CN107589850A - A kind of recognition methods of gesture moving direction and system - Google Patents

A kind of recognition methods of gesture moving direction and system Download PDF

Info

Publication number
CN107589850A
CN107589850A CN201710884796.5A CN201710884796A CN107589850A CN 107589850 A CN107589850 A CN 107589850A CN 201710884796 A CN201710884796 A CN 201710884796A CN 107589850 A CN107589850 A CN 107589850A
Authority
CN
China
Prior art keywords
frame
accumulative
hand
palm
gesture
Prior art date
Application number
CN201710884796.5A
Other languages
Chinese (zh)
Inventor
黄恩武
Original Assignee
深圳睛灵科技有限公司
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 深圳睛灵科技有限公司 filed Critical 深圳睛灵科技有限公司
Priority to CN201710884796.5A priority Critical patent/CN107589850A/en
Publication of CN107589850A publication Critical patent/CN107589850A/en

Links

Abstract

The invention discloses a kind of recognition methods of gesture moving direction, including:In detection cycle, multiframe images of gestures is obtained, and hand region is identified from every frame images of gestures;Palm or the back of the hand center of the positioning per frame hand region, and obtain palm or the back of the hand radius;According to the change of palm or the back of the hand center and palm or the back of the hand radius between the multiframe images of gestures, the moving direction of gesture is identified.In addition, the invention discloses a kind of identifying system of gesture moving direction.The present invention can improve the recognition accuracy of gesture moving direction, and stability is good, and recognition efficiency is high.

Description

A kind of recognition methods of gesture moving direction and system

Technical field

The present invention relates to recognition methods and the system of technical field of image processing, more particularly to a kind of gesture moving direction.

Background technology

Gesture identification is that the body language of people is identified using technologies such as computer graphics, and is converted into order and is set to operate Standby, this is new man-machine interaction mode after mouse, keyboard and touch-screen.Gesture is due to nature, random, abundant, square Just etc. characteristic, the man-machine interaction mode based on gesture identification have received more and more extensive attention.

Gesture identification can be divided into static gesture identification and dynamic hand gesture recognition.Static gesture identification can only identify single frames figure The state of hand as in, and the lasting change of hand can not be perceived;Dynamic hand gesture recognition can follow the trail of the motion of gesture and then identification will The compound movement that hand-type is combined together with hand exercise.In dynamic gesture identification, generally require to identify the motion of hand Direction, such as up and down and fore-and-aft direction movement.Therefore how to stablize and rapidly identify the moving direction of gesture for hand Gesture identifying system is quite important.

Gesture identification method of the prior art is that the human hand focus point in multiple continuous images is identified to judge The moving direction of human hand.But movement of this method None- identified human hand on the direction vertical with plane where image, separately Outside, palm and arm can all be remained when hand is identified, the human hand focus point when hand moves straight up Moving down can cause to misidentify.

The content of the invention

The present invention is directed to problems of the prior art, there is provided a kind of recognition methods of gesture moving direction, can The recognition accuracy of gesture moving direction is improved, and stability is good, recognition efficiency is high.

The technical scheme that the present invention proposes with regard to above-mentioned technical problem is as follows:

On the one hand, the present invention provides a kind of recognition methods of gesture moving direction, including:

In detection cycle, multiframe images of gestures is obtained, and hand region is identified from every frame images of gestures;

Hand palm or the back of the hand center of the positioning per frame hand region, and obtain palm or the back of the hand radius;

According to the change of palm or the back of the hand center and palm or the back of the hand radius between the multiframe images of gestures, identification Go out the moving direction of gesture.

Further, it is described to identify hand region from every frame images of gestures, specifically include:

Medium filtering and gaussian filtering are carried out to every frame images of gestures, then skin color segmentation is carried out to filtered image, so The morphological operation that the image after segmentation is expanded and corroded afterwards obtains the contour images after binaryzation;By the profile diagram Contour area is more than hand region of the pixel region of predetermined threshold value as the images of gestures as in.

Further, palm or the back of the hand center of the positioning per frame hand region, and obtain palm or the back of the hand half Footpath, specifically include:

Obtained based on distance transform algorithm per the maximum inscribed circle of radius in frame hand region, by the center of circle of the inscribed circle As the palm or the back of the hand center, and using the radius of the inscribed circle as the palm or the back of the hand radius.

Further, it is described according to palm or the back of the hand center between the multiframe images of gestures and palm or the back of the hand half The change in footpath, the moving direction of gesture is identified, is specifically included:

According to the palm of the multiframe images of gestures or the back of the hand center and palm or the back of the hand radius, palm or hand are calculated First accumulative frame number, palm or the back of the hand center of the center in the horizontal direction with Similar trend is carried on the back vertical On direction the second accumulative frame number and palm or the back of the hand radius with Similar trend with Similar trend the 3rd Accumulative frame number, and calculate the starting two field picture in the multiframe images of gestures and end two field picture in the horizontal direction first inclined Shifting value and the second deviant in vertical direction;

According to the described first accumulative frame number, the second accumulative frame number, the 3rd accumulative frame number, first deviant With second deviant, the moving direction of gesture is identified.

