characteristic information from face location region, and the face characteristic information is transmitted
Sort out unit 32 to direction of visual lines, direction of visual lines sorts out unit 32 by carrying out operation to the face characteristic of extraction, obtains the people
Face feature belongs to the confidence level of all pre-set categories, and is sorted out face characteristic information according to the confidence level.
Signal transmission module 4 receives the face direction of visual lines categorizing information of direction of visual lines processing unit 3, and is uploaded to
Upper layer points-scoring system 5, furthermore signal transmission module 4 can also directly receive the signal from facial image detection module 2, will be interior
The information of no facial image is transferred to points-scoring system 5.Wherein, it is 0 that Face datection state is reported if face is not detected;Such as
Fruit detects face and face direction of visual lines classification is k, then reporting Face datection state is 1, and direction of visual lines state is k.
The judgment method process of this programme is as shown in Figure 5, comprising:
S1, the interior image for obtaining driver's seat region;
S2, the interior image got is carried out using the preset human face target detection model based on deep learning
Detection, and obtain testing result;
S3, facial image is determined whether there is according to testing result, if it exists facial image, then obtain the face figure
Picture;
S4, the facial image is input in the preset face direction of visual lines disaggregated model based on deep learning, is obtained
To a 8 dimension groups [x1, x2, x3, x4, x5, x6, x7, x8];Wherein k-th of value represents current face image direction of visual lines people
Face direction of visual lines is kth kind shape probability of state;;
S5, by face direction of visual lines kind judging locating for the maximum value in array be facial image in face sight side
To classification, and then obtain the face direction of visual lines categorizing information of facial image.
Wherein, it before the determination method of this programme starts, also needs to do following preparation process:
The human face target detection model of A1, prebuild based on deep learning;
The face direction of visual lines disaggregated model of A2, prebuild based on deep learning.
Wherein, prebuild based on deep learning human face target detection model the step of include:
A11, human face target detection infrastructure network is built;The human face target detection infrastructure network includes: by 8
The face characteristic extractor of a convolutional layer, 8 ReLU active coatings and 3 pond layers composition, and extracted with the face characteristic
The first convolutional layer and the second convolutional layer of device connection;
A12, the interior image pattern for obtaining different angle, illumination and picture quality;
A13, the facial image in the interior image pattern is marked using rectangle frame;
A14, the location information [x ', y ', width ', height '] for recording the rectangle frame;Wherein, x ', y ' indicate rectangle
Frame top left co-ordinate point, width ' indicate that rectangle width, height ' indicate rectangular elevation;
It A15, take the location information of the rectangle frame in every interior image pattern and the car image pattern as one group of training
Data form target detection training dataset;
A16, infrastructure network is detected using the target detection training dataset training human face target, parameter is more
The human face target detection model based on deep learning is obtained after new.
Wherein, prebuild based on deep learning face direction of visual lines disaggregated model the step of include:
A21, face direction of visual lines basis of classification network structure is built;Wherein, face direction of visual lines basis of classification network knot
Structure is by 8 convolutional layers, 7 PRelu active coatings, 3 pond layers and 1 Dropout layers of network structure constituted;
A22, the facial image marked in every interior image pattern by rectangle frame is obtained;
A23, N number of face direction of visual lines label is pre-seted, each face direction of visual lines label represents a kind of face sight side
To;Wherein, face direction of visual lines includes 8 classes, be respectively as follows: observation left side car door direction, observation left-hand mirror direction, in observation after
Visor direction, scope dial side, observation right side car door direction, observation right rear view mirror direction, observation front and observation shelves
Position direction;
A24, the facial image is marked using face direction of visual lines label;
A25, using each facial image and corresponding face direction of visual lines label as one group of training data, shape
At face direction of visual lines classification based training data set;
A26, face direction of visual lines classification based training data set training face direction of visual lines basis of classification network knot is used
Structure learns facial image Pixel Information to the mapping relations of face direction of visual lines information, updates and save in deep neural network
Each network layer parameter, obtain the face direction of visual lines disaggregated model based on deep learning.
