CN104978570A - Incremental learning based method for detecting and identifying traffic sign in traveling video - Google Patents

Incremental learning based method for detecting and identifying traffic sign in traveling video Download PDF

Info

Publication number
CN104978570A
CN104978570A CN201510359501.3A CN201510359501A CN104978570A CN 104978570 A CN104978570 A CN 104978570A CN 201510359501 A CN201510359501 A CN 201510359501A CN 104978570 A CN104978570 A CN 104978570A
Authority
CN
China
Prior art keywords
traffic sign
detection
training
detecting device
target
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201510359501.3A
Other languages
Chinese (zh)
Other versions
CN104978570B (en
Inventor
袁媛
王�琦
熊志同
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Northwestern Polytechnical University
Original Assignee
Northwestern Polytechnical University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Northwestern Polytechnical University filed Critical Northwestern Polytechnical University
Priority to CN201510359501.3A priority Critical patent/CN104978570B/en
Publication of CN104978570A publication Critical patent/CN104978570A/en
Application granted granted Critical
Publication of CN104978570B publication Critical patent/CN104978570B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Multimedia (AREA)
  • Image Analysis (AREA)
  • Traffic Control Systems (AREA)
  • Image Processing (AREA)

Abstract

The invention discloses an incremental learning based method for detecting and identifying a traffic sign in a traveling video, and is used for solving the technical problem of poor robustness in an existing traffic sign detection and identification method. The technical scheme is as follows: polymerization channel characteristics are adopted for training an Adaboost classifier, and a detector is improved; then a detection result of the detector serves as an observation value of a Kalman filter to perform motion model based tracing; in the tracing process, a new incremental SVM detector is trained on line; when the detection of an original Adaboost detector is failed due to apparent change of the sign, the online incremental detector is used for carrying out detection, the detection result serves as the observation value of the Kalman filter to be input, and targets incapable of being continuously detected are filtered; and finally, a tracing result of the same physical traffic sign is subjected to weighted voting of scale based Gaussian weight, a final identification result is obtained, and the detection and identification robustness are improved.