Further, the images of gestures is obtained by camera;

It is described inclined according to the described first accumulative frame number, the second accumulative frame number, the 3rd accumulative frame number, the first deviant and second Shifting value, the moving direction of gesture is identified, is specifically included:

Described first accumulative frame number, the second accumulative frame number and the 3rd accumulative frame number are distinguished corresponding pre- If frame number threshold value is compared;

If the described first accumulative frame number and the second accumulative frame number are respectively smaller than its corresponding frame number threshold value, the described 3rd Accumulative frame number is more than its corresponding frame number threshold value, and the palm or the back of the hand radius be in the trend that diminishes, then judgement gesture is to vertical Plane where the camera simultaneously moves away from the direction of the camera;

If the described first accumulative frame number and the second accumulative frame number are respectively smaller than its corresponding frame number threshold value, the described 3rd Accumulative frame number is more than its corresponding frame number threshold value, and the palm or the back of the hand radius are in change trend, then judgement gesture is to vertical Plane where the camera simultaneously moves towards the direction of the camera;

If the described second accumulative frame number and the 3rd accumulative frame number are respectively smaller than its corresponding frame number threshold value, described first Accumulative frame number is more than its corresponding frame number threshold value, and first deviant is negative value, then judges that gesture level is moved to the left;

If the described second accumulative frame number and the 3rd accumulative frame number are respectively smaller than its corresponding frame number threshold value, described first Accumulative frame number is more than its corresponding frame number threshold value, and first deviant is on the occasion of then judgement gesture level moves right;

If the described first accumulative frame number and the 3rd accumulative frame number are respectively smaller than its corresponding frame number threshold value, described second Accumulative frame number is more than its corresponding frame number threshold value, and second deviant is negative value, then judges that gesture moves straight up;

If the described first accumulative frame number and the 3rd accumulative frame number are respectively smaller than its corresponding frame number threshold value, described second Accumulative frame number is more than its corresponding frame number threshold value, and second deviant is on the occasion of then judgement gesture moves straight down.

On the other hand, the present invention provides a kind of identifying system of gesture moving direction, including:

Image collection module, in detection cycle, obtaining multiframe images of gestures, and identified from every frame images of gestures Go out hand region;

Locating module, for positioning palm or the back of the hand center of each hand region, and obtain palm or the back of the hand half Footpath;And

Gesture recognition module, for according to palm or the back of the hand center and palm or hand between the multiframe images of gestures The change of radius is carried on the back, identifies the moving direction of gesture.

Further, described image acquisition module specifically includes:

Image segmentation unit, for carrying out medium filtering and gaussian filtering to every frame images of gestures, then to the figure after filtering As carrying out skin color segmentation, the morphological operation that then image after segmentation is expanded and corroded obtains the profile after binaryzation Image;And

Hand region acquiring unit, the pixel region for the contour images contour area to be more than to predetermined threshold value are made For the hand region of the images of gestures.

Further, the locating module is specifically used for:

The inscribed circle that radius is maximum in each hand region is obtained based on distance transform algorithm, by the center of circle of the inscribed circle As the palm or the back of the hand center, and using the radius of the inscribed circle as the palm or the back of the hand radius.

Further, the gesture recognition module specifically includes:

Computing unit, for the palm according to the multiframe images of gestures or the back of the hand center and palm or the back of the hand half Footpath, calculate the first accumulative frame number, palm or the hand of palm or the back of the hand center in the horizontal direction with Similar trend Second accumulative frame number and palm or the back of the hand radius of the center in the vertical direction with Similar trend is carried on the back with phase The 3rd with variation tendency adds up frame number, and calculates starting two field picture with terminating the first deviant of two field picture in the horizontal direction And the second deviant in vertical direction;And

Recognition unit, for according to the described first accumulative frame number, the second accumulative frame number, the 3rd accumulative frame number, First deviant and second deviant, identify the moving direction of gesture.

Further, the images of gestures is obtained by camera;

The recognition unit specifically includes:

Comparing subunit, for the described first accumulative frame number, the second accumulative frame number and the described 3rd to be added up into frame number Corresponding default frame number threshold value is compared respectively;

First judges subelement, if it is corresponding to be respectively smaller than it for the described first accumulative frame number and the second accumulative frame number Frame number threshold value, the described 3rd accumulative frame number is more than its corresponding frame number threshold value, and the palm or the back of the hand radius are in diminish Gesture, then plane where judging gesture to the vertical camera simultaneously move away from the direction of the camera;

Second judges subelement, if it is corresponding to be respectively smaller than it for the described first accumulative frame number and the second accumulative frame number Frame number threshold value, the described 3rd accumulative frame number is more than its corresponding frame number threshold value, and the palm or the back of the hand radius become greatly in change Gesture, then plane where judging gesture to the vertical camera simultaneously move towards the direction of the camera;

3rd judges subelement, if it is corresponding to be respectively smaller than it for the described second accumulative frame number and the 3rd accumulative frame number Frame number threshold value, the described first accumulative frame number is more than its corresponding frame number threshold value, and first deviant is negative value, then judges The moving direction of gesture is moved to the left for level;

4th judges subelement, if it is corresponding to be respectively smaller than it for the described second accumulative frame number and the 3rd accumulative frame number Frame number threshold value, the described first accumulative frame number is more than its corresponding frame number threshold value, and first deviant is on the occasion of then judging The moving direction of gesture moves right for level;

5th judges subelement, if it is corresponding to be respectively smaller than it for the described first accumulative frame number and the 3rd accumulative frame number Frame number threshold value, the described second accumulative frame number is more than its corresponding frame number threshold value, and second deviant is negative value, then judges The moving direction of gesture is to move straight up;And

6th judges subelement, if it is corresponding to be respectively smaller than it for the described first accumulative frame number and the 3rd accumulative frame number Frame number threshold value, the described second accumulative frame number is more than its corresponding frame number threshold value, and second deviant is on the occasion of then judging The moving direction of gesture is to move straight down.

