CN102200834B - Television control-oriented finger-mouse interaction method - Google Patents

Television control-oriented finger-mouse interaction method Download PDF

Info

Publication number
CN102200834B
CN102200834B CN2011101380409A CN201110138040A CN102200834B CN 102200834 B CN102200834 B CN 102200834B CN 2011101380409 A CN2011101380409 A CN 2011101380409A CN 201110138040 A CN201110138040 A CN 201110138040A CN 102200834 B CN102200834 B CN 102200834B
Authority
CN
China
Prior art keywords
step
fist
forefinger
image
finger tip
Prior art date
Application number
CN2011101380409A
Other languages
Chinese (zh)
Other versions
CN102200834A (en
Inventor
姜小波
石任重
Original Assignee
华南理工大学
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 华南理工大学 filed Critical 华南理工大学
Priority to CN2011101380409A priority Critical patent/CN102200834B/en
Publication of CN102200834A publication Critical patent/CN102200834A/en
Application granted granted Critical
Publication of CN102200834B publication Critical patent/CN102200834B/en

Links

Abstract

The invention discloses a television control-oriented finger-mouse interaction method. A television interaction system is characterized by comprising the following steps of: acquiring an image by using a built-in camera of a television; performing image detection on the acquired image to obtain a gesture area; and locating a finger and obtaining a motion coordinate of the finger by using a characteristic of the gesture area, so that an execution device in the television completes the corresponding movement and click operation of a mouse pointer according to the coordinate motion trajectory and residence time of the finger. By the method, the remote control operation of the television can be performed through gesture movement, so a man-machine interaction manner of the television is more intelligent and natural. The long-distance natural remote control manner has remarkable promotion effect on the field of television application under a trend of tri-networks integration.

Description

Finger tip mouse exchange method towards TV control

Technical field

The present invention relates to a kind of TV remote technology, be specifically related to towards the finger tip mouse exchange method and the system of TV control.

Background technology

In recent years, along with the integration of three networks and television set intelligently development trend, TV becomes the intelligent information terminal of family rapidly.This human-computer interaction technology to TV has proposed new requirement.TV has possessed online, the function of information processing.The TV man-machine interaction needs literal input and mouse click function.Traditional ultrared button telepilot that utilizes can't satisfy the new interaction demand that the TV development brings.More intelligent, natural man-machine interaction mode is the active demand of TV, and the research of TV man-machine interaction becomes a research focus.

The problem that prior art exists:

1. present TV remote controller all is stereotyped in the operation that profile is still imported, the characteristic that the product of each big TV production company can not project oneself.

2. along with the diversification of TV functions, the design of telepilot also becomes increasingly complex, and the rapid increase of button makes that the straighforward operation of TV is more and more inconvenient.

3. along with the popularizing of electrification equipment, various device can be furnished with different telepilots, and such as TV, air-conditioning, DVD, projector etc., TV user can be in a rush in order to seek telepilot.

4. present gesture telepilot also only is satisfied with and is changed platform, regulates these shirtsleeve operations of volume, also can't satisfy accurate finger tip mouse control function.

Therefore, providing a kind of utilizes the finger tip mouse to carry out the TV remote-control method of operating and utilizes this method to realize that the man-machine interaction telechirics of nature will be very important.

Summary of the invention

The objective of the invention is to overcome the above-mentioned defective that prior art exists; A kind of finger tip mouse exchange method towards TV control is provided; The system that constitutes by built-in process chip of TV and image collecting device, through images of gestures collection, images of gestures detect, forefinger finger tip location, operational order carry out four steps and realize that a kind of auxiliary devices that need not directly use the man-machine interaction method of staff as the nature of input tool.

A kind of finger tip mouse exchange method towards TV control, it may further comprise the steps:

Step 1 is utilized the camera collection image of television internal, obtains the RGB image;

Step 2 is carried out gray-scale map to said RGB image and is converted to gray level image, keeps original image simultaneously;

Step 3 is carried out statistics with histogram to said gray level image, carry out gray-level histogram equalizationization then;

Step 4, the method for utilizing template matches detects the fist image of gesture, if do not find the fist image, then returns step 1, if find the fist image then to note the coordinate of the peripheral rectangle frame of fist;

Step 5 is mapped to the fist zone in the gray level image in the RGB image that keeps in the step 2;

Step 6 is carried out the HSV color space transformation to the forefinger zone and the fist zone of RGB image fist top;

Step 7 is carried out H histogram of component statistics to the fist zone;

Step 8 according to the H histogram of component statistics setting threshold in fist zone, is carried out threshold process to the pixel in forefinger zone, fist top then; The total number of pixel according to remaining after the threshold process judges whether to exist forefinger, if the remaining total number of pixel does not reach pre-set threshold, thinks that then forefinger does not exist; Return step 1 simultaneously; If the remaining total number of pixel has reached pre-set threshold, think that then forefinger exists, and notes forefinger profile coordinate then;

Step 9 is utilized the geometric properties of forefinger, and the forefinger profile is carried out the finger tip location, obtains the coordinate figure of finger tip;

Step 10 according to the variation of finger tip coordinate, produces the virtual mouse instruction, accomplishes corresponding man-machine interactive operation.

Above-mentioned is a kind of in the finger tip mouse exchange method of TV control, and the purpose of gray-level histogram equalizationization is on the basis that does not influence the gray level image overall contrast, to strengthen local contrast in the step 3.

Above-mentioned is a kind of in the finger tip mouse exchange method of TV control; Fist image detection in the step 4 comprises feature extraction and characteristic matching; Adopt Haar characteristic frame traversal entire image to detect, detection algorithm then adopts the AdaBoost algorithm, and the traversal frame is searched for the fist image; In case characteristic matching success then jump out search step gets into next step.

Above-mentioned is a kind of in the finger tip mouse exchange method of TV control, in the said HSV color space of step 6, and the color type of H component statement object, i.e. tone; S is a saturation degree; V is brightness.

Above-mentioned is a kind of in the finger tip mouse exchange method of TV control; The setting threshold as a result that step 8 is at first added up according to fist zone H histogram of component; Utilize the H component of fist and the forefinger conforming characteristic that distributes, can carry out threshold process to the forefinger area pixel of fist top then.

Above-mentioned is a kind of in the finger tip mouse exchange method of TV control, geometric properties type of the being character column of forefinger in the step 9.

Above-mentioned a kind of in the finger tip mouse exchange method of TV control, in the step 10, according to the motion of the slip simulation mouse pointer of finger tip, according to finger tip in the residence time of certain some analog mouse clicking operation.

Above-mentioned is a kind of in the finger tip mouse exchange method of TV control, and the processing procedure of step 2 to step 9 all is to accomplish through FPGA.

In the said method, step 1 is used to obtain user's images of gestures data, as gesture to be detected through the continuous acquisition of image data of the image collecting device of the televisor buffer memory that Updates Information.In the step 2, the light of TV screen changes easily, and this utilizes the half-tone information of image can reduce this influence to having produced very big influence cutting apart with detection of coloured image.Gray-level histogram equalizationization in the step 3 can strengthen local contrast, helps the detection of subsequent gesture image; Images of gestures detects in the step 4; The mode of employing off-line training obtains the feature database of images of gestures; The characteristic of template in gesture feature to be detected and the feature database is mated; Thereby the testing result of drawing because the process that detects is the target search of carrying out through the traversal entire image, has therefore been saved the step of image segmentation.Owing to the nonreversibility of greyscale image transitions, the zone that the gray level image fist detects must be mapped in the RGB coloured image, for the detection of forefinger is prepared in the step 5.Only the HSV color space transformation is carried out in the forefinger zone and the fist zone of RGB image fist top in the step 6, other zone all is set to background, can reduce calculated amount like this, improves processing speed.In same two field picture, do not have the influence of illumination variation in the step 8, so the colour of skin information of fist and forefinger can be consistent highly, utilize fist region histogram statistics preset threshold to cut apart like this forefinger zone.Step 9: finger tip location, the class character column of finger shape for the location of finger tip provides the simple geometric characteristic, thereby obtains the coordinate figure of finger tip point.Step 10 through judging the clicking operation of the time simulation mouse that the finger tip coordinate stops, is finally carried out corresponding instruction through the man-machine interactive system in the televisor according to the motion of the situation of change analog mouse pointer of finger tip coordinate.

Compared with prior art; The present invention has following advantage and technique effect: the present invention proposes a kind of natural finger tip mouse exchange method towards TV control; This method has the high detection rate, receives that illumination effect is little, the AdaBoost algorithm based on the haar characteristic of real-time detection through employing; And realize the location with finger tip of cutting apart of forefinger, and produce corresponding television control signals according to the features of skin colors of fist.This method has reduced the influence that TV screen light changes to greatest extent.This remote control mode will promote the development of TV man-machine interaction application significantly.

Description of drawings

Fig. 1 is the gesture synoptic diagram of finger tip mouse exchange method, and S1 refers to that fist zone and S2 refer to the forefinger zone of fist top.

Fig. 2 is the structured flowchart of the graphics processing unit of finger tip mouse exchange method.

Fig. 3 example be the flow chart of data processing figure of finger tip mouse exchange method.

Embodiment

Below in conjunction with accompanying drawing enforcement of the present invention is described further, but enforcement of the present invention and protection domain are not limited thereto.

Fig. 1 is the gesture synoptic diagram of finger tip mouse exchange method, is a fist iconic model of erectting forefinger among the figure.S1 refers to the fist zone; Method through template matches detects fist; Note the coordinate figure of the peripheral rectangle frame of fist then, S2 refers to the forefinger zone of fist top, and it is through fist image among the S1 and consistent realization of forefinger colour of skin distribution that the forefinger in the S2 zone detects.The rectangle frame that two with dashed lines among the figure surround is to carry out the zone that successive image is handled after fist detects completion.

The graphics processing unit of Fig. 2 comprises that fist detection, off-line training fist eigenwert, forefinger cut apart three processes in finger tip location.

One, fist detects

Comprise three steps: gray-scale map conversion, gray-level histogram equalizationization, fist zone are detected.Wherein gray-scale map conversion belongs to preprocessing part with histogram equalization, and the fist detection method of this paper has been skipped traditional step that detects again of cutting apart earlier, adopts the method for traversal search directly to get into the detection step from pre-treatment step.This is because for remote gestures detection, the target area has only accounted for very little a part of ratio of entire image, therefore utilizes traditional dividing method to fall the target area as background process; Simultaneously because the colour of skin consistance of people's face and staff also can be brought very big interference when cutting apart.The gray-level histogram equalizationization of preprocessing part can strengthen local contrast, for choosing of eigenwert provides advantage.

The fist zone is detected the link employing and is detected based on Haar characteristic AdaBoost sorting algorithm, and this algorithm needs off-line to set up the feature templates storehouse, and wherein eigenwert is through setting up the sample storehouse, utilizing the AdaBoost algorithm training to obtain.This sample storehouse comprises the colour of skin and the positive sample of the fist at age of the different people that 1383 width of cloth are gathered under the different light environment, 7820 width of cloth do not comprise the negative sample of arbitrary image of fist.The verification and measurement ratio of AdaBoost algorithm can reach more than 95%, and satisfies the requirement of real-time, theoretical proof, as long as the sample storehouse is enough big, verification and measurement ratio can be infinitely close to 1.The advantage of AdaBoost algorithm self makes finger tip mouse exchange method have very strong robustness.

The Haar eigenwert is based on adjacent area respectively with the summation of the gray-scale value of each pixel, then relatively each area grayscale and difference, it belongs to appearance features.The Haar feature class is similar to rectangle frame; But because the scale feature of the translation of rectangle frame and scaling; So Haar characteristic enormous amount of this picture of duplicate sample; The Haar characteristic that the purpose of training is utilized the AdaBoost algorithm to filter out to meet the demands just, these characteristics that meet the demands be usually located at those adjacent area gray scales and the zone that differs greatly of difference, the tangible zone of those local contrast just.The calculating of Haar eigenwert is the thought of utilizing integrogram, the integration that this method only need be known rectangle frame corresponding vertex pixel with, utilize simple plus and minus calculation just can obtain then.

It is to utilize a kind of search box to realize through traveling through the method for coming that fist detects, rectangle frame of this search box, and there is the Haar characteristic frame that meets the demands that obtains through the AdaBoost algorithm training the inside.The traversal of search box is to accomplish through the translation of yardstick and scaling, in case the feature templates storehouse coupling that the zone that search box detects and training obtain, then the zone detected of expression is the fist image, immediately with rectangle frame with its sign and jump out the detection step; If search box has traveled through entire image, all do not find the target area that meets the demands, then there is not fist in the presentation video, system comes back to step 1.

Two, forefinger is cut apart with finger tip and is located

Can be known that by Fig. 1 forefinger is positioned at the S2 zone, the present invention utilizes the colour of skin consistance characteristic of fist and forefinger to realize cutting apart forefinger according to detected fist (S1 zone) information.The detection of forefinger is to detect through fist earlier, and then carries out forefinger and cut apart that two steps accomplish.Why not the whole gesture of direct Training; Only just realize that through this step of gestures detection gesture (S1 and S2 zone) detects? The experiment proof; Because forefinger only accounts for the very little part in S2 zone, remaining part all is a background, and positive sample can be introduced a large amount of backgrounds when therefore training; A large amount of negative samples is trained as positive sample together, and final result can cause verification and measurement ratio seriously to reduce.

Because people's the colour of skin has the characteristic of cluster in the HSV color space, therefore in the HSV color space, can realize cutting apart to forefinger.At first S1 and S2 zone are carried out the HSV color space transformation, the characteristic that only adopts tonal content (H component) to cut apart then as forefinger, the computing formula that is transformed into the HSV color space by rgb color space is following:

if , then

Through after the HSV color space transformation, then H histogram of component statistics is carried out in the S1 zone, because the S1 zone nearly all occupies by fist, again because the cluster characteristic of the colour of skin, so the statistics with histogram result in S1 zone can present tangible cluster feature.According to the frequency statistics of each H component value, just can obtain one and be used to carry out the threshold value that S2 zone forefinger is cut apart.Because the S2 zone only accounts for the very little part of entire image, it is simple to operate therefore to carry out Threshold Segmentation for the S2 zone, and processing speed is fast.Simultaneously, because the colour of skin consistance of fist and forefinger so just can be cut apart forefinger accurately.Fist and forefinger are arranged in same secondary figure, and the detection algorithm of fist has very strong robustness again, like this variation of illumination to forefinger to cut apart influence very little.

The forefinger of final type cylinder model is the similar club model of meeting on the 2D projection plane, and finger tip always is distributed on the summit of club, utilizes this geometric properties just can realize the location of finger tip.So just can obtain the coordinate figure that finger tip is arranged in image, thereby accomplish corresponding with the virtual mouse pointer in the TV.Like this, just can realize the pen travel of mouse, simultaneously, can also utilize residence time of finger tip to judge the click action of virtual mouse through moving of finger tip.

The finger tip mouse exchange method flow chart of data processing figure of Fig. 3 specifically comprises the steps:

Step 1, IMAQ utilizes the built-in camera images acquired in the televisor, obtains the RGB image, is used to obtain user's images of gestures data, as gesture to be detected.

Step 2, the gray-scale map conversion is carried out gray-scale map to said RGB image and is converted to gray level image, keeps original image simultaneously; The light of TV screen changes easily, and this utilizes the half-tone information of image can reduce this influence to having produced very big influence cutting apart with detection of coloured image.

Step 3, gray-level histogram equalizationization is carried out statistics with histogram to said gray level image, carry out gray-level histogram equalizationization then; This step can strengthen local contrast, helps the detection of subsequent gesture image.

Step 4, images of gestures detect, and the method for utilizing template matches detects the fist image of gesture, if do not find the fist image, then returns step 1, if find the fist image then to note the coordinate of the peripheral rectangle frame of fist; Because the process that detects is the target search of carrying out through the traversal entire image, therefore saved the step of image segmentation.

Step 5, the fist zone marker is mapped to the fist zone in the gray level image in the RGB image that keeps in the step 2; Because the nonreversibility of greyscale image transitions, the zone that the gray level image fist detects must be mapped in the RGB coloured image, for the detection of forefinger is prepared.

Step 6, marked region HSV color space transformation is carried out the HSV color space transformation to the forefinger zone and the fist zone of RGB image fist top; Here only the HSV color space transformation is carried out in the forefinger zone and the fist zone of RGB image fist top, other zone all is set to background, can reduce calculated amount like this, improves processing speed.

Step 7, H histogram of component statistics is carried out H histogram of component statistics to the fist zone, according to the threshold parameter of automatic setting as a result of statistics.

Step 8, the forefinger Region Segmentation is according to the H histogram of component statistics setting threshold in fist zone; Then the pixel in forefinger zone, fist top is carried out threshold process; The total number of pixel according to remaining after the threshold process judges whether to exist forefinger, if the remaining total number of pixel does not reach pre-set threshold, thinks that then forefinger does not exist; Return step 1 simultaneously; If the remaining total number of pixel has reached pre-set threshold, think that then forefinger exists, and notes forefinger profile coordinate then; Here because in same two field picture, do not have the influence of illumination variation, so the colour of skin information of fist and forefinger can be consistent highly, utilizing the fist region histogram to add up preset threshold like this can well cut apart the forefinger zone.

Step 9, finger tip is located, and utilizes the geometric properties of forefinger, and the forefinger profile is carried out the finger tip location, obtains the coordinate figure of finger tip; The class character column of finger shape is for the location of finger tip provides the simple geometric characteristic.

Step 10, the finger tip mouse instructions according to the situation of change of finger tip coordinate, produces the virtual mouse instruction, accomplishes corresponding man-machine interactive operation.

Utilize above-mentioned method, just can carry out the straighforward operation of TV,, make the straighforward operation of TV become simple and nature simultaneously by the remote handling system in the TV with finger tip.Therefore, towards the finger tip mouse exchange method of TV control, be a kind of contactless nature and TV remote controlling method easily.

Claims (8)

1. finger tip mouse exchange method towards TV control is characterized in that may further comprise the steps:
Step 1 is utilized the built-in camera images acquired in the televisor, obtains the RGB image;
Step 2 is carried out gray-scale map to said RGB image and is converted to gray level image, keeps original image simultaneously;
Step 3 is carried out statistics with histogram to said gray level image, carry out gray-level histogram equalizationization then;
Step 4, the method for utilizing template matches detects the fist image of gesture, if do not find the fist image, then returns step 1, if find the fist image then to note the coordinate of the peripheral rectangle frame of fist;
Step 5 is mapped to the fist zone in the gray level image in the RGB image that keeps in the step 2;
Step 6 is carried out the HSV color space transformation to the forefinger zone and the fist zone of RGB image fist top;
Step 7 is carried out H histogram of component statistics to the fist zone;
Step 8 according to the H histogram of component statistics setting threshold in fist zone, is carried out threshold process to the pixel in forefinger zone, fist top then; The total number of pixel according to remaining after the threshold process judges whether to exist forefinger, if the remaining total number of pixel does not reach pre-set threshold, thinks that then forefinger does not exist; Return step 1 simultaneously; If the remaining total number of pixel has reached pre-set threshold, think that then forefinger exists, and notes forefinger profile coordinate then;
Step 9 is utilized the geometric properties of forefinger, and the forefinger profile is carried out the finger tip location, obtains the coordinate figure of finger tip;
Step 10 according to the variation of finger tip coordinate, produces the virtual mouse instruction, accomplishes corresponding man-machine interactive operation.
2. a kind of finger tip mouse exchange method towards TV control according to claim 1 is characterized in that the purpose of gray-level histogram equalizationization is on the basis that does not influence the gray level image overall contrast, to strengthen local contrast in the step 3.
3. a kind of finger tip mouse exchange method according to claim 1 towards TV control; It is characterized in that the fist image detection in the step 4 comprises feature extraction and characteristic matching; Adopt Haar characteristic frame traversal entire image to detect, detection algorithm then adopts the AdaBoost algorithm, and the traversal frame is searched for the fist image; In case characteristic matching success then jump out search step gets into next step.
4. a kind of finger tip mouse exchange method towards TV control according to claim 1 is characterized in that in the said HSV color space of step 6 the color type of H component statement object, i.e. tone; S is a saturation degree; V is brightness.
5. a kind of finger tip mouse exchange method according to claim 1 towards TV control; It is characterized in that the setting threshold as a result that step 8 is at first added up according to fist zone H histogram of component; Utilize the H component of fist and the forefinger conforming characteristic that distributes, can carry out threshold process to the forefinger area pixel of fist top then.
6. a kind of finger tip mouse exchange method towards TV control according to claim 1 is characterized in that the geometric properties of forefinger in the step 9 is a type character column.
7. a kind of finger tip mouse exchange method towards TV control according to claim 1 is characterized in that in the step 10, according to the motion of the slip simulation mouse pointer of finger tip, according to finger tip in the residence time of certain some analog mouse clicking operation.
8. according to each described a kind of finger tip mouse exchange method towards TV control of claim 1 ~ 7, the processing procedure that it is characterized in that step 2 to step 9 all is to accomplish through FPGA.
CN2011101380409A 2011-05-26 2011-05-26 Television control-oriented finger-mouse interaction method CN102200834B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN2011101380409A CN102200834B (en) 2011-05-26 2011-05-26 Television control-oriented finger-mouse interaction method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN2011101380409A CN102200834B (en) 2011-05-26 2011-05-26 Television control-oriented finger-mouse interaction method

Publications (2)

Publication Number Publication Date
CN102200834A CN102200834A (en) 2011-09-28
CN102200834B true CN102200834B (en) 2012-10-31

Family

ID=44661561

Family Applications (1)

Application Number Title Priority Date Filing Date
CN2011101380409A CN102200834B (en) 2011-05-26 2011-05-26 Television control-oriented finger-mouse interaction method

Country Status (1)

Country Link
CN (1) CN102200834B (en)

Families Citing this family (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102508544A (en) * 2011-10-24 2012-06-20 四川长虹电器股份有限公司 Intelligent television interactive method based on projection interaction
TWI571772B (en) 2011-10-31 2017-02-21 財團法人資訊工業策進會 Virtual mouse driving apparatus and virtual mouse simulation method
CN102521567B (en) * 2011-11-29 2013-10-23 Tcl集团股份有限公司 Human-computer interaction fingertip detection method, device and television
TWI540461B (en) * 2011-12-05 2016-07-01 緯創資通股份有限公司 Gesture input method and system
CN102662577B (en) * 2012-03-29 2016-08-10 华为终端有限公司 A kind of cursor operating method based on three dimensional display and mobile terminal
US9141198B2 (en) * 2013-01-08 2015-09-22 Infineon Technologies Ag Control of a control parameter by gesture recognition
CN103179445A (en) * 2013-03-26 2013-06-26 Tcl集团股份有限公司 Method, device and television (TV) for receiving external input signals
CN103366188B (en) * 2013-07-08 2017-07-07 中科创达软件股份有限公司 It is a kind of to be detected as the gesture tracking method of auxiliary information based on fist
JP6349800B2 (en) * 2014-03-12 2018-07-04 オムロン株式会社 Gesture recognition device and method for controlling gesture recognition device
CN105323619A (en) * 2014-08-04 2016-02-10 深圳市同方多媒体科技有限公司 Gesture control method and gesture control television based on analog button board
CN105373213A (en) * 2014-08-25 2016-03-02 乐视致新电子科技(天津)有限公司 Operation identification method and device in human-computer interaction
CN105635776B (en) * 2014-11-06 2019-03-01 深圳Tcl新技术有限公司 Pseudo operation graphical interface remoting control method and system
CN104461008B (en) * 2014-12-23 2017-12-08 山东建筑大学 A kind of multimedia teaching control system and control method
CN104484040B (en) * 2014-12-23 2017-12-08 山东建筑大学 A kind of multimedia interactive teaching control system and learning control mode
CN107872657A (en) * 2017-11-29 2018-04-03 杭州电子科技大学 The method of color gamut space conversion based on FPGA

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101344816A (en) * 2008-08-15 2009-01-14 华南理工大学 Human-machine interaction method and device based on sight tracing and gesture discriminating
CN101437124A (en) * 2008-12-17 2009-05-20 三星电子(中国)研发中心 Method for processing dynamic gesture identification signal facing (to)television set control
CN101729808A (en) * 2008-10-14 2010-06-09 Tcl集团股份有限公司 Remote control method for television and system for remotely controlling television by same
CN101853071A (en) * 2010-05-13 2010-10-06 重庆大学 Gesture identification method and system based on visual sense

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101344816A (en) * 2008-08-15 2009-01-14 华南理工大学 Human-machine interaction method and device based on sight tracing and gesture discriminating
CN101729808A (en) * 2008-10-14 2010-06-09 Tcl集团股份有限公司 Remote control method for television and system for remotely controlling television by same
CN101437124A (en) * 2008-12-17 2009-05-20 三星电子(中国)研发中心 Method for processing dynamic gesture identification signal facing (to)television set control
CN101853071A (en) * 2010-05-13 2010-10-06 重庆大学 Gesture identification method and system based on visual sense

Also Published As

Publication number Publication date
CN102200834A (en) 2011-09-28

Similar Documents

Publication Publication Date Title
Yi et al. Age estimation by multi-scale convolutional network
Ibraheem et al. Survey on various gesture recognition technologies and techniques
TWI454968B (en) Three-dimensional interactive device and operation method thereof
CN105353634B (en) Utilize the home appliance and method of gesture identification control operation
CN105718878B (en) The aerial hand-written and aerial exchange method in the first visual angle based on concatenated convolutional neural network
CN101763515B (en) Real-time gesture interaction method based on computer vision
JP3863809B2 (en) Input system by hand image recognition
CN102576259B (en) Hand position detection method
CN105160310A (en) 3D (three-dimensional) convolutional neural network based human body behavior recognition method
CN101561710B (en) Man-machine interaction method based on estimation of human face posture
Murthy et al. Hand gesture recognition using neural networks
Kim et al. Simultaneous gesture segmentation and recognition based on forward spotting accumulative HMMs
Zhou et al. A novel finger and hand pose estimation technique for real-time hand gesture recognition
CN102402680B (en) Hand and indication point positioning method and gesture confirming method in man-machine interactive system
KR101526644B1 (en) Method system and software for providing image sensor based human machine interfacing
CN105069413A (en) Human body gesture identification method based on depth convolution neural network
CN101777116B (en) Method for analyzing facial expressions on basis of motion tracking
CN102122350B (en) Skeletonization and template matching-based traffic police gesture identification method
RU2711029C2 (en) Touch classification
CN103472916B (en) A kind of man-machine interaction method based on human body gesture identification
US20120027263A1 (en) Hand gesture detection
US20120027252A1 (en) Hand gesture detection
WO2015000286A1 (en) Three-dimensional interactive learning system and method based on augmented reality
KR101017936B1 (en) The system controled a action of the display device, based a gesture information recognition of the user
CN102567727B (en) Method and device for replacing background target

Legal Events

Date Code Title Description
PB01 Publication
C06 Publication
SE01 Entry into force of request for substantive examination
C10 Entry into substantive examination
GR01 Patent grant
C14 Grant of patent or utility model
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20121031

Termination date: 20180526

CF01 Termination of patent right due to non-payment of annual fee