Wherein, the facial image detection model using preset based on deep learning examines the interior image got
It surveys, and obtains testing result and include the following steps:
S21, the interior image that will acquire, which are input in the face characteristic extractor, obtains face characteristic V;
S22, the face characteristic V is inputted into calculating processing in first convolutional layer, obtaining N number of size is 2*13*13
Array;Wherein, each array represents original image is projected in the region of 13*13 size after each position there are face and
There is no the confidence levels of face, can detecte the various sizes of face of N class altogether;
S23, the face characteristic V is inputted into calculating processing in second convolutional layer, obtaining N number of size is 4*13*13
Array, wherein after the representative of each array projects to original image in the region of 13*13 size, a left side for face in the area
The offset at upper angle and two, lower right corner point and the regional center point;
S24, by the parsing to first convolutional layer and the calculated result of the second convolutional layer, finally obtain N number of one-dimensional
Array [score, x, y, width, height];
Wherein, x, y indicate that rectangle frame top left co-ordinate point, width indicate that rectangle width, height indicate rectangular elevation,
Score indicates that there are the probability of face in the region for the rectangle frame that [x, y, width, height] is represented.
Wherein, facial image is determined whether there is according to testing result to include the following steps:
Size between S31, comparison score value and the threshold value of setting;
If S32, score value are greater than or equal to threshold value, it is determined that the location information institute in array corresponding to the score value
There are facial images in the rectangle frame of mark;
If S33, score value are less than threshold value, it is determined that the location information in array corresponding to the score value was identified
Facial image is not present in rectangle frame.
It, therefore, need to be by face in order to assist examiner to be set with points-scoring system in Driving Test system to the Driving Test scoring of examinee
Direction of visual lines categorizing information is uploaded to upper layer points-scoring system, to assist examiner to make scoring judgement.In addition, if in the car in image
Facial image is not detected, then will directly feed back to upper layer points-scoring system without face information.
In addition, hardware configuration involved in determination method includes: memory in the present invention, it is configured as storing data and meter
Calculation machine program instruction;
The processor communicated is established with memory, the processor executes the computer program instructions to execute above-mentioned
The step of one determination method.
The advantages of basic principles and main features and this programme of this programme have been shown and described above.The technology of the industry
Personnel are it should be appreciated that this programme is not restricted to the described embodiments, and the above embodiments and description only describe this
The principle of scheme, under the premise of not departing from this programme spirit and scope, this programme be will also have various changes and improvements, these changes
Change and improvement is both fallen within the scope of claimed this programme.This programme be claimed range by appended claims and its
Equivalent thereof.
Claims (10)
1. the determination method of examinee's face direction of visual lines in a kind of Driving Test, which comprises the steps of:
S1, the interior image for obtaining driver's seat region;
S2, the interior image got is examined using the preset human face target detection model based on deep learning
It surveys, and obtains testing result;
S3, facial image is determined whether there is according to testing result, if it exists facial image, then obtain the facial image;
S4, the facial image is input in the preset face direction of visual lines disaggregated model based on deep learning, to calculate
The facial image confidence level on preset N class face direction of visual lines respectively out;
S5, the facial image is ranged to the maximum face direction of visual lines of confidence value, and then obtains the face of facial image
Direction of visual lines categorizing information.
2. the determination method of examinee's face direction of visual lines in a kind of Driving Test as described in claim 1, which is characterized in that described to sentence
Determine method further include:
The human face target detection model of A1, prebuild based on deep learning;
The face direction of visual lines disaggregated model of A2, prebuild based on deep learning.
3. the determination method of examinee's face direction of visual lines in a kind of Driving Test as claimed in claim 2, which is characterized in that described pre-
The step of constructing human face target detection model based on deep learning include:
A11, human face target detection infrastructure network is built;
A12, the interior image pattern for obtaining different angle, illumination and picture quality;
A13, the facial image in the interior image pattern is marked using rectangle frame;
A14, the location information [x ', y ', width ', height '] for recording the rectangle frame;Wherein, x ', y ' indicate that rectangle frame is left
Upper angular coordinate point, width ' indicate that rectangle width, height ' indicate rectangular elevation;
A15, with the location information of the rectangle frame in every interior image pattern and the car image pattern for one group of training data,
Form target detection training dataset;
A16, infrastructure network is detected using the target detection training dataset training human face target, after parameter updates
Obtain the human face target detection model based on deep learning.
4. the determination method of examinee's face direction of visual lines in a kind of Driving Test as claimed in claim 3, which is characterized in that
The human face target detection infrastructure network includes: to be made of 8 convolutional layers, 8 ReLU active coatings and 3 pond layers
Face characteristic extractor, and the first convolutional layer and the second convolutional layer that are connect with the face characteristic extractor.
5. the determination method of examinee's face direction of visual lines in a kind of Driving Test as claimed in claim 4, which is characterized in that the benefit
The interior image got is detected with the preset facial image detection model based on deep learning, and obtains inspection
Result is surveyed to include the following steps:
S21, the interior image that will acquire, which are input in the face characteristic extractor, obtains face characteristic V;
S22, the face characteristic V is inputted into calculating processing in first convolutional layer, obtains the number that N number of size is 2*13*13
Group;Wherein, each position is deposited there are face and not after each array representative projects to original image in the region of 13*13 size
In the confidence level of face, the various sizes of face of N class can detecte altogether;
S23, the face characteristic V is inputted into calculating processing in second convolutional layer, obtains the number that N number of size is 4*13*13
Group, wherein after the representative of each array projects to original image in the region of 13*13 size, the upper left corner of face in the area
With the offset of two, lower right corner point and the regional center point;
S24, by the parsing to first convolutional layer and the calculated result of the second convolutional layer, finally obtain N number of one-dimension array
[score,x,y,width,height];
Wherein, x, y indicate that rectangle frame top left co-ordinate point, width indicate that rectangle width, height indicate rectangular elevation, score
Indicate that there are the probability of face in the region for the rectangle frame that [x, y, width, height] is represented.
6. the determination method of examinee's face direction of visual lines in a kind of Driving Test as claimed in claim 5, which is characterized in that described
Facial image is determined whether there is according to testing result to include the following steps:
Size between S31, comparison score value and the threshold value of setting;
If S32, score value are greater than or equal to threshold value, it is determined that the location information in array corresponding to the score value is identified
Rectangle frame in there are facial images;
If S33, score value are less than threshold value, it is determined that the rectangle that the location information in array corresponding to the score value is identified
Facial image is not present in frame.
7. the determination method of examinee's face direction of visual lines in a kind of Driving Test as claimed in claim 3, which is characterized in that described pre-
The step of constructing face direction of visual lines disaggregated model based on deep learning include:
A21, face direction of visual lines basis of classification network structure is built;
A22, the facial image marked in every interior image pattern by rectangle frame is obtained;
A23, N number of face direction of visual lines label is pre-seted, each face direction of visual lines label represents a kind of face direction of visual lines;
A24, the facial image is marked using face direction of visual lines label;
A25, using each facial image and corresponding face direction of visual lines label as one group of training data, form people
Face direction of visual lines classification based training data set;
A26, face direction of visual lines classification based training data set training face direction of visual lines basis of classification network structure, are used
Facial image Pixel Information is practised to the mapping relations of face direction of visual lines information, each of updates and saves deep neural network
Network layer parameter obtains the face direction of visual lines disaggregated model based on deep learning.
8. the determination method of examinee's face direction of visual lines in a kind of Driving Test as claimed in claim 7, which is characterized in that the people
Face direction of visual lines includes 8 classes, be respectively as follows: observation left side car door direction, observation left-hand mirror direction, observation inside rear-view mirror direction,
Scope dial side, observation right side car door direction, observation right rear view mirror direction, observation front and observation gear direction;
The face direction of visual lines basis of classification network structure is by 8 convolutional layers, 7 PRelu active coatings, 3 pond layers and 1
The network structure of a Dropout layers of composition.
9. the determination method of examinee's face direction of visual lines, feature exist in a kind of Driving Test as claimed in any one of claims 1 to 8
In the determination method further include: face direction of visual lines categorizing information is uploaded to upper layer points-scoring system, to assist examiner to make
Scoring judgement.
10. the decision-making system of examinee's face direction of visual lines in a kind of Driving Test characterized by comprising memory is configured as depositing
Store up data and computer program instructions;
The processor communicated is established with memory, the processor executes the computer program instructions and requires 1 with perform claim
The step of to any one of 9 determination method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910475340.2A CN110276276A (en) | 2019-06-03 | 2019-06-03 | The determination method and system of examinee's face direction of visual lines in a kind of Driving Test |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910475340.2A CN110276276A (en) | 2019-06-03 | 2019-06-03 | The determination method and system of examinee's face direction of visual lines in a kind of Driving Test |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110276276A true CN110276276A (en) | 2019-09-24 |
Family
ID=67960399
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910475340.2A Pending CN110276276A (en) | 2019-06-03 | 2019-06-03 | The determination method and system of examinee's face direction of visual lines in a kind of Driving Test |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110276276A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110705509A (en) * | 2019-10-16 | 2020-01-17 | 上海眼控科技股份有限公司 | Face direction recognition method and device, computer equipment and storage medium |
WO2023272635A1 (en) * | 2021-06-30 | 2023-01-05 | 华为技术有限公司 | Target position determining method, determining apparatus and determining system |
-
2019
- 2019-06-03 CN CN201910475340.2A patent/CN110276276A/en active Pending
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110705509A (en) * | 2019-10-16 | 2020-01-17 | 上海眼控科技股份有限公司 | Face direction recognition method and device, computer equipment and storage medium |
WO2023272635A1 (en) * | 2021-06-30 | 2023-01-05 | 华为技术有限公司 | Target position determining method, determining apparatus and determining system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107229930B (en) | Intelligent identification method for numerical value of pointer instrument | |
CN106128115B (en) | A kind of fusion method based on twin camera detection Traffic Information | |
CN110232389A (en) | A kind of stereoscopic vision air navigation aid based on green crop feature extraction invariance | |
CN110850723B (en) | Fault diagnosis and positioning method based on transformer substation inspection robot system | |
CN110142785A (en) | A kind of crusing robot visual servo method based on target detection | |
CN109460709A (en) | The method of RTG dysopia analyte detection based on the fusion of RGB and D information | |
CN110097109A (en) | A kind of road environment obstacle detection system and method based on deep learning | |
CN109506628A (en) | Object distance measuring method under a kind of truck environment based on deep learning | |
CN107844797A (en) | A kind of method of the milking sow posture automatic identification based on depth image | |
CN109284674A (en) | A kind of method and device of determining lane line | |
CN108596087B (en) | Driving fatigue degree detection regression model based on double-network result | |
CN106886216A (en) | Robot automatic tracking method and system based on RGBD Face datections | |
WO2020029444A1 (en) | Method and system for detecting attention of driver while driving | |
CN101726498B (en) | Intelligent detector and method of copper strip surface quality on basis of vision bionics | |
CN105868690A (en) | Method and apparatus for identifying mobile phone use behavior of driver | |
WO2023155483A1 (en) | Vehicle type identification method, device, and system | |
CN110189375A (en) | A kind of images steganalysis method based on monocular vision measurement | |
CN107452015A (en) | A kind of Target Tracking System with re-detection mechanism | |
CN105022999A (en) | Man code company real-time acquisition system | |
CN109948433A (en) | A kind of embedded human face tracing method and device | |
CN110633671A (en) | Bus passenger flow real-time statistical method based on depth image | |
CN110276276A (en) | The determination method and system of examinee's face direction of visual lines in a kind of Driving Test | |
CN106570440A (en) | People counting method and people counting device based on image analysis | |
CN110135277A (en) | A kind of Human bodys' response method based on convolutional neural networks | |
CN111709365A (en) | Automatic human motion posture detection method based on convolutional neural network |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190924 |
|
WD01 | Invention patent application deemed withdrawn after publication |