The beneficial effect that technical scheme provided in an embodiment of the present invention is brought is:

After the hand region in identifying images of gestures, the palm or the back of the hand center and palm of hand region are obtained Or the back of the hand radius, and then by the change of palm or the back of the hand center and palm between multiframe images of gestures in detection cycle Or the change of the back of the hand radius, you can judge the moving direction of gesture, gesture can be identified in front and rear all directions up and down Moving direction, and recognition accuracy is high, stability is good, recognition efficiency is high.

Brief description of the drawings

Technical scheme in order to illustrate the embodiments of the present invention more clearly, make required in being described below to embodiment Accompanying drawing is briefly described, it should be apparent that, drawings in the following description are only some embodiments of the present invention, for For those of ordinary skill in the art, on the premise of not paying creative work, other can also be obtained according to these accompanying drawings Accompanying drawing.

Fig. 1 is the schematic flow sheet of the recognition methods of gesture moving direction provided in an embodiment of the present invention;

Fig. 2 is the structural representation of the identifying system of gesture moving direction provided in an embodiment of the present invention.

Embodiment

To make the object, technical solutions and advantages of the present invention clearer, below in conjunction with accompanying drawing to embodiment party of the present invention Formula is described in further detail.

The embodiments of the invention provide a kind of recognition methods of gesture moving direction, referring to Fig. 1, this method includes:

S1, in detection cycle, obtain multiframe images of gestures, and identify hand region from every frame images of gestures;

The palm or the back of the hand center of S2, positioning per frame hand region, and obtain palm or the back of the hand radius;

S3, the change according to palm or the back of the hand center and palm or the back of the hand radius between the multiframe images of gestures, Identify the moving direction of gesture.

It should be noted that the present embodiment can pre-set a detection cycle, it is more to be continuously acquired in detection cycle Frame images of gestures is identified.Wherein, obtained per frame images of gestures by single camera (monocular) or dual camera (binocular), inspection The survey cycle may be configured as less value, for example, 10 frame images of gestures are obtained in a detection cycle, to improve the identification of gesture Efficiency.

A frame images of gestures is often obtained, i.e., hand region is identified from the images of gestures, and then from the hand region Orient palm or the back of the hand center and its corresponding palm or the back of the hand radius.Pass through palm or the back of the hand in every frame images of gestures The variation tendency and palm of center or the variation tendency of the back of the hand radius, you can identify the moving direction of gesture.This implementation Example may recognize that movement of the gesture in fore-and-aft direction, by palm by the detection to palm or the back of the hand radius change trend Or the detection of the back of the hand center variation tendency, the accuracy rate of gesture moving direction can be improved.

Further, it is described to identify hand region from every frame images of gestures, specifically include:

Medium filtering and gaussian filtering are carried out to every frame images of gestures, then skin color segmentation is carried out to the image after filtering, so The morphological operation that the image after segmentation is expanded and corroded afterwards obtains the contour images after binaryzation, by the profile diagram Contour area is more than hand region of the pixel region of predetermined threshold value as the images of gestures as in.

Specifically, the detailed step of the hand region identification is as follows:

Medium filtering and gaussian filtering are carried out to every frame images of gestures;

Filtered image is transformed into YCrCb color spaces and carries out skin color segmentation, if the Color Channel of pixel meets Cr ∈ [133,173] and Cb ∈ [77,127], then be set to 255 by the pixel value of the pixel, be otherwise set to 0;

The morphological operation that image after segmentation is expanded and corroded obtains the contour images after binaryzation;

Contour area is more than to hand region of the pixel region as the images of gestures of predetermined threshold value 10000.

Further, the palm of each hand region of positioning or the back of the hand center, and obtain palm or the back of the hand half Footpath, specifically include:

The inscribed circle that radius is maximum in each hand region is obtained based on distance transform algorithm, by the center of circle of the inscribed circle As the palm or the back of the hand center, and using the radius of the inscribed circle as the palm or the back of the hand radius.

It should be noted that distance transform algorithm be calculate a bianry image in non-zero pixels point to recently zero pixel The distance of point.After the bianry image of hand region is obtained, radius maximum in hand region is found using distance change algorithm Inscribed circle, the center of circle of inscribed circle is palm or the back of the hand center, and the radius of inscribed circle is palm or the back of the hand radius.

Further, it is described according to palm or the back of the hand center between the multiframe images of gestures and palm or the back of the hand half The change in footpath, the moving direction of gesture is identified, is specifically included:

According to the palm of the multiframe images of gestures or the back of the hand center and palm or the back of the hand radius, palm or hand are calculated First accumulative frame number, palm or the back of the hand center of the center in the horizontal direction with Similar trend is carried on the back vertical On direction the second accumulative frame number and palm or the back of the hand radius with Similar trend with Similar trend the 3rd Accumulative frame number, and calculate the starting two field picture in the multiframe images of gestures and end two field picture in the horizontal direction first inclined Shifting value and the second deviant in vertical direction;

According to the described first accumulative frame number, the second accumulative frame number, the 3rd accumulative frame number, first deviant With second deviant, the moving direction of gesture is identified.

Further, the images of gestures is obtained by camera;

It is described inclined according to the described first accumulative frame number, the second accumulative frame number, the 3rd accumulative frame number, the first deviant and second Shifting value, the moving direction of gesture is identified, is specifically included:

Described first accumulative frame number, the second accumulative frame number and the 3rd accumulative frame number are distinguished corresponding pre- If frame number threshold value is compared;

If the described first accumulative frame number and the second accumulative frame number are respectively smaller than its corresponding frame number threshold value, the described 3rd Accumulative frame number is more than its corresponding frame number threshold value, and the palm or the back of the hand radius be in the trend that diminishes, then judgement gesture is to vertical Plane where the camera simultaneously moves away from the direction of the camera;

If the described first accumulative frame number and the second accumulative frame number are respectively smaller than its corresponding frame number threshold value, the described 3rd Accumulative frame number is more than its corresponding frame number threshold value, and the palm or the back of the hand radius are in change trend, then judgement gesture is to vertical Plane where the camera simultaneously moves towards the direction of the camera;

If the described second accumulative frame number and the 3rd accumulative frame number are respectively smaller than its corresponding frame number threshold value, described first Accumulative frame number is more than its corresponding frame number threshold value, and first deviant is negative value, then judges that gesture level is moved to the left;

If the described second accumulative frame number and the 3rd accumulative frame number are respectively smaller than its corresponding frame number threshold value, described first Accumulative frame number is more than its corresponding frame number threshold value, and first deviant is on the occasion of then judgement gesture level moves right;

If the described first accumulative frame number and the 3rd accumulative frame number are respectively smaller than its corresponding frame number threshold value, described second Accumulative frame number is more than its corresponding frame number threshold value, and second deviant is negative value, then judges that gesture moves straight up;

If the described first accumulative frame number and the 3rd accumulative frame number are respectively smaller than its corresponding frame number threshold value, described second Accumulative frame number is more than its corresponding frame number threshold value, and second deviant is on the occasion of then judgement gesture moves straight down.

It should be noted that palm or the back of the hand move up in the side vertical with plane where camera, palm or the back of the hand Size change occurs, so as to cause palm or the back of the hand radius that large change occurs, and palm or the back of the hand are horizontal and perpendicular Palm or the back of the hand radius change are smaller when Nogata moves up.Therefore, in a detection cycle, a frame gesture figure is often obtained Picture, then present frame images of gestures and the palm of former frame images of gestures or the variable quantity of the back of the hand radius are recorded, and then obtain detection There is the accumulative change frame number of Similar trend, the as the 3rd accumulative frame number in cycle.If palm or the back of the hand radius are diminishing The 3rd accumulative frame number in trend reaches default frame number threshold value, such as 6 frames, then judges that gesture plane where vertical camera is carried on the back Move, that is, be moved rearwards from camera;If the 3rd accumulative frame number of palm or the back of the hand radius in change trend reaches default Frame number threshold value, such as 6 frames, then judge that gesture plane where vertical camera moves towards camera, that is, move forward.

For movement of the gesture in direction up and down deviant can be calculated to realize by establishing rectangular coordinate system.Example Such as, using the palm of start frame images of gestures or the back of the hand center as origin, using horizontal right direction as x-axis direction, with vertical It is in downward direction y-axis direction, establishes rectangular coordinate system.In a detection cycle, a frame images of gestures is often obtained, then record is worked as The variable quantity of the palm or the back of the hand center of previous frame images of gestures and former frame images of gestures in the direction of the x axis, and then obtain There is the accumulative change frame number of Similar trend, the as first accumulative frame number in detection cycle in x-axis direction.Meanwhile record Terminate the palm of two field picture or the x coordinate of the back of the hand center in the detection cycle, the first deviant can be used as.In a detection In cycle, a frame images of gestures is often obtained, then in the palm or the back of the hand that record present frame images of gestures and former frame images of gestures The variable quantity of heart position in the y-axis direction, and then obtain the accumulative change in y-axis direction in detection cycle with Similar trend Change frame number, the as second accumulative frame number.Meanwhile record the palm for terminating two field picture in the detection cycle or the back of the hand center Y-coordinate, the second deviant can be used as.