Description

Based on the detection and Identification method of traffic sign in the driving video of incremental learning
Technical field
The present invention relates to a kind of detection and Identification method of traffic sign, particularly relate to a kind of detection and Identification method based on traffic sign in the driving video of incremental learning.
Background technology
Document " Andreas dongran Liu, Mohan M.Trivedi, Traffic Sign Detectionfor U.S.Roads:Remaining Challenges and a case for Tracking.IEEE IntelligentTransportation Systems Conference, pp.1394-1399.2014. " in adopt cascade classifier Adaboost method to detect.This article selects integrating channel feature to carry out feature extraction to input training data, then trains the Adaboost cascade classifier of 3 layers, and detects inputting video data by the method for moving window.This method for traffic sign detection too relies on training data and feature extraction mode, easily occurs flase drop for situation about not occurring in training data, or shines, the change such as to block and cause undetected due to target light.And this detection method does not utilize the traffic sign of driving video to occur that the prior imformation of position is improved detecting device.The document carries out the tracking based on motion model to the traffic sign employing kalman filter method detected, by the observed reading of testing result as Kalman filter, thus estimates the next frame position of traffic sign.And this track algorithm only utilizes motion model information, once in motion model changes suddenly will cause tracking drift.And cannot process traffic sign illumination, the tracking failure problem changing and cause such as to block.For the identification of traffic sign, existing method does not consider that the recognition result of the different scale size of traffic sign in tracing process is to the Weight of the ballot of net result, can not improve the robustness of Traffic Sign Recognition.
Summary of the invention
In order to overcome the deficiency of the detection and Identification method poor robustness of existing traffic sign, the invention provides a kind of detection and Identification method based on traffic sign in the driving video of incremental learning.The method adopts converging channels features training Adaboost sorter, and by the spatial prior distribution of the traffic sign of video of driving a vehicle, detecting device is improved, then using tracking that the testing result of this detecting device is carried out based on motion model as the observed reading of Kalman filter, simultaneously, in tracing process, train new increment SVM detecting device online, when former Adaboost detecting device causes detecting failed due to the apparent change indicated, detected by this online incremental detector, and the observed reading of testing result as Kalman filter is inputted, reach and utilize motion model and apparent model to carry out the effect of following the tracks of simultaneously.In tracing process, the target of failing to be consecutively detected is filtered, improve the reliability of tracking results.Finally the tracking results of same physics traffic sign is carried out to the Nearest Neighbor with Weighted Voting of the Gauss's weight based on yardstick, obtain the final output that final recognition result is system.For the apparent state change due to target in traffic sign tracing process, or the problem such as the sudden change of motion model, the inventive method adopts the on-line checkingi device based on incremental learning, when the state change of target, in the enterprising Mobile state adjustment of detecting device model, and with this testing result, former motion model is upgraded, the final robustness improving detection and Identification.
The technical solution adopted for the present invention to solve the technical problems is: a kind of detection and Identification method based on traffic sign in the driving video of incremental learning, is characterized in adopting following steps:
Step one, to training dataset normalization, and sampled images block, calculates channel characteristics pond, use Adaboost algorithm training detecting device model, this detecting device is obtained by 4 layers strong detecting device cascades, and these 4 layers of strong detecting devices comprise 32 respectively, 128,512,2048 Weak Classifiers.Be candidate target by the detection target of whole 4 layers of strong classifier, and the strong classifier model of last one deck is:
Wherein, α tfor training the weight of each Weak Classifier obtained, h tfor Weak Classifier, T is the number of Weak Classifier.
Step 2, use training data carry out off-line training to online increment SVM detecting device, complete initialization procedure.
Step 3, the positive sample position of training data is carried out to the Parzen window density Estimation of gaussian kernel, obtaining probability density function is:
Wherein, x is vector (x, y), i.e. the position coordinates of traffic sign appearance.N is the positive number of samples of training data, V nfor the window size of window function, h nfor the length of side of window. function is Gaussian function.That is:
Step 4, formula manipulation is below adopted to the strong classifier model in step one:
By the cross-validation experiments determination parameter lambda on training set.
Step 5, the detecting device model utilizing step 2 to determine detect input video, using testing result position as observed reading input card Thalmann filter, by renewal and the tracking of forecast period realization to traffic sign of Kalman filtering.
Step 6, the tracking results of step 5 carried out to the test of positive sample possibility, trial function is as follows:
L positive(f k)＝symm(f k)·p ph(f kmean,f k)·p pos(f k-1,f k) (5)
High the exporting as tracking results of positive sample possibility, and parameter and the incremental update on-line checkingi device of Kalman filtering is upgraded with this result.Wherein L positive(f k) for kth frame becomes the inverse of positive sample possibility size, namely symmetry is higher, and its value is less.Symm (f k) be the large small magnitude of symmetry of the target of kth frame, p ph(f kmean, f k) be the Euclidean distance of the perception Hash of kth frame target and the mean value of front k frame.P pos(f k-1, f k) be the Euclidean distance of the coordinate of kth frame target and k-1 frame target.On-line checkingi device is then utilized to carry out on-line checkingi at this position neighborhood for the frame of video that positive sample possibility is low, using result as Output rusults, simultaneously for upgrading Kalman filtering parameter.
The invention has the beneficial effects as follows: the method adopts converging channels features training Adaboost sorter, and by the spatial prior distribution of the traffic sign of video of driving a vehicle, detecting device is improved, then using tracking that the testing result of this detecting device is carried out based on motion model as the observed reading of Kalman filter, simultaneously, in tracing process, train new increment SVM detecting device online, when former Adaboost detecting device causes detecting failed due to the apparent change indicated, detected by this online incremental detector, and the observed reading of testing result as Kalman filter is inputted, reach and utilize motion model and apparent model to carry out the effect of following the tracks of simultaneously.In tracing process, the target of failing to be consecutively detected is filtered, improve the reliability of tracking results.Finally the tracking results of same physics traffic sign is carried out to the Nearest Neighbor with Weighted Voting of the Gauss's weight based on yardstick, obtain the final output that final recognition result is system.For the apparent state change due to target in traffic sign tracing process, or the problem such as the sudden change of motion model, the inventive method adopts the on-line checkingi device based on incremental learning, when the state change of target, in the enterprising Mobile state adjustment of detecting device model, and with this testing result, former motion model is upgraded, finally improve the robustness of detection and Identification.
Below in conjunction with the drawings and specific embodiments, the present invention is elaborated.
Accompanying drawing explanation
Fig. 1 is the process flow diagram of the detection and Identification method of traffic sign in the driving video that the present invention is based on incremental learning.
Fig. 2 is prior probability distribution curve in space in the inventive method.
Fig. 3 is the Traffic Sign Recognition accuracy curve that the inventive method obtains.
Fig. 4 is that the real shooting photo with illumination variation is blocked in the inventive method process.
Embodiment
With reference to Fig. 1-4.The detection and Identification method concrete steps that the present invention is based on traffic sign in the driving video of incremental learning are as follows:
Step 1, first, to training dataset normalization, and sampled images block, calculate channel characteristics pond, use Adaboost algorithm training detecting device model, this detecting device is obtained by 4 layers strong detecting device cascades, and these 4 layers of strong detecting devices comprise 32 respectively, 128,512,2048 Weak Classifiers.Be candidate target by the detection target of whole 4 layers of strong classifier, and the strong classifier model of last one deck is:
Wherein, α tfor training the weight of each Weak Classifier obtained, h tfor Weak Classifier, T is the number of Weak Classifier.
Step 2, uses training data to carry out off-line training to online increment SVM detecting device, completes initialization procedure.
Step 3, carries out the Parzen window density Estimation of gaussian kernel to the positive sample position of training data, obtaining probability density function is:
Wherein, x is vector (x, y), i.e. the position coordinates of traffic sign appearance.N is the positive number of samples of training data, V nfor the window size of window function, h nfor the length of side of window. function is Gaussian function.That is:
Step 4, adopts formula manipulation below to the model in step 1:
By the cross-validation experiments determination parameter lambda on training set.
Step 5, the detecting device model utilizing step 2 to determine detects input video, using testing result position as observed reading input card Thalmann filter, by renewal and the tracking of forecast period realization to traffic sign of Kalman filtering.
Step 6, carry out the test of positive sample possibility to the tracking results of step 5, trial function is as follows:
L positive(f k)＝symm(f k)·p ph(f kmean,f k)·p pos(f k-1,f k) (10)
High the exporting as tracking results of positive sample possibility, and parameter and the incremental update on-line checkingi device of Kalman filtering is upgraded with this result.Wherein L positive(f k) for kth frame becomes the inverse of positive sample possibility size, namely symmetry is higher, and its value is less.Symm (f k) be the large small magnitude of symmetry of the target of kth frame, p ph(f kmean, f k) be the Euclidean distance of the perception Hash of kth frame target and the mean value of front k frame.P pos(f k-1, f k) be the Euclidean distance of the coordinate of kth frame target and k-1 frame target.On-line checkingi device is then utilized to carry out on-line checkingi at this position neighborhood for the frame of video that positive sample possibility is low, using result as Output rusults, simultaneously for upgrading Kalman filtering parameter.
Effect of the present invention can be described further by following emulation experiment.
1. simulated conditions.
The inventive method is in i5-3470 3.2GHz CPU, internal memory 4G, WINDOWS7 operating system, use the emulation that MATLAB software carries out.
The data used in emulation are public data collection and the data set that gathers voluntarily.
2. emulate content.
First, the distribution of experimental verification spatial prior probabilities, for the raising of road traffic sign detection, shows experimental result by Fig. 2.
Secondly, the Traffic Sign Recognition accuracy Dependence Results obtained as shown in Figure 3.
Finally, Fig. 4 proves that method of the present invention blocks the validity with illumination variation for process.
As seen from Figure 2, along with parameter change, Detection accuracy significantly increases, and recall rate is substantially unaffected simultaneously, therefore can prove that the spatial prior probabilities distribution in the present invention can reduce flase drop effectively, improves detection perform.
Can see from Fig. 3 curve, the Gauss's Weight method based on yardstick in the inventive method can improve recognition accuracy effectively.
The inventive method is for traffic sign illumination variation and the validity of to a certain degree blocking as seen from Figure 4.

Claims (1)

1., based on a detection and Identification method for traffic sign in the driving video of incremental learning, it is characterized in that comprising the following steps:
Step one, to training dataset normalization, and sampled images block, calculates channel characteristics pond, use Adaboost algorithm training detecting device model, this detecting device is obtained by 4 layers strong detecting device cascades, and these 4 layers of strong detecting devices comprise 32 respectively, 128,512,2048 Weak Classifiers; Be candidate target by the detection target of whole 4 layers of strong classifier, and the strong classifier model of last one deck is:
H ( x ) = s i g n ( Σ t = 1 T α t h t ( x ) ) - - - ( 1 )
Wherein, α tfor training the weight of each Weak Classifier obtained, h tfor Weak Classifier, T is the number of Weak Classifier;
Step 2, use training data carry out off-line training to online increment SVM detecting device, complete initialization procedure;
Step 3, the positive sample position of training data is carried out to the Parzen window density Estimation of gaussian kernel, obtaining probability density function is:
Wherein, x is vector (x, y), i.e. the position coordinates of traffic sign appearance; N is the positive number of samples of training data, V nfor the window size of window function, h nfor the length of side of window; function is Gaussian function; That is:
Step 4, formula manipulation is below adopted to the strong classifier model in step one:
H ( x ) = s i g n ( Σ t = 1 T α t h t ( x ) + ( P s p a t i a l ( x , y ) - 1 2 ) · λ ) - - - ( 4 )
By the cross-validation experiments determination parameter lambda on training set;
Step 5, the detecting device model utilizing step 2 to determine detect input video, using testing result position as observed reading input card Thalmann filter, by renewal and the tracking of forecast period realization to traffic sign of Kalman filtering;
Step 6, the tracking results of step 5 carried out to the test of positive sample possibility, trial function is as follows:
L positive(f k)＝symm(f k)·p ph(f kmean,f k)·p pos(f k-1,f k) (5)
High the exporting as tracking results of positive sample possibility, and parameter and the incremental update on-line checkingi device of Kalman filtering is upgraded with this result; Wherein L positive(f k) for kth frame becomes the inverse of positive sample possibility size, namely symmetry is higher, and its value is less; Symm (f k) be the large small magnitude of symmetry of the target of kth frame, p ph(f kmean, f k) be the Euclidean distance of the perception Hash of kth frame target and the mean value of front k frame; p pos(f k-1, f k) be the Euclidean distance of the coordinate of kth frame target and k-1 frame target; On-line checkingi device is then utilized to carry out on-line checkingi at this position neighborhood for the frame of video that positive sample possibility is low, using result as Output rusults, simultaneously for upgrading Kalman filtering parameter.
CN201510359501.3A 2015-06-25 2015-06-25 The detection and recognition methods of traffic sign in driving video based on incremental learning Active CN104978570B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510359501.3A CN104978570B (en) 2015-06-25 2015-06-25 The detection and recognition methods of traffic sign in driving video based on incremental learning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510359501.3A CN104978570B (en) 2015-06-25 2015-06-25 The detection and recognition methods of traffic sign in driving video based on incremental learning

Publications (2)

Publication Number Publication Date
CN104978570A true CN104978570A (en) 2015-10-14
CN104978570B CN104978570B (en) 2018-10-23

Family

ID=54275055

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510359501.3A Active CN104978570B (en) 2015-06-25 2015-06-25 The detection and recognition methods of traffic sign in driving video based on incremental learning

Country Status (1)

Country Link
CN (1) CN104978570B (en)

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105787470A (en) * 2016-03-25 2016-07-20 黑龙江省电力科学研究院 Method for detecting power transmission line tower in image based on polymerization multichannel characteristic
CN107157717A (en) * 2016-03-07 2017-09-15 维看公司 Object detection from visual information to blind person, analysis and prompt system for providing
CN107220583A (en) * 2016-03-21 2017-09-29 伊莱比特汽车有限责任公司 Method and apparatus for recognizing traffic sign
CN108154102A (en) * 2017-12-21 2018-06-12 安徽师范大学 A kind of traffic sign recognition method
CN108351962A (en) * 2015-12-01 2018-07-31 英特尔公司 Object detection with adaptivity channel characteristics
CN108805187A (en) * 2018-05-29 2018-11-13 北京佳格天地科技有限公司 Celestial spectrum sequence automatic classification system and method
CN109993207A (en) * 2019-03-01 2019-07-09 华南理工大学 A kind of image method for secret protection and system based on target detection
CN110245565A (en) * 2019-05-14 2019-09-17 东软集团股份有限公司 Wireless vehicle tracking, device, computer readable storage medium and electronic equipment
CN110415270A (en) * 2019-06-17 2019-11-05 广东第二师范学院 A kind of human motion form evaluation method based on double study mapping increment dimensionality reduction models
CN110555466A (en) * 2019-08-13 2019-12-10 创新奇智(南京)科技有限公司 Cascade identification network algorithm capable of being dynamically increased
CN112686203A (en) * 2021-01-12 2021-04-20 重庆大学 Vehicle safety warning device detection method based on space prior

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101023436A (en) * 2004-08-16 2007-08-22 西门子共同研究公司 Method for traffic sign detection
CN102902955A (en) * 2012-08-30 2013-01-30 中国科学技术大学 Method and system for intelligently analyzing vehicle behaviour
CN103390167A (en) * 2013-07-18 2013-11-13 奇瑞汽车股份有限公司 Multi-characteristic layered traffic sign identification method

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101023436A (en) * 2004-08-16 2007-08-22 西门子共同研究公司 Method for traffic sign detection
CN102902955A (en) * 2012-08-30 2013-01-30 中国科学技术大学 Method and system for intelligently analyzing vehicle behaviour
CN103390167A (en) * 2013-07-18 2013-11-13 奇瑞汽车股份有限公司 Multi-characteristic layered traffic sign identification method

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
ANDREAS MØGELMOSE ETC，: ""Traffic Sign Detection for U.S. Roads: Remaining Challenges and a case for Tracking"", 《2014 IEEE 17TH INTERNATIONAL CONFERENCE ON INTELLIGENT TRANSPORTATION SYSTEMS (ITSC)》 *
萧嵘 等，: ""一种SVM增量学习算法"", 《南京大学学报(自然科学)》 *

Cited By (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108351962A (en) * 2015-12-01 2018-07-31 英特尔公司 Object detection with adaptivity channel characteristics
CN108351962B (en) * 2015-12-01 2022-05-10 英特尔公司 Object detection with adaptive channel features
CN107157717A (en) * 2016-03-07 2017-09-15 维看公司 Object detection from visual information to blind person, analysis and prompt system for providing
CN107220583A (en) * 2016-03-21 2017-09-29 伊莱比特汽车有限责任公司 Method and apparatus for recognizing traffic sign
CN107220583B (en) * 2016-03-21 2020-11-13 伊莱比特汽车有限责任公司 Method and device for recognizing traffic signs
CN105787470A (en) * 2016-03-25 2016-07-20 黑龙江省电力科学研究院 Method for detecting power transmission line tower in image based on polymerization multichannel characteristic
CN108154102A (en) * 2017-12-21 2018-06-12 安徽师范大学 A kind of traffic sign recognition method
CN108805187B (en) * 2018-05-29 2022-07-19 北京佳格天地科技有限公司 Astronomical spectrum sequence automatic classification system and method
CN108805187A (en) * 2018-05-29 2018-11-13 北京佳格天地科技有限公司 Celestial spectrum sequence automatic classification system and method
CN109993207A (en) * 2019-03-01 2019-07-09 华南理工大学 A kind of image method for secret protection and system based on target detection
CN109993207B (en) * 2019-03-01 2022-10-25 华南理工大学 Image privacy protection method and system based on target detection
CN110245565A (en) * 2019-05-14 2019-09-17 东软集团股份有限公司 Wireless vehicle tracking, device, computer readable storage medium and electronic equipment
US11106924B2 (en) 2019-05-14 2021-08-31 Neusoft Corporation Vehicle tracking method, computer readable storage medium, and electronic device
CN110415270A (en) * 2019-06-17 2019-11-05 广东第二师范学院 A kind of human motion form evaluation method based on double study mapping increment dimensionality reduction models
CN110415270B (en) * 2019-06-17 2020-06-26 广东第二师范学院 Human motion form estimation method based on double-learning mapping incremental dimension reduction model
CN110555466A (en) * 2019-08-13 2019-12-10 创新奇智(南京)科技有限公司 Cascade identification network algorithm capable of being dynamically increased
CN112686203A (en) * 2021-01-12 2021-04-20 重庆大学 Vehicle safety warning device detection method based on space prior
CN112686203B (en) * 2021-01-12 2023-10-31 重庆大学 Vehicle safety warning device detection method based on space priori

Also Published As

Publication number Publication date
CN104978570B (en) 2018-10-23

Similar Documents

Publication Publication Date Title
CN104978570A (en) Incremental learning based method for detecting and identifying traffic sign in traveling video
Jensen et al. Evaluating state-of-the-art object detector on challenging traffic light data
Li et al. Road network extraction via deep learning and line integral convolution
CN112200161A (en) Face recognition detection method based on mixed attention mechanism
CN106599827A (en) Small target rapid detection method based on deep convolution neural network
CN105320966A (en) Vehicle driving state recognition method and apparatus
CN106096561A (en) Infrared pedestrian detection method based on image block degree of depth learning characteristic
CN105513354A (en) Video-based urban road traffic jam detecting system
CN103455797A (en) Detection and tracking method of moving small target in aerial shot video
CN103150740A (en) Method and system for moving target tracking based on video
CN102945374B (en) Method for automatically detecting civil aircraft in high-resolution remote sensing image
CN101281648A (en) Method for tracking dimension self-adaption video target with low complex degree
CN103593672A (en) Adaboost classifier on-line learning method and Adaboost classifier on-line learning system
CN103593679A (en) Visual human-hand tracking method based on online machine learning
Li et al. Vehicle classification with single multi-functional magnetic sensor and optimal MNS-based CART
CN105654516A (en) Method for detecting small moving object on ground on basis of satellite image with target significance
CN110991397B (en) Travel direction determining method and related equipment
Lan et al. Vehicle detection and recognition based on a MEMS magnetic sensor
CN104463907A (en) Self-adaptation target tracking method based on vision saliency characteristics
CN106887012A (en) A kind of quick self-adapted multiscale target tracking based on circular matrix
Pyo et al. Front collision warning based on vehicle detection using CNN
Wang et al. Vehicle reidentification with self-adaptive time windows for real-time travel time estimation
CN104268574A (en) SAR image change detecting method based on genetic kernel fuzzy clustering
CN104268584A (en) Human face detection method based on hierarchical filtration
CN104463909A (en) Visual target tracking method based on credibility combination map model

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant