CN104915000A - Multisensory biological recognition interaction method for naked eye 3D advertisement - Google Patents
Multisensory biological recognition interaction method for naked eye 3D advertisement Download PDFInfo
- Publication number
- CN104915000A CN104915000A CN201510282106.XA CN201510282106A CN104915000A CN 104915000 A CN104915000 A CN 104915000A CN 201510282106 A CN201510282106 A CN 201510282106A CN 104915000 A CN104915000 A CN 104915000A
- Authority
- CN
- China
- Prior art keywords
- feature
- human
- advertisement
- classification
- crowd
- 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
Landscapes
- Image Analysis (AREA)
Abstract
The invention relates to a multisensory biological recognition interaction method for a naked eye 3D advertisement. The method comprises the steps of 1, detection of a human body in a scene, wherein in the process of building a platform automatically delivering targeted advertisement content, the moving human body in the fixed scene is monitored and detected through a camera; 2, feature detection of crowds, wherein the human face features, hair style features and morphological features of the crowds are extracted, and used for distinguishing the ages and sexes of different crowds; 3, intelligent recognition and classification of the crowds, wherein the extracted human face features and morphological features of the crowds are utilized for intelligently recognizing and classifying the sexes and ages of the crowds; 4, advertisement recommending delivering deciding, wherein different advertisements are classified to be delivered according to the classification result of the crowds, and different information is transmitted to the crowds to be specified based on a personalized recommending technology; 5, intelligent human-computer interaction, wherein based on a somatosensory technology and a speech recognition technology, captured human movement and speech signals are converted into control signals, and intelligent human-computer interaction of human beings and an advisement player is carried out.
Description
Technical field
The invention belongs to Intelligent Recognition field, particularly a kind of many perception bio-identification exchange method for bore hole 3D advertisement.
Background technology
3D display technique, particularly bore hole 3D technology and product are promoting one turns to stereo display technological revolution from flat pannel display.International research is like a raging fire, as Sony, Samsung, LG company have released bore hole 3D product.Current China 3D industry is still at an early stage of development, bore hole 3D industry is also at the early-stage, there have been many enterprises actively to put in bore hole 3D technology and research and development of products although domestic, but China bore hole 3D technical market development time is short, and there is no the scientific research institution that specialty is regular, simultaneously because stereoscopic imaging technology is a kind of overlapping edges science and technology, the people of research is few, on market, bore hole 3D product price is higher, and market is in education and breeding phase.In the bore hole 3D equipment gone on the market at present, parallax barrier formula 3D technology is more common type, and this technology is used on 3D mobile phone (Sharp SH8168U), portable game machine (Nintendo 3DS) this kind of undersized equipment more.56 inch 3D TVs of Toshiba then employ lens pillar technology, because the requirement of televisor to brightness, resolution is higher, the former is also difficult to competent harsher applied environment.
Bore hole 3D advertisement machine is the characteristic utilizing people two to have parallax, when without any need for utility appliance (as 3D glasses, the helmet etc.), can obtain the lifelike stereoscopic image with space, the degree of depth.In picture, namely things can protrude from outside picture, also can deeply be hidden among picture.Beautiful in colour, well arranged, vivid, life-like, be 3 D stereoscopic image truly.Naked-eye stereoscopic image based is with its truly lively expressive force, and graceful graceful environmental infection power, the visual impact of strong shock is deeply by the favor of consumers in general.The audio and video products of this new, special, the strange technique of expression, are widely applied to multiple industries such as advertisement media, display and demonstration, tourism trade and investment promotion, wedding photo, research and teaching, Entertainment, commercial Application, architectural design, mobile phone.
Digital advertising machine (hereinafter referred to as advertisement machine) is a kind of new media concept, refer in market, airport and other people flow the public place of convergence, by giant-screen terminal presentation facility, issue the multimedia specialty audiovisual system of business, finance and economics or entertainment information.As the replacer of traditional outdoor advertising, the quantity of information that advertisement machine comprises is very large, can upgrade easily at any time, and can network, this be all traditional outdoor advertising incomparable.Advertisement machine not only can to public's issuing multimedia advertisement information, bring brand-new business innovative value, and be considered to the power that display industries newly grows up.Through the development of nearly 5 years, advertisement machine was passed by unit, network, several stage such as mutual, and progressively strides forward to intellectuality.Intel China chief inspector Shi Yangwei represents: " traditional, based on the content of Plane performance form and the attractive force of advertisement pushing to consumer more and more less, interactive broadcasting media and marketing mode are about to become main flow ".
The realization of applying for interactive broadcasting media, as " the 5th media ", is provided requisite technology platform by intelligentized bore hole 3D advertisement machine product." intelligentized bore hole 3D advertisement machine refers to that terminal presentation facility releases bore hole 3D advertising message targetedly automatically according to identified different consumers' characteristics.Based on the analysis to this trend, Intel is thought: intelligent bore hole 3D advertisement machine system is not only wanted to support that more clear, lively bore hole 3D picture display is to attract consumer, have more friendly touch screen interactive operation interface, also need to identify consumers' characteristics, the instant advertising message automatically releasing customization.The realization of these application needs the support of powerful processor technology, hardware platform and software solution.
Along with the development of bore hole 3D advertisement machine Internet technology, new application and the growth needing mass data to be processed thereof, the bore hole 3D advertisement machine processing scheme based on " intelligent interaction platform " seems particularly important.Hardware resource for intelligent bore hole 3D advertisement machine can meet the demand of different application pattern, simultaneously, the hardware such as the controller developed for bore hole 3D advertisement machine industry and display are also increasingly abundant, and relatively weak link is corresponding software solution, i.e. intelligent identification technology.Although report some intelligent identification technologies in the world at present, they are in laboratory development mostly, and consider who object and environment also more single.But, in fact the application of bore hole 3D advertisement machine is completely open complex environment, service object and environment are all extremely uncertain: enter crowd's quantity of camera coverage, Sex, Age distributes, stand or walking postures etc. all cannot retrain, environment before video camera, as vehicle walk, light and shade change, weather effect etc. be also uncontrollable.Everything brings difficulty all to human detection, tracking, identification.In addition, the many face features based on personage of more existing recognition technologies, minority is based on hair color and hearing feature, but the above-mentioned characteristic of bore hole 3D advertisement machine applied environment makes the accurate detection of single features almost cannot realize, and therefore only can not rely on a few feature and make a policy.
In addition, current body sense technology, speech recognition technologies etc. support that human-computer interaction technology occurs in succession, by body propagated sensation sensor and voice collector, catch action or the voice of human body, by the information processing technology, be corresponding instruction by the work of human body or speech conversion, be used for controlling the broadcasting of intelligent advisement player and carry out corresponding man-machine interaction and have great importance.
Summary of the invention
The object of the invention is to the blank filling up prior art, provide a kind of and manifold extraction and fusion are had a many perception bio-identification exchange method being used for bore hole 3D advertisement of intelligent recommendation technology.
The technical solution adopted for the present invention to solve the technical problems is:
For many perception bio-identification exchange method of bore hole 3D advertisement, the steps include:
(1) the human detection in scene: put in the construction of ad content platform targetedly at automatic switching, by camera, inspection and monitoring is carried out to the movement human in fixed scene, movement human detects and comprises human detection and tracking, and namely crowd in video appears in detection and tracking; The number of people, namely in video, the number of crowd is added up;
(2) the feature detection of crowd: extract the face characteristic of crowd, hair style feature, sound characteristic and morphological feature, is used for distinguishing age and the sex of different crowd;
(3) crowd's Intelligent Recognition and classification: the feature utilizing the crowd extracted, carries out the Intelligent Recognition classification of sex and age bracket to crowd;
(4) advertisement is recommended to throw in decision-making: according to the result of listener clustering, different advertisements put on by classification, adopt personalized recommended technology, different information is passed to the crowd wanting to specify;
(5) intelligent human-machine interaction: by body sense technology, speech recognition technology changes the human action of catching and voice signal into control signal, carries out the human-computer intellectualization of people and advertisement machine.
And, step (1) described in human detection and follow the tracks of and adopt still image human detection, it is in feature selecting, adopt Haar-like feature, EOH feature, textural characteristics, color characteristic, in the design of sorter, employing be statistical learning classification method.
And, the number of the people that step is (1) described adopts and dynamically updates model, utilize Gaussian distribution to describe the probability distribution of background dot color, the detection of new movement human contrasts according to current frame image and the background model set up by mixed Gauss model, obtain all movement humans in the foreground image of present frame, then, then with the movement human data recorded compare.Like this, just obtain the target with kinetic property in many present frames, have overlapping target to abandon less moving target and tracked human body, and remaining part is sorted according to size, just newly add 10 maximum human bodies afterwards and continue to follow the trail of.
And the face characteristic that step is (2) described and morphological feature extraction are divided into global feature to extract and local shape factor two kinds of modes, and global feature extraction algorithm carries out rough identification to face, and local shape factor supplements from details.
And the (2) described hair style feature extraction key step of step is as follows:
A, from video, obtain a two field picture detect, time image acquisitions or detected in scene that timing acquisition list note image processes when having personage.
B, Image semantic classification.
C, hair style are split.
D, hair style model are formed.
And crowd's Intelligent Recognition that step is (3) described and classification adopt multi-categorizer to realize Decision fusion.
And the step of the algorithm of the sex identification that step is (3) described is as follows:
The various features such as a, the face extracting human body, the colour of skin, hair style, decoration, face characteristic comprises whole and part feature;
B, utilize linear discriminant method to extract the feature of human body of training and detecting, utilize Fisher method to extract the feature of training sample and target image simultaneously;
C, utilize dynamic cluster method to classify to target image, and calculate the average recognition rate of single classifier;
D, utilize men and women's training sample to train SVR, obtain one group of parameter value, utilize the SVR trained to classify to test sample book;
E, various fusion rule is utilized to carry out assembled classification, the nicety of grading of more different fusion rule.
And the step of the algorithm of the age bracket identification that step is (3) described is as follows:
A, the face extracting human body, the colour of skin, hair style, the various feature of form, to face, choose distinguishing ability stronger left and right eyes, nose and face as regional area, reduce computation complexity to a certain extent, feature point for calibration is partitioned into the left and right eyes of face, nose and face region;
B, utilize linear discriminant method to extract the feature of human body of training and detecting, utilize Fisher method to extract the feature of training sample and target image simultaneously;
C, utilize dynamic cluster method to classify to target image, and calculate the average recognition rate of single classifier;
D, utilize Different age group training sample to train SVR, obtain one group of parameter value, utilize the SVR trained to classify to test sample book;
E, various fusion rule is utilized to carry out assembled classification, the nicety of grading of more different fusion rule.Increase class categories according to the classification iteration of target image, reduce age bracket class interval, to improve age estimated accuracy.
And the (4) described advertisement of step is recommended to throw in decision-making and is adopted demographic recommended technology, and advertisement putting adopts manual type classification.
Advantage of the present invention and good effect are:
The present invention is directed to intelligent bore hole 3D advertisement machine, by to manifold extraction of the target group under complex environment and fusion, the intelligent classification finally realizing the advertisement of bore hole 3D advertisement machine with throw in decision-making and the man-machine many perception intelligent interaction based on recognition of face, body sense technology and speech recognition technology.Thus by the extraction of multiple features, calculating and classification, important data basis is provided to the input decision-making of bore hole 3D advertisement machine, has filled up the blank of intelligent identification technology in bore hole 3D advertisement machine application.
The present invention is by body propagated sensation sensor and voice collector, catch action or the voice signal of human body, by the information processing technology, the action of human body or voice signal are converted to corresponding instruction, be used for control intelligent advisement player broadcasting and carry out corresponding man-machine interaction.Thus recognition of face, body sense technology are merged mutually with speech recognition technology, realize the application of many perception biological identification technology in intelligent advisement player, can be different colony's services, attract the notice of client.
Accompanying drawing explanation
Fig. 1 is many perception bio-identification exchange method process flow diagram of the present invention;
Fig. 2 is the human detection block diagram based on still image;
Fig. 3 is the human detection block diagram based on video;
Fig. 4 is the disposal system block diagram of voice signal;
Fig. 5 is the calculation block figure of MFCC characteristic parameter;
Fig. 6 is the Intelligent Recognition block diagram of sex, age bracket.
Embodiment
Below by specific embodiment, the invention will be further described, and following examples are descriptive, is not determinate, can not limit protection scope of the present invention with this.
For many perception bio-identification exchange method of bore hole 3D advertisement, the steps include:
(1) the human detection in complex scene: put in the construction of ad content platform targetedly at automatic switching, by camera, inspection and monitoring is carried out to the movement human in fixed scene, be the basis automatically identifying sex, age bracket etc. further, mainly contain following two sub-contents of research:
Human detection and tracking: crowd in video appears in detection and tracking, can carry out the detection and tracking of multiple target simultaneously,
The number of people: the number statistics of crowd in video.
Still image human detection, problem is actually the pattern recognition problem of a standard as shown in Figure 1, two wherein the most key broad aspect contents are: the selection of feature and the design of sorter, in feature selecting, adopt Haar-like feature, EOH feature, textural characteristics, color characteristic etc., in the design of sorter, what adopt is the method that statistical learning is classified, consider to use the Adaboost learning algorithm being successfully applied to Face datection field, this algorithm feature select with classifier design on achieve gratifying effect, for human detection and tracking, the present invention is based on the movement human detection and tracking of video, the scene background General Transformations of camera is little, this brings great convenience to movement human detection and tracking, greatly reduced the complexity of detection and tracking simultaneously.
Based on video human detection block diagram as shown in Figure 2, detect before movement human at process video, utilize the movable information in video image to mark its moving region, remove the region do not met the demands.Then the sorter trained is used to detect to obtain final movement human example to the region screened.Video image is mainly divided into the human body foreground area of motion by foreground detection, with the background area of static part in camera.Background updating is that static background sets up background model, by current image frame and background model being compared, determines that foreground area is thought in the region changed greatly namely.The computing velocity of this method is very fast, the complete accurate description about motion target area can be obtained, but more responsive to illumination condition, large area motion and noise ratio in scene, so need in actual applications to adopt certain algorithm to carry out the change dynamically updating to conform of background model.
The present invention adopts and dynamically updates model is the probability distribution describing background dot color by Gaussian distribution (normal distribution).The detection of new movement human contrasts according to current frame image and the background model set up by mixed Gauss model, obtains all movement humans in the foreground image of present frame.Next, then compare with the movement human data recorded.Like this, just obtain the target with kinetic property in many present frames, overlapping target is had to abandon with tracked human body less moving target (may be caused by noise), and remaining part is sorted according to size, just newly add 10 maximum human bodies afterwards to continue to follow the trail of, the number for crowd in video is added up.
(2) the feature detection of crowd: extract some features of crowd, be used for distinguishing age and the sex of different crowd, these features can comprehensively use, and to improve accuracy of identification and the reliability of system, conventional feature has following several:
A, face characteristic: the entirety of face or local feature.
B, hair style feature: mainly refer to the hair style such as length hair and shaven head.
C, sound characteristic: voice, cough, pant, footsteps etc.
B, morphological feature: the height of human body, the feature of crucial articulation point.
Face characteristic and morphological feature extraction can be divided into global feature to extract and local shape factor two kinds of modes, global feature extraction algorithm carries out rough identification to face, local shape factor supplements from details, but when identifying local feature (as shaven head, the whiskers etc.) of some uniqueness, then these local features can directly be used for determining sex, and global feature and local shape factor complement each other and make recognition of face obtain better effect.
In addition, Fusion Features also successfully can be incorporated in local shape factor, by repeatedly processing piece image, as processed with local binary patterns, can obtain several characteristic images, then carrying out merging and feature extraction, carrying out face coupling the most afterwards.
Four-dimensional local binary patterns feature extraction is a new method improved based on local binary patterns, and local binary patterns only considers the partial structurtes relation of the size of central pixel point and field pixel, well describes the Local textural feature of image.But three-dimensional local binary patterns but have ignored the impact of pixel value for local detail feature of the central point of image own, four-dimensional local binary patterns feature is exactly the pixel value adding central point on the basis of three-dimensional local binary patterns feature, and consideration gradient information is taken into account, form a new fusion local binary patterns feature.
Wherein, for the extraction of hair style feature,
The key step of hair style feature extraction is as follows:
The first step: obtain a two field picture and detect from video.Can time image acquisitions or detected in scene that timing acquisition list note image processes when having personage.
Second step: Image semantic classification.The effect of this step process directly affects the effect of hair style model extraction below.Hair style extract with face extraction some distinguish, the rigidity of faceform is stronger, and the reflectance of face comparatively hair is low, namely the susceptibility of face to illumination is lower than hair.Therefore, the process of this step is most important.
3rd step: hair style is split.This step is the difficult point of whole process, and mainly the flexibility of hair is larger, and its shape is not fixed, and comparatively large with illumination variation impact, no matter is intensity of illumination or direction of illumination.
4th step: hair style model is formed.After splitting, can extract hair style model from segmentation figure, the hair style model of taking-up is used for sex or Age estimation.
For the extraction of sound characteristic,
The feature extraction of sound occupies very important effect, has several link that must experience to comprise digitizing, pre-filtering before recognition system, sampling, and A/D converts, and the pre-service (comprising framing and windowing) of voice signal, idiographic flow as shown in Figure 4.
What the feature extracting method for sound adopted is Mel Frequency Cepstral parameter (MFCC), this method takes the auditory properties of people's ear into account, namely in the middle of the non-linear spectrum region frequency spectrum of sound being transformed into Mel frequency marking, and then be transformed on spectrum domain by certain Homomorphic Processing process, its calculation flow chart is as shown in Figure 5.
Compared with the sound characteristic extracting method that other are traditional, MFCC parameter has good recognition performance and noise resisting ability, and this parameter is compared than the LPCC based on channel model and is had better Shandong nation property, more meet the auditory properties of people's ear, and when signal to noise ratio (S/N ratio) reduces, still there is good recognition performance.
As for the method for classifying to phonetic feature, we carry out Classification and Identification by SVM (support vector machine), by will obtained MFCC parameter transmission be trained to SVM classifier, then higher dimensional space is projected to, them are made to become linear separability, the principle of recycling linear partition judges classification boundaries, finally reaches the object of speech recognition.
(3) crowd's Intelligent Recognition and classification: utilize the crowd characteristic extracted, to the Intelligent Recognition classification that crowd carries out sex and age bracket, utilizes the method for statistics or artificial intelligence and so on to carry out sex and age bracket discriminator to the crowd characteristic extracted.
Sex identification: men and women is classified.
Age bracket identification: old, middle age, teenager are distinguished and classified.
Adopting sorter to carry out character classification by age to the human body detected is main method about change of age research in field of face identification, and such age estimation method is effective.But because single sorter has its respective relative merits, thus the recognition result of different sorter may differ very large under some circumstances.Multi-categorizer Decision fusion is a kind of trend of mode development, and its object is intended to improve nicety of grading, and the dense degree of measurement pattern.
Fisher linear discriminant analysis.As single classifier, Fisher linear discriminant analysis recognition performance is all good on different databases.But utilize Fisher linear discriminant analysis when carrying out dimension-reduction treatment, still can lose some useful informations, and these information may be very important for later step, this is also the weak point of the method.
Dynamic Clustering Algorithm.Dynamic Clustering Algorithm adopts C-mean algorithm, and it is a kind of conventional unsupervised learning method based on neighbour's rule.Its basic thought is very simple, first determines the group's number c needed, chooses c representative point, with these representative points as starting type, then finds out at a distance of nearest representative point sample X each in sample set H, is grouped into by X in the group at this nearest representative point place and goes.Like this, H is just tentatively divided into c group by neighbour's rule by iteration for the first time.H, just on this basis using the mean vector of each group of last iteration gained as new representative point, is divided into c group by neighbour's rule by next iteration again, until it is stable to hive off.
Support vector regression (Support Vector Regression is called for short SVR).Support vector regression algorithm is the popularization of support vector machine method on regression problem.By introducing insensitive loss function and kernel function, can nonlinear regression analysis be advantageously applied to, and to small sample size problem, there is good estimated performance.
On the basis of extraction feature noted earlier, we can use sorter to carry out the sex and age identification of people, as shown in Figure 6.Introduce detailed algorithm below:
1. the key step of the algorithm of sex identification is as follows:
The method of a, basis introduction above extracts the various feature such as face, the colour of skin, hair style, decoration of human body.Face characteristic comprises whole and part feature.
B, utilize linear discriminant method to extract the feature of human body of training and detecting, utilize Fisher method to extract the feature of training sample and target image simultaneously.
C, utilize dynamic cluster method to classify to target image, and calculate the average recognition rate of single classifier.
D, utilize men and women's training sample to train SVR, obtain one group of parameter value.The SVR trained is utilized to classify to test sample book.
E, various fusion rule is utilized to carry out assembled classification, the nicety of grading of more different fusion rule.
2. the key step of the algorithm of age bracket identification is as follows:
The method of a, basis introduction above extracts the various feature such as face, the colour of skin, hair style, decoration of human body.To face, choose distinguishing ability stronger left and right eyes, nose and face as regional area, reduce computation complexity to a certain extent.Feature point for calibration is partitioned into the left and right eyes of face, nose and face region.
B, utilize linear discriminant method to extract the feature of human body of training and detecting, utilize Fisher method to extract the feature of training sample and target image simultaneously.
C, utilize dynamic cluster method to classify to target image, and calculate the average recognition rate of single classifier.
D, utilize Different age group (old, in, few etc.) training sample to train SVR, obtain one group of parameter value.The SVR trained is utilized to classify to test sample book.
E, various fusion rule is utilized to carry out assembled classification, the nicety of grading of more different fusion rule.Increase class categories according to the classification iteration of target image, reduce age bracket class interval, to improve age estimated accuracy.(4) advertisement is recommended to throw in decision-making: according to the result of listener clustering, different advertisements put on by classification.Adopt personalized recommended technology, at reasonable time, appropriate information is passed to the crowd wanting to transmit, make advertisement expect to reach better effect.
Adopt demographic recommended technology (Demographic-based), the most frequently used demographic characteristics's variable of this technology is age, sex and geographic position, sometimes takes in variable and also can be used (client as in high-grade hotel).Age and sex are obtained by system identification, and geographic position is determined by system terminal location.
The advertisement starting stage is classified by manual type, and marks suitable population's attribute (age, sex, geographic position and income etc.).If likely, system increases observed quantity: user is to the viewing time of a certain advertisement.Utilize this time, obtain the favorable rating of certain class crowd to each advertisement, to adjust the broadcast strategy of each advertisement.
(5) intelligent human-machine interaction: by body sense technology, speech recognition technology etc. change the human action of catching and voice signal into control signal, carries out the human-computer intellectualization of advertisement machine.By face recognition technology, the identity of people, age and sex can be identified, then play different advertisements for different crowds, such as, be the advertisement of child's playing animation, for Ms plays advertisement for cosmetics, realize man-machine interaction, to a certain extent can attracting notice.
By body propagated sensation sensor and voice collector, catch action or the voice signal of human body, by the information processing technology, the action of human body or voice signal are converted to corresponding instruction, be used for control intelligent advisement player broadcasting and carry out corresponding man-machine interaction.
Wherein, can body propagated sensation sensor be passed through, catch the actuating signal of human body, as left, to the right, upwards, downwards, and palm closes and the signal such as to close, and passes through signal transacting, the actuating signal of human body is converted to the control signal of intelligent advisement player, controls the broadcasting of advertisement video, suspend, F.F., rewind, the replacing of film source, screen magnifying, the function such as to reduce, this technology can attract the advertisement colony of child and youth.
In addition, by speech recognition technology, by the voice of people as started, suspending, playing, F.F., rewind, the replacing of film source, screen magnifying, the voice signal such as to reduce, be converted to corresponding control signal, control the broadcasting of advertisement video.This technology will facilitate the disadvantaged group such as disabled person greatly.
Thus recognition of face, body sense technology are merged mutually with speech recognition technology, realize the application of many perception biological identification technology in intelligent advisement player, can be different colony's services, attract the notice of client.
Claims (9)
1., for many perception bio-identification exchange method of bore hole 3D advertisement, it is characterized in that: the steps include:
(1) the human detection in scene: put in the construction of ad content platform targetedly at automatic switching, by camera, inspection and monitoring is carried out to the movement human in fixed scene, movement human detects and comprises human detection and tracking, and namely crowd in video appears in detection and tracking; The number of people, namely in video, the number of crowd is added up;
(2) the feature detection of crowd: extract the face characteristic of crowd, hair style feature, sound characteristic and morphological feature, is used for distinguishing age and the sex of different crowd;
(3) crowd's Intelligent Recognition and classification: the feature utilizing the crowd extracted, carries out the Intelligent Recognition classification of sex and age bracket to crowd;
(4) advertisement is recommended to throw in decision-making: according to the result of listener clustering, different advertisements put on by classification, adopt personalized recommended technology, different information is passed to the crowd wanting to specify;
(5) intelligent human-machine interaction: by body sense technology, speech recognition technology changes the human action of catching and voice signal into control signal, carries out the human-computer intellectualization of people and advertisement machine.
2. the many perception bio-identification exchange method for bore hole 3D advertisement according to claim 1, it is characterized in that: step (1) described in human detection and follow the tracks of and adopt still image human detection, it is in feature selecting, adopt Haar-like feature, EOH feature, textural characteristics, color characteristic, in the design of sorter, employing be statistical learning classification method.
3. the many perception bio-identification exchange method for bore hole 3D advertisement according to claim 1, it is characterized in that: the number of the people that step is (1) described adopts and dynamically updates model, utilize Gaussian distribution to describe the probability distribution of background dot color, the detection of new movement human contrasts according to current frame image and the background model set up by mixed Gauss model, obtain all movement humans in the foreground image of present frame, then, compare with the movement human data recorded again, obtain the target with kinetic property in many present frames thus, overlapping target is had to abandon less moving target and tracked human body, and remaining part is sorted according to size, just newly add 10 maximum human bodies afterwards to continue to follow the trail of.
4. the many perception bio-identification exchange method for bore hole 3D advertisement according to claim 1, it is characterized in that: the face characteristic that step is (2) described and morphological feature extraction are divided into global feature to extract and local shape factor two kinds of modes, global feature extraction algorithm carries out rough identification to face, and local shape factor supplements from details.
5. the many perception bio-identification exchange method for bore hole 3D advertisement according to claim 1, is characterized in that: the (2) described hair style feature extraction key step of step is as follows:
A, from video, obtain a two field picture detect, time image acquisitions or detected in scene that timing acquisition list note image processes when having personage.
B, Image semantic classification.
C, hair style are split.
D, hair style model are formed.
6. the many perception bio-identification exchange method for bore hole 3D advertisement according to claim 1, is characterized in that: crowd's Intelligent Recognition that step is (3) described and classification adopt multi-categorizer to realize Decision fusion.
7. the many perception bio-identification exchange method for bore hole 3D advertisement according to claim 6, is characterized in that: the step of the algorithm of the sex identification that step is (3) described is as follows:
The various features such as a, the face extracting human body, the colour of skin, hair style, decoration, face characteristic comprises whole and part feature;
B, utilize linear discriminant method to extract the feature of human body of training and detecting, utilize Fisher method to extract the feature of training sample and target image simultaneously;
C, utilize dynamic cluster method to classify to target image, and calculate the average recognition rate of single classifier;
D, utilize men and women's training sample to train SVR, obtain one group of parameter value, utilize the SVR trained to classify to test sample book;
E, various fusion rule is utilized to carry out assembled classification, the nicety of grading of more different fusion rule.
8. the many perception bio-identification exchange method for bore hole 3D advertisement according to claim 6, is characterized in that: the step of the algorithm of the age bracket identification that step is (3) described is as follows:
A, the face extracting human body, the colour of skin, hair style, the various feature of form, to face, choose distinguishing ability stronger left and right eyes, nose and face as regional area, reduce computation complexity to a certain extent, feature point for calibration is partitioned into the left and right eyes of face, nose and face region;
B, utilize linear discriminant method to extract the feature of human body of training and detecting, utilize Fisher method to extract the feature of training sample and target image simultaneously;
C, utilize dynamic cluster method to classify to target image, and calculate the average recognition rate of single classifier;
D, utilize Different age group training sample to train SVR, obtain one group of parameter value, utilize the SVR trained to classify to test sample book;
E, various fusion rule is utilized to carry out assembled classification, the nicety of grading of more different fusion rule.Increase class categories according to the classification iteration of target image, reduce age bracket class interval, to improve age estimated accuracy.
9. the many perception bio-identification exchange method for bore hole 3D advertisement according to claim 1, is characterized in that: the (4) described advertisement of step is recommended to throw in decision-making and adopted demographic recommended technology, and advertisement putting adopts manual type classification.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510282106.XA CN104915000A (en) | 2015-05-27 | 2015-05-27 | Multisensory biological recognition interaction method for naked eye 3D advertisement |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510282106.XA CN104915000A (en) | 2015-05-27 | 2015-05-27 | Multisensory biological recognition interaction method for naked eye 3D advertisement |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104915000A true CN104915000A (en) | 2015-09-16 |
Family
ID=54084136
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510282106.XA Pending CN104915000A (en) | 2015-05-27 | 2015-05-27 | Multisensory biological recognition interaction method for naked eye 3D advertisement |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104915000A (en) |
Cited By (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105550679A (en) * | 2016-02-29 | 2016-05-04 | 深圳前海勇艺达机器人有限公司 | Judgment method for robot to cyclically monitor recording |
| WO2017049918A1 (en) * | 2015-09-25 | 2017-03-30 | 杭州海康威视数字技术股份有限公司 | Information publishing terminal and method |
| CN106599656A (en) * | 2016-11-28 | 2017-04-26 | 深圳超多维科技有限公司 | Display method, device and electronic equipment |
| CN106625711A (en) * | 2016-12-30 | 2017-05-10 | 华南智能机器人创新研究院 | Method for positioning intelligent interaction of robot |
| CN106650601A (en) * | 2016-10-18 | 2017-05-10 | 武汉慧能机器人科技有限公司 | Human-machine interaction method and intelligent robot |
| CN107146096A (en) * | 2017-03-07 | 2017-09-08 | 浙江工业大学 | Intelligent video advertisement display method and device |
| CN107221252A (en) * | 2017-06-22 | 2017-09-29 | 蚌埠市麦沃特信息科技有限公司 | A kind of energy-saving signboard |
| CN107369398A (en) * | 2017-06-22 | 2017-11-21 | 蚌埠市麦沃特信息科技有限公司 | A kind of voice advertisement method |
| CN107678545A (en) * | 2017-09-26 | 2018-02-09 | 深圳市维冠视界科技股份有限公司 | A kind of information interactive terminal and method |
| WO2018035831A1 (en) * | 2016-08-25 | 2018-03-01 | 武克易 | Advertisement video playback system |
| CN107766782A (en) * | 2016-08-23 | 2018-03-06 | 中兴通讯股份有限公司 | A kind of method and device of age-colony classification |
| CN108022543A (en) * | 2017-11-27 | 2018-05-11 | 深圳中科呼图电子商务有限公司 | A kind of advertisement autonomous demenstration method, system and advertisement machine and application |
| CN108076128A (en) * | 2016-12-28 | 2018-05-25 | 北京市商汤科技开发有限公司 | User property extracting method, device and electronic equipment |
| CN108269133A (en) * | 2018-03-23 | 2018-07-10 | 深圳悠易阅科技有限公司 | A kind of combination human bioequivalence and the intelligent advertisement push method and terminal of speech recognition |
| CN109117833A (en) * | 2018-10-29 | 2019-01-01 | 天津市汇诚智慧体育科技有限公司 | A kind of full crowd movement's intelligent management system based on big data |
| CN109886730A (en) * | 2019-01-15 | 2019-06-14 | 北京奇艺世纪科技有限公司 | Put-on method, equipment and system, service equipment and method, processing equipment, medium |
| CN109934149A (en) * | 2019-03-06 | 2019-06-25 | 百度在线网络技术（北京）有限公司 | Method and apparatus for output information |
| CN110002302A (en) * | 2018-08-09 | 2019-07-12 | 浙江新再灵科技股份有限公司 | A kind of elevator switch door detection system and method based on deep learning |
| CN110070865A (en) * | 2019-04-03 | 2019-07-30 | 北京容联易通信息技术有限公司 | A kind of guidance robot with voice and image identification function |
| CN110674760A (en) * | 2019-09-27 | 2020-01-10 | 广州云从人工智能技术有限公司 | Interactive guiding method, device, equipment and medium based on mobile robot |
| CN113158834A (en) * | 2021-03-31 | 2021-07-23 | 中北大学南通智能光机电研究院 | Multi-perception face recognition magic cube |
| CN113435518A (en) * | 2021-06-29 | 2021-09-24 | 青岛海尔科技有限公司 | Feature fusion interaction method and device based on multiple modes |
| CN115099866A (en) * | 2022-07-07 | 2022-09-23 | 悦诚智慧(厦门)科技有限公司 | Advertisement delivery system based on AI glasses |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070092245A1 (en) * | 2005-10-20 | 2007-04-26 | Honeywell International Inc. | Face detection and tracking in a wide field of view |
| CN101236599A (en) * | 2007-12-29 | 2008-08-06 | 浙江工业大学 | Human face recognition detection device based on multi- video camera information integration |
| CN102289672A (en) * | 2011-06-03 | 2011-12-21 | 天津大学 | Infrared gait identification method adopting double-channel feature fusion |
| CN102298709A (en) * | 2011-09-07 | 2011-12-28 | 江西财经大学 | Energy-saving intelligent identification digital signage fused with multiple characteristics in complicated environment |
| CN102368294A (en) * | 2011-09-07 | 2012-03-07 | 常州蓝城信息科技有限公司 | Method for extracting sex characteristics under complex environment |
| CN102368295A (en) * | 2011-09-07 | 2012-03-07 | 常州蓝城信息科技有限公司 | Method for extracting age characteristics under complex environment |
| CN102446327A (en) * | 2010-10-15 | 2012-05-09 | 汉王科技股份有限公司 | Advertisement delivery device and method based on face recognition |
| CN102542252A (en) * | 2011-11-18 | 2012-07-04 | 江西财经大学 | Intelligent advertisement delivery system |
| CN102930454A (en) * | 2012-10-07 | 2013-02-13 | 乐配(天津)科技有限公司 | Intelligent 3D (Three Dimensional) advertisement recommendation method based on multiple perception technologies |
-
2015
- 2015-05-27 CN CN201510282106.XA patent/CN104915000A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070092245A1 (en) * | 2005-10-20 | 2007-04-26 | Honeywell International Inc. | Face detection and tracking in a wide field of view |
| CN101236599A (en) * | 2007-12-29 | 2008-08-06 | 浙江工业大学 | Human face recognition detection device based on multi- video camera information integration |
| CN102446327A (en) * | 2010-10-15 | 2012-05-09 | 汉王科技股份有限公司 | Advertisement delivery device and method based on face recognition |
| CN102289672A (en) * | 2011-06-03 | 2011-12-21 | 天津大学 | Infrared gait identification method adopting double-channel feature fusion |
| CN102298709A (en) * | 2011-09-07 | 2011-12-28 | 江西财经大学 | Energy-saving intelligent identification digital signage fused with multiple characteristics in complicated environment |
| CN102368294A (en) * | 2011-09-07 | 2012-03-07 | 常州蓝城信息科技有限公司 | Method for extracting sex characteristics under complex environment |
| CN102368295A (en) * | 2011-09-07 | 2012-03-07 | 常州蓝城信息科技有限公司 | Method for extracting age characteristics under complex environment |
| CN102542252A (en) * | 2011-11-18 | 2012-07-04 | 江西财经大学 | Intelligent advertisement delivery system |
| CN102930454A (en) * | 2012-10-07 | 2013-02-13 | 乐配(天津)科技有限公司 | Intelligent 3D (Three Dimensional) advertisement recommendation method based on multiple perception technologies |
Non-Patent Citations (2)
| Title |
|---|
| 中央文化企业国有资产监督管理领导小组办公室等: "《文化产业专题研究报告 （下）》", 31 December 2014 * |
| 李学庆等编: "《和谐人机环境 2007》", 31 May 2008 * |
Cited By (30)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106557940B (en) * | 2015-09-25 | 2019-09-17 | 杭州海康威视数字技术股份有限公司 | Information release terminal and method |
| WO2017049918A1 (en) * | 2015-09-25 | 2017-03-30 | 杭州海康威视数字技术股份有限公司 | Information publishing terminal and method |
| CN106557940A (en) * | 2015-09-25 | 2017-04-05 | 杭州海康威视数字技术股份有限公司 | Information release terminal and method |
| CN105550679B (en) * | 2016-02-29 | 2019-02-15 | 深圳前海勇艺达机器人有限公司 | A kind of judgment method of robot cycle monitoring recording |
| CN105550679A (en) * | 2016-02-29 | 2016-05-04 | 深圳前海勇艺达机器人有限公司 | Judgment method for robot to cyclically monitor recording |
| CN107766782A (en) * | 2016-08-23 | 2018-03-06 | 中兴通讯股份有限公司 | A kind of method and device of age-colony classification |
| WO2018035831A1 (en) * | 2016-08-25 | 2018-03-01 | 武克易 | Advertisement video playback system |
| CN106650601A (en) * | 2016-10-18 | 2017-05-10 | 武汉慧能机器人科技有限公司 | Human-machine interaction method and intelligent robot |
| CN106599656A (en) * | 2016-11-28 | 2017-04-26 | 深圳超多维科技有限公司 | Display method, device and electronic equipment |
| WO2018121541A1 (en) * | 2016-12-28 | 2018-07-05 | 北京市商汤科技开发有限公司 | User attribute extraction method and apparatus, and electronic device |
| CN108076128A (en) * | 2016-12-28 | 2018-05-25 | 北京市商汤科技开发有限公司 | User property extracting method, device and electronic equipment |
| CN106625711A (en) * | 2016-12-30 | 2017-05-10 | 华南智能机器人创新研究院 | Method for positioning intelligent interaction of robot |
| CN107146096B (en) * | 2017-03-07 | 2020-08-18 | 浙江工业大学 | Intelligent video advertisement display method and device |
| CN107146096A (en) * | 2017-03-07 | 2017-09-08 | 浙江工业大学 | Intelligent video advertisement display method and device |
| CN107221252A (en) * | 2017-06-22 | 2017-09-29 | 蚌埠市麦沃特信息科技有限公司 | A kind of energy-saving signboard |
| CN107369398A (en) * | 2017-06-22 | 2017-11-21 | 蚌埠市麦沃特信息科技有限公司 | A kind of voice advertisement method |
| CN107678545A (en) * | 2017-09-26 | 2018-02-09 | 深圳市维冠视界科技股份有限公司 | A kind of information interactive terminal and method |
| CN108022543A (en) * | 2017-11-27 | 2018-05-11 | 深圳中科呼图电子商务有限公司 | A kind of advertisement autonomous demenstration method, system and advertisement machine and application |
| CN108269133A (en) * | 2018-03-23 | 2018-07-10 | 深圳悠易阅科技有限公司 | A kind of combination human bioequivalence and the intelligent advertisement push method and terminal of speech recognition |
| CN110002302A (en) * | 2018-08-09 | 2019-07-12 | 浙江新再灵科技股份有限公司 | A kind of elevator switch door detection system and method based on deep learning |
| CN109117833A (en) * | 2018-10-29 | 2019-01-01 | 天津市汇诚智慧体育科技有限公司 | A kind of full crowd movement's intelligent management system based on big data |
| CN109886730A (en) * | 2019-01-15 | 2019-06-14 | 北京奇艺世纪科技有限公司 | Put-on method, equipment and system, service equipment and method, processing equipment, medium |
| CN109934149A (en) * | 2019-03-06 | 2019-06-25 | 百度在线网络技术（北京）有限公司 | Method and apparatus for output information |
| CN109934149B (en) * | 2019-03-06 | 2022-08-09 | 百度在线网络技术（北京）有限公司 | Method and apparatus for outputting information |
| CN110070865A (en) * | 2019-04-03 | 2019-07-30 | 北京容联易通信息技术有限公司 | A kind of guidance robot with voice and image identification function |
| CN110674760A (en) * | 2019-09-27 | 2020-01-10 | 广州云从人工智能技术有限公司 | Interactive guiding method, device, equipment and medium based on mobile robot |
| CN113158834A (en) * | 2021-03-31 | 2021-07-23 | 中北大学南通智能光机电研究院 | Multi-perception face recognition magic cube |
| CN113435518A (en) * | 2021-06-29 | 2021-09-24 | 青岛海尔科技有限公司 | Feature fusion interaction method and device based on multiple modes |
| CN113435518B (en) * | 2021-06-29 | 2024-03-22 | 青岛海尔科技有限公司 | Multi-mode-based interaction method and device for feature fusion |
| CN115099866A (en) * | 2022-07-07 | 2022-09-23 | 悦诚智慧(厦门)科技有限公司 | Advertisement delivery system based on AI glasses |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104915000A (en) | Multisensory biological recognition interaction method for naked eye 3D advertisement | |
| CN102930454A (en) | Intelligent 3D (Three Dimensional) advertisement recommendation method based on multiple perception technologies | |
| CN107197384B (en) | The multi-modal exchange method of virtual robot and system applied to net cast platform | |
| CN102081918B (en) | Video image display control method and video image display device | |
| KR20160012902A (en) | Method and device for playing advertisements based on associated information between audiences | |
| CN104486680A (en) | Video-based advertisement pushing method and system | |
| CN102298709A (en) | Energy-saving intelligent identification digital signage fused with multiple characteristics in complicated environment | |
| CN109614849A (en) | Remote teaching method, apparatus, equipment and storage medium based on bio-identification | |
| CN106682946A (en) | Advertisement content analysis method and device | |
| CN105426850A (en) | Human face identification based related information pushing device and method | |
| US20030126013A1 (en) | Viewer-targeted display system and method | |
| CN107153976A (en) | The Instant Ads play system and player method recognized based on characteristics of human body | |
| Yargıç et al. | A lip reading application on MS Kinect camera | |
| Yu et al. | AI-based targeted advertising system | |
| Kishore et al. | Continuous sign language recognition from tracking and shape features using fuzzy inference engine | |
| Zhou et al. | Now you shake me: Towards automatic 4D cinema | |
| Zhang et al. | Teaching chinese sign language with a smartphone | |
| US20220375223A1 (en) | Information generation method and apparatus | |
| CN117541321B (en) | Advertisement making and publishing method and system based on virtual digital person | |
| CN102368295A (en) | Method for extracting age characteristics under complex environment | |
| CN103218829A (en) | Foreground extracting method suitable for dynamic background | |
| KR20200004014A (en) | Digital signage system for clothing fitting and method for displaying items using thereof | |
| CN112764530A (en) | Ammunition identification method based on touch handle and augmented reality glasses | |
| CN116910302A (en) | Multi-mode video content effectiveness feedback visual analysis method and system | |
| CN106571108A (en) | Advisement player having voice interaction function |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150916 |