If hand does not move where camera on the direction of plane, palm or the back of the hand center x coordinate are in Diminish trend, and the first accumulative frame number reaches default frame number threshold value, such as 6 frames, and the first deviant reaches -50 pixels, then judges hand Gesture is moved to the left in the horizontal direction;If hand does not move where camera on the direction of plane, in palm or the back of the hand Heart position x coordinate is in change trend, and the first accumulative frame number reaches default frame number threshold value, such as 6 frames, and the first deviant reaches+50 Pixel, then judge that gesture moves right in the horizontal direction;If hand does not move where camera on the direction of plane, Palm or the back of the hand center y-coordinate reach default frame number threshold value in the trend that diminishes, the second accumulative frame number, such as 6 frames, and second is inclined Shifting value reaches -50 pixels, then judges that gesture moves up in vertical direction;If hand is in the direction of the plane where camera On do not move, palm or the back of the hand center y-coordinate are in change trend, and the second accumulative frame number reaches default frame number threshold value, such as 6 Frame, and the second deviant reaches+50 pixels, then judges that gesture moves down in vertical direction.The present embodiment is a detection week The accumulative frame number with Similar trend reaches certain frame number threshold value (generally certain proportion of a detection cycle in phase Value, configurable) moving direction is just can determine that, it can not only prevent because shake causes to misidentify, improve stability, and calculate letter It is single, improve recognition efficiency.

The embodiment of the present invention is after the hand region in identifying images of gestures, in the palm or the back of the hand that obtain hand region Heart position and palm or the back of the hand radius, and then by palm between multiframe images of gestures in detection cycle or the back of the hand center Change and palm or the back of the hand radius change, you can judge the moving direction of gesture, gesture can be identified up and down The moving direction of front and rear all directions, and recognition accuracy is high, stability is good, recognition efficiency is high

Accordingly, the embodiments of the invention provide a kind of identifying system of gesture moving direction, above-mentioned gesture can be realized All flows of the recognition methods of moving direction, referring to Fig. 2, the system includes:

Image collection module 1, in detection cycle, obtaining multiframe images of gestures, and know from every frame images of gestures Hand region is not gone out;

Locating module 2, for positioning palm or the back of the hand center of each hand region, and obtain palm or the back of the hand half Footpath;And

Gesture recognition module 3, for according to palm or the back of the hand center and palm between the multiframe images of gestures or The change of the back of the hand radius, identify the moving direction of gesture.

Further, described image acquisition module specifically includes:

Image segmentation unit, for carrying out medium filtering and gaussian filtering to every frame images of gestures, then skin color segmentation is carried out, Then the morphological operation for the image after segmentation being expanded and being corroded obtains the contour images after binaryzation;And

Hand region acquiring unit, for contour area to be more than to the pixel region of predetermined threshold value as the gesture figure The hand region of picture.

Further, the locating module is specifically used for:

Obtained based on distance transform algorithm per the maximum inscribed circle of radius in frame hand region, by the center of circle of the inscribed circle As the palm or the back of the hand center, and using the radius of the inscribed circle as the palm or the back of the hand radius.

Further, the gesture recognition module specifically includes:

Computing unit, for the palm according to the multiframe images of gestures or the back of the hand center and palm or the back of the hand half Footpath, calculate the first accumulative frame number, palm or the hand of palm or the back of the hand center in the horizontal direction with Similar trend Second accumulative frame number and palm or the back of the hand radius of the center in the vertical direction with Similar trend is carried on the back with phase The 3rd with variation tendency adds up frame number, and calculates starting two field picture with terminating the first deviant of two field picture in the horizontal direction And the second deviant in vertical direction;And

Recognition unit, for according to the described first accumulative frame number, the second accumulative frame number, the 3rd accumulative frame number, First deviant and second deviant, identify the moving direction of gesture.

Further, the images of gestures is obtained by camera;

The recognition unit specifically includes:

Comparing subunit, for the described first accumulative frame number, the second accumulative frame number and the described 3rd to be added up into frame number Corresponding default frame number threshold value is compared respectively;

First judges subelement, if it is corresponding to be respectively smaller than it for the described first accumulative frame number and the second accumulative frame number Frame number threshold value, the described 3rd accumulative frame number is more than its corresponding frame number threshold value, and the palm or the back of the hand radius are in diminish Gesture, then plane where judging gesture to the vertical camera simultaneously move away from the direction of the camera;

Second judges subelement, if it is corresponding to be respectively smaller than it for the described first accumulative frame number and the second accumulative frame number Frame number threshold value, the described 3rd accumulative frame number is more than its corresponding frame number threshold value, and the palm or the back of the hand radius become greatly in change Gesture, then plane where judging gesture to the vertical camera simultaneously move towards the direction of the camera;

3rd judges subelement, if it is corresponding to be respectively smaller than it for the described second accumulative frame number and the 3rd accumulative frame number Frame number threshold value, the described first accumulative frame number is more than its corresponding frame number threshold value, and first deviant is negative value, then judges Gesture level is moved to the left;

4th judges subelement, if it is corresponding to be respectively smaller than it for the described second accumulative frame number and the 3rd accumulative frame number Frame number threshold value, the described first accumulative frame number is more than its corresponding frame number threshold value, and first deviant is on the occasion of then judging Gesture level moves right;

5th judges subelement, if it is corresponding to be respectively smaller than it for the described first accumulative frame number and the 3rd accumulative frame number Frame number threshold value, the described second accumulative frame number is more than its corresponding frame number threshold value, and second deviant is negative value, then judges Gesture moves straight up;And

6th judges subelement, if it is corresponding to be respectively smaller than it for the described first accumulative frame number and the 3rd accumulative frame number Frame number threshold value, the described second accumulative frame number is more than its corresponding frame number threshold value, and second deviant is on the occasion of then judging Gesture moves straight down.

The embodiment of the present invention is after the hand region in identifying images of gestures, in the palm or the back of the hand that obtain hand region Heart position and palm or the back of the hand radius, and then by palm between multiframe images of gestures in detection cycle or the back of the hand center Change and palm or the back of the hand radius change, you can judge the moving direction of gesture, gesture can be identified up and down The moving direction of front and rear all directions, and recognition accuracy is high, stability is good, recognition efficiency is high.

The foregoing is only presently preferred embodiments of the present invention, be not intended to limit the invention, it is all the present invention spirit and Within principle, any modification, equivalent substitution and improvements made etc., it should be included in the scope of the protection.

Claims (10)

  1. A kind of 1. recognition methods of gesture moving direction, it is characterised in that including:
    In detection cycle, multiframe images of gestures is obtained, and hand region is identified from every frame images of gestures;
    Palm or the back of the hand center of the positioning per frame hand region, and obtain palm or the back of the hand radius;
    According to the change of palm or the back of the hand center and palm or the back of the hand radius between the multiframe images of gestures, identification is sold The moving direction of gesture.
  2. 2. the recognition methods of gesture moving direction as claimed in claim 1, it is characterised in that described from every frame images of gestures Hand region is identified, is specifically included:
    Medium filtering and gaussian filtering are carried out to every frame images of gestures, and skin color segmentation is carried out to filtered image;
    The morphological operation that image after segmentation is expanded and corroded obtains the contour images after binaryzation;
    Contour area in the contour images is more than to hand area of the pixel region as the images of gestures of predetermined threshold value Domain.
  3. 3. the recognition methods of gesture moving direction as claimed in claim 1, it is characterised in that the positioning is per frame hand region Palm or the back of the hand center, and obtain palm or the back of the hand radius, specifically include:
    The maximum inscribed circle of radius in being obtained based on distance transform algorithm per frame hand region, using the center of circle of the inscribed circle as Palm or the back of the hand center, and using the radius of the inscribed circle as the palm or the back of the hand radius.
  4. 4. the recognition methods of gesture moving direction as claimed in claim 1, it is characterised in that described according to the multiframe gesture The change of palm or the back of the hand center and palm or the back of the hand radius between image, the moving direction of gesture is identified, specific bag Include:
    According to the palm of the multiframe images of gestures or the back of the hand center and palm or the back of the hand radius, calculate in palm or the back of the hand First accumulative frame number, palm or the back of the hand center of the heart position in the horizontal direction with Similar trend is in vertical direction Upper the second accumulative frame number and palm or the back of the hand radius the with Similar trend the 3rd with Similar trend is accumulative Frame number, and the starting two field picture in the multiframe images of gestures is calculated with terminating the first deviant of two field picture in the horizontal direction And the second deviant in vertical direction;
    According to the described first accumulative frame number, the second accumulative frame number, the 3rd accumulative frame number, first deviant and the institute The second deviant is stated, identifies the moving direction of gesture.
  5. 5. the recognition methods of gesture moving direction as claimed in claim 4, it is characterised in that the images of gestures passes through shooting Head obtains;
    It is described to be offset according to the described first accumulative frame number, the second accumulative frame number, the 3rd accumulative frame number, the first deviant and second Value, identifies the moving direction of gesture, specifically includes:
    Described first accumulative frame number, the second accumulative frame number and the 3rd accumulative frame number are distinguished into corresponding default frame Number threshold value is compared;
    If the described first accumulative frame number and the second accumulative frame number are respectively smaller than its corresponding frame number threshold value, the described 3rd is accumulative Frame number is more than its corresponding frame number threshold value, and the palm or the back of the hand radius be in the trend that diminishes, then judges gesture to vertical described Plane where camera simultaneously moves away from the direction of the camera;
    If the described first accumulative frame number and the second accumulative frame number are respectively smaller than its corresponding frame number threshold value, the described 3rd is accumulative Frame number is more than its corresponding frame number threshold value, and the palm or the back of the hand radius are in change trend, then judges gesture to vertical described Plane where camera simultaneously moves towards the direction of the camera;
    If the described second accumulative frame number and the 3rd accumulative frame number are respectively smaller than its corresponding frame number threshold value, described first is accumulative Frame number is more than its corresponding frame number threshold value, and first deviant is negative value, then judges that gesture level is moved to the left;
    If the described second accumulative frame number and the 3rd accumulative frame number are respectively smaller than its corresponding frame number threshold value, described first is accumulative Frame number is more than its corresponding frame number threshold value, and first deviant is on the occasion of then judgement gesture level moves right;
    If the described first accumulative frame number and the 3rd accumulative frame number are respectively smaller than its corresponding frame number threshold value, described second is accumulative Frame number is more than its corresponding frame number threshold value, and second deviant is negative value, then judges that gesture moves straight up;
    If the described first accumulative frame number and the 3rd accumulative frame number are respectively smaller than its corresponding frame number threshold value, described second is accumulative Frame number is more than its corresponding frame number threshold value, and second deviant is on the occasion of then judgement gesture moves straight down.
  6. A kind of 6. identifying system of gesture moving direction, it is characterised in that including:
    Image collection module, in detection cycle, obtaining multiframe images of gestures, and identify and sell from every frame images of gestures Portion region;
    Locating module, for positioning palm or the back of the hand center of every frame hand region, and obtain palm or the back of the hand radius;With And
    Gesture recognition module, for according to palm or the back of the hand center between the multiframe images of gestures and palm or the back of the hand half The change in footpath, identify the moving direction of gesture.
  7. 7. the identifying system of gesture moving direction as claimed in claim 6, it is characterised in that described image acquisition module is specific Including:
    Image segmentation unit, for carrying out medium filtering and gaussian filtering to every frame images of gestures, and filtered image is entered Row skin color segmentation, and then the morphological operation that the image after segmentation is expanded and corroded obtains the profile diagram after binaryzation Picture;And
    Hand region acquiring unit, for using contour area in the contour images be more than predetermined threshold value pixel region as The hand region of the images of gestures.
  8. 8. the identifying system of gesture moving direction as claimed in claim 6, it is characterised in that the locating module is specifically used In:
    The inscribed circle that radius is maximum in each hand region is obtained based on distance transform algorithm, using the center of circle of the inscribed circle as Palm or the back of the hand center, and using the radius of the inscribed circle as the palm or the back of the hand radius.
  9. 9. the identifying system of gesture moving direction as claimed in claim 6, it is characterised in that the gesture recognition module is specific Including:
    Computing unit, for the palm according to the multiframe images of gestures or the back of the hand center and palm or the back of the hand radius, meter Calculate the first accumulative frame number, palm or the back of the hand center of palm or the back of the hand center in the horizontal direction with Similar trend Second accumulative frame number and palm or the back of the hand radius of the position in the vertical direction with Similar trend is with identical change 3rd accumulative frame number of trend, and calculate starting two field picture with terminate two field picture the first deviant in the horizontal direction and Second deviant of vertical direction;And
    Recognition unit, for according to the described first accumulative frame number, the second accumulative frame number, the 3rd accumulative frame number, described First deviant and second deviant, identify the moving direction of gesture.
  10. 10. the identifying system of gesture moving direction as claimed in claim 9, it is characterised in that the images of gestures is by taking the photograph As head obtains;
    The recognition unit specifically includes:
    Comparing subunit, for the described first accumulative frame number, the second accumulative frame number and the 3rd accumulative frame number to be distinguished Corresponding default frame number threshold value is compared;
    First judges subelement, if being respectively smaller than its corresponding frame for the described first accumulative frame number and the second accumulative frame number Number threshold value, the described 3rd accumulative frame number is more than its corresponding frame number threshold value, and the palm or the back of the hand radius are in the trend that diminishes, then Plane where judging gesture to the vertical camera simultaneously moves away from the direction of the camera;
    Second judges subelement, if being respectively smaller than its corresponding frame for the described first accumulative frame number and the second accumulative frame number Number threshold value, the described 3rd accumulative frame number is more than its corresponding frame number threshold value, and the palm or the back of the hand radius are in change trend, then Plane where judging gesture to the vertical camera simultaneously moves towards the direction of the camera;
    3rd judges subelement, if being respectively smaller than its corresponding frame for the described second accumulative frame number and the 3rd accumulative frame number Number threshold value, the described first accumulative frame number is more than its corresponding frame number threshold value, and first deviant is negative value, then judges gesture Level is moved to the left;
    4th judges subelement, if being respectively smaller than its corresponding frame for the described second accumulative frame number and the 3rd accumulative frame number Number threshold values, the described first accumulative frame number are more than its corresponding frame number threshold value, and first deviant is on the occasion of then judging gesture Level moves right;
    5th judges subelement, if being respectively smaller than its corresponding frame for the described first accumulative frame number and the 3rd accumulative frame number Number threshold value, the described second accumulative frame number is more than its corresponding frame number threshold value, and second deviant is negative value, then judges gesture Move straight up;And
    6th judges subelement, if being respectively smaller than its corresponding frame for the described first accumulative frame number and the 3rd accumulative frame number Number threshold values, the described second accumulative frame number are more than its corresponding frame number threshold value, and second deviant is on the occasion of then judging gesture Move straight down.
CN201710884796.5A 2017-09-26 2017-09-26 A kind of recognition methods of gesture moving direction and system CN107589850A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710884796.5A CN107589850A (en) 2017-09-26 2017-09-26 A kind of recognition methods of gesture moving direction and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710884796.5A CN107589850A (en) 2017-09-26 2017-09-26 A kind of recognition methods of gesture moving direction and system

Publications (1)

Publication Number Publication Date
CN107589850A true CN107589850A (en) 2018-01-16

Family

ID=61047884

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710884796.5A CN107589850A (en) 2017-09-26 2017-09-26 A kind of recognition methods of gesture moving direction and system

Country Status (1)

Country Link
CN (1) CN107589850A (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108197596A (en) * 2018-01-24 2018-06-22 京东方科技集团股份有限公司 A kind of gesture identification method and device
CN109409277A (en) * 2018-10-18 2019-03-01 北京旷视科技有限公司 Gesture identification method, device, intelligent terminal and computer storage medium
CN109598198A (en) * 2018-10-31 2019-04-09 深圳市商汤科技有限公司 The method, apparatus of gesture moving direction, medium, program and equipment for identification
US10261593B2 (en) * 2016-04-13 2019-04-16 Volkswagen Aktiengesellschaft User interface, means of movement, and methods for recognizing a user's hand
CN109635665A (en) * 2018-11-16 2019-04-16 惠州拓邦电气技术有限公司 A kind of electric appliance gestural control method, device and kitchen appliance
CN111263084A (en) * 2018-11-30 2020-06-09 北京字节跳动网络技术有限公司 Video-based gesture jitter detection method, device, terminal and medium

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040141634A1 (en) * 2002-10-25 2004-07-22 Keiichi Yamamoto Hand pattern switch device
CN102368290A (en) * 2011-09-02 2012-03-07 华南理工大学 Hand gesture identification method based on finger advanced characteristic
CN104331151A (en) * 2014-10-11 2015-02-04 中国传媒大学 Optical flow-based gesture motion direction recognition method
CN105393281A (en) * 2013-08-02 2016-03-09 三菱电机株式会社 Gesture determination device and method, gesture-operated device, program, and recording medium
CN105825193A (en) * 2016-03-25 2016-08-03 乐视控股(北京)有限公司 Method and device for position location of center of palm, gesture recognition device and intelligent terminals
CN106971131A (en) * 2016-01-14 2017-07-21 芋头科技(杭州)有限公司 A kind of gesture identification method based on center

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040141634A1 (en) * 2002-10-25 2004-07-22 Keiichi Yamamoto Hand pattern switch device
CN102368290A (en) * 2011-09-02 2012-03-07 华南理工大学 Hand gesture identification method based on finger advanced characteristic
CN105393281A (en) * 2013-08-02 2016-03-09 三菱电机株式会社 Gesture determination device and method, gesture-operated device, program, and recording medium
CN104331151A (en) * 2014-10-11 2015-02-04 中国传媒大学 Optical flow-based gesture motion direction recognition method
CN106971131A (en) * 2016-01-14 2017-07-21 芋头科技(杭州)有限公司 A kind of gesture identification method based on center
CN105825193A (en) * 2016-03-25 2016-08-03 乐视控股(北京)有限公司 Method and device for position location of center of palm, gesture recognition device and intelligent terminals

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10261593B2 (en) * 2016-04-13 2019-04-16 Volkswagen Aktiengesellschaft User interface, means of movement, and methods for recognizing a user's hand
CN108197596A (en) * 2018-01-24 2018-06-22 京东方科技集团股份有限公司 A kind of gesture identification method and device
US10803304B2 (en) 2018-01-24 2020-10-13 Boe Technology Group Co., Ltd. Gesture recognition method, device, apparatus, and storage medium
CN109409277A (en) * 2018-10-18 2019-03-01 北京旷视科技有限公司 Gesture identification method, device, intelligent terminal and computer storage medium
CN109598198A (en) * 2018-10-31 2019-04-09 深圳市商汤科技有限公司 The method, apparatus of gesture moving direction, medium, program and equipment for identification
CN109635665A (en) * 2018-11-16 2019-04-16 惠州拓邦电气技术有限公司 A kind of electric appliance gestural control method, device and kitchen appliance
CN111263084A (en) * 2018-11-30 2020-06-09 北京字节跳动网络技术有限公司 Video-based gesture jitter detection method, device, terminal and medium

Similar Documents

Publication Publication Date Title
US9785245B2 (en) Image processing apparatus, image processing method, and program for recognizing a gesture
AU2013200807B2 (en) Method and portable terminal for correcting gaze direction of user in image
Biswas et al. Gesture recognition using microsoft kinect®
Khan et al. Hand gesture recognition: a literature review
CN102360421B (en) Face identification method and system based on video streaming
CN103353935B (en) A kind of 3D dynamic gesture identification method for intelligent domestic system
Itoh et al. Interaction-free calibration for optical see-through head-mounted displays based on 3d eye localization
US9934573B2 (en) Technologies for adjusting a perspective of a captured image for display
US7391888B2 (en) Head pose assessment methods and systems
JP6255486B2 (en) Method and system for information recognition
US7324693B2 (en) Method of human figure contour outlining in images
CN105488815B (en) A kind of real-time objects tracking for supporting target size to change
KR101877570B1 (en) Apparatus for setting parking position based on around view image and method thereof
US20130128001A1 (en) Method and system for detecting object on a road
JP5699788B2 (en) Screen area detection method and system
US9014459B2 (en) Identification method for valuable file and identification device thereof
US9025875B2 (en) People counting device, people counting method and people counting program
WO2015187294A1 (en) Technologies for viewer attention area estimation
WO2014189685A1 (en) Motion-assisted visual language for human computer interfaces
CN104751142B (en) A kind of natural scene Method for text detection based on stroke feature
CN103116896B (en) Visual saliency model based automatic detecting and tracking method
US9710109B2 (en) Image processing device and image processing method
US20060281969A1 (en) System and method for operation without touch by operators
SE1550243A1 (en) Improved identification of a gesture
CN102063618B (en) Dynamic gesture identification method in interactive system

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination