CN103544469A - Fingertip detection method and device based on palm ranging - Google Patents
Fingertip detection method and device based on palm ranging Download PDFInfo
- Publication number
- CN103544469A CN103544469A CN201310312446.3A CN201310312446A CN103544469A CN 103544469 A CN103544469 A CN 103544469A CN 201310312446 A CN201310312446 A CN 201310312446A CN 103544469 A CN103544469 A CN 103544469A
- Authority
- CN
- China
- Prior art keywords
- point
- finger tip
- palm
- barycenter
- gesture
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Image Analysis (AREA)
- Collating Specific Patterns (AREA)
Abstract
The invention discloses a fingertip detection method and device based on palm ranging. The method includes marking initially screened fingertip points after cluster, corroding the gesture communicating area, extracting the centroid of the palm area after corrosion, making a circumcircle on the palm area after corrosion, and acquiring the radius of the circumcircle; calculating a first ratio of distance between an initially-screened fingertip point to the centroid and the radius of the circumcircle, and comparing the first ratio with a preset first threshold; when the first ratio is smaller than the first threshold, determining the corresponding initially-screened fingertip point to be a false minutiae point and filtering; when the first ratio is larger than or equal to the first threshold, determining the corresponding initially-screened fingertip point to be a detected fingertip point; and outputting detection results. Thus, detection accuracy of real-time fingertip detection is improved, and fingertip false detection rate is decreased.
Description
Technical field
The present invention relates to Machine Vision Recognition field, in particular a kind of Fingertip Detection and device based on centre of the palm range finding.
Background technology
In existing finger tip context of detection, after generally tracing into gesture by certain track algorithm, then by after retaining hand connected domain after image opening and closing operation, binary conversion treatment and connected domain judgement, start to find finger tip point.By curvature method edge point, carry out curvature calculating again, will be marked compared with deep camber point, be decided to be possible finger position of cusp, then filter out finger tip point by the whole bag of tricks such as cluster etc.But when actual machine visual identity gesture, because the freedom of movement of staff is large, change of shape is large, usually clench fist inadequate because pointing degree of holding, cause some small salient points to be mistaken for the situation of finger tip, also have when having finger to stretch out the error detection problem that half bending of the hand refers to, these problems greatly reduce the degree of accuracy that finger tip detects, and are difficult to meet the demand of many real time human-machine interaction application.
Therefore, prior art has yet to be improved and developed.
Summary of the invention
The technical problem to be solved in the present invention is, for the above-mentioned defect of prior art, provides a kind of Fingertip Detection and device based on centre of the palm range finding, is intended to solve the not high problem of accuracy of detection that existing finger tip detects.
The technical scheme that technical solution problem of the present invention adopts is as follows:
A Fingertip Detection based on centre of the palm range finding, wherein, comprises the following steps:
A, the original image that comprises gesture gathering is carried out to noise reduction process, and by the image through noise reduction process by unrestrained water completion method judgement connected domain, find out gesture connected region;
The marginal point of B, traversal gesture connected region obtains marginal point sequence, the curvature value of every bit in edge calculation point sequence, and the curvature value of every bit and default curvature threshold are compared, find out the marginal point that curvature value is greater than curvature threshold, removing in the marginal point that curvature value is greater than curvature threshold is the point of finger recess again, and the marginal point that remaining curvature value is greater than to curvature threshold is labeled as possible finger tip point;
C, the salient point sequence at each possible finger tip point place is independently divided into new point set, respectively new point set is carried out to cluster, from described point, concentrate and find out finger tip point, and be labeled as preliminary screening finger tip point out;
D, gesture connected region is corroded, extract the barycenter of the palm area after corrosion, and the palm area after corrosion is done to circumscribed circle, obtain external radius of a circle;
E, calculate the first ratio that each preliminary screening finger tip is out put distance and the circumradius of barycenter, and described the first ratio and predetermined first threshold are compared, when the first ratio is less than first threshold, judge that corresponding preliminary screening finger tip point is out as erroneous judgement point filtering; When the first ratio is more than or equal to first threshold, judge that corresponding preliminary screening finger tip point is out as the finger tip point detecting; And output detections result.
The described Fingertip Detection based on centre of the palm range finding wherein, also comprises after described step e:
F, calculate the distance that each preliminary screening finger tip is out put barycenter, find out the maximum finger tip point of distance, calculating other finger tip except the finger tip point of ultimate range is put the second ratio of distance and the ultimate range of barycenter, described the second ratio and predetermined Second Threshold are compared, when the second ratio is less than Second Threshold, judge that corresponding preliminary screening finger tip point is out as erroneous judgement point filtering; When the second ratio is more than or equal to Second Threshold, judge that corresponding preliminary screening finger tip point is out as the finger tip point detecting; And output detections result.
The described Fingertip Detection based on centre of the palm range finding, wherein, described steps A specifically comprises:
A1, the original image that comprises gesture by camera collection, carry out smoothing processing to described original image, to remove unnecessary noise; Carry out again binary conversion treatment and image opening and closing operation, to eliminate cavity and further to eliminate noise;
A2, by unrestrained water completion method, the image through noise reduction process is carried out to connected domain judgement, calculate the area of each connected region;
A3, find out the connected region of area maximum and be left gesture connected region, remove small size connected domain.
The described Fingertip Detection based on centre of the palm range finding, wherein, described step B specifically comprises:
B1, by edge detection method, detect the marginal point of gesture connected region, the marginal point of traversal gesture connected region obtains marginal point sequence M;
B2, the size by gesture connected region judge the distance between staff and camera, according to the distance between staff and camera, a shaping are set and count k, and wherein, k is positive integer, then calculates any point in gesture marginal point sequence M
curvature value: get in gesture marginal point sequence with
2 equidistant points
with
, compute vector
with vector
the cosine value of angle is
the approximate curvature value of point, curvature value formula is
;
B3, the curvature value of every bit in marginal point sequence and default curvature threshold are compared, find out the marginal point that curvature value is greater than curvature threshold;
B4, judgement
whether the pixel value of line segment mid point is 0; If this pixel value is 0, be judged as finger recess filtering; The marginal point that remaining curvature value is greater than to curvature threshold is labeled as possible finger tip point.
The described Fingertip Detection based on centre of the palm range finding, wherein, described step C specifically comprises:
C1, the salient point sequence at each possible finger tip point place is independently divided into new point set, respectively new point set is carried out to cluster;
C2, get the some P of each some set mean curvature value maximum
max, the method for employing ellipse fitting, the oval external matrix four limit mid points after digital simulation are from P
maxnearest point, is labeled as preliminary screening finger tip point out by this point.
The described Fingertip Detection based on centre of the palm range finding, wherein, described step D specifically comprises:
D1, gesture connected region is repeatedly corroded, obtain one for the image of palm area, corrosion palm area is later extracted to barycenter O, this barycenter O point is the centre of the palm of palm area, and the formula of barycenter O is
,
, wherein, i, j is respectively x axial coordinate and the y axial coordinate of any one pixel in image; M, n are respectively the row and column of image,
for the x axial coordinate of barycenter,
y axial coordinate for barycenter;
D2, palm area is made to circumscribed circle, obtain the radius R of palm circumscribed circle.
A finger tip pick-up unit for centre of the palm range finding, wherein, comprising:
Gesture connected region generation module, for the original image that comprises gesture gathering is carried out to noise reduction process, and judges connected domain by the image through noise reduction process by unrestrained water completion method, finds out gesture connected region;
Curvature value computing module, obtains marginal point sequence for traveling through the marginal point of gesture connected region, the curvature value of every bit in edge calculation point sequence;
Curvature value contrast module, for comparing the curvature value of marginal point sequence every bit and default curvature threshold;
Groove is removed module, for removing curvature value, is greater than the marginal point of curvature threshold for the point of finger recess;
Cluster screening module, for the salient point sequence at each possible finger tip point place is independently divided into new point set, carries out cluster to new point set respectively, concentrates and finds out finger tip point, and be labeled as preliminary screening finger tip point out from described point;
Barycenter and circumradius acquisition module, for gesture connected region is corroded, extract the barycenter of the palm area after corrosion, and the palm area after corrosion done to circumscribed circle, obtains external radius of a circle;
The first ratio calculation module, puts the first ratio of distance and the circumradius of barycenter for calculating each preliminary screening finger tip out;
The first ratio contrast module, for comparing described the first ratio and predetermined first threshold.
The described finger tip pick-up unit based on centre of the palm range finding, wherein, also comprises:
The second ratio calculation module, puts the distance of barycenter for calculating each preliminary screening finger tip out, find out the maximum finger tip point of distance, calculates the second ratio that other finger tip except the finger tip point of ultimate range is put distance and the ultimate range of barycenter;
The second ratio contrast module, for comparing described the second ratio and predetermined Second Threshold.
The described finger tip pick-up unit based on centre of the palm range finding, wherein, described curvature value computing module comprises:
Marginal point sequence generating unit, for detect the marginal point of gesture connected region by edge detection method, the marginal point of traversal gesture connected region obtains marginal point sequence M;
Curvature value computing unit, for the size by gesture connected region, judge the distance between staff and camera, according to the distance between staff and camera, a shaping is set and counts k, wherein, k is positive integer, then calculates any point in gesture marginal point sequence M
curvature value: get in gesture marginal point sequence with
2 equidistant points
with
, compute vector
with vector
the cosine value of angle is
the approximate curvature value of point, curvature value formula is
.
The described finger tip pick-up unit based on centre of the palm range finding, wherein, described barycenter and circumradius acquisition module comprise:
Erosion unit, for gesture connected region is repeatedly corroded, obtains one for the image of palm area;
Centroid calculation unit, for corrosion palm area is later extracted to barycenter O, this barycenter O point is the centre of the palm of palm area, and the formula of barycenter O is
,
, wherein, i, j is respectively x axial coordinate and the y axial coordinate of any one pixel in image; M, n are respectively the row and column of image,
for the x axial coordinate of barycenter,
y axial coordinate for barycenter;
Circumradius computing unit, for to corrosion palm area is later made to circumscribed circle, obtains the radius R of palm circumscribed circle.
A kind of Fingertip Detection and device based on centre of the palm range finding provided by the present invention, some misjudgment phenomenon that exist in existing finger tip detection have effectively been solved, for example when clenching fist because the reasons such as finger degree of holding is inadequate can exist the problem that some small salient points are finger tip by error detection, its method is by mark preliminary screening finger tip point out after cluster, gesture connected region is corroded, extract the barycenter of the palm area after corrosion, and the palm area after corrosion is done to circumscribed circle, obtain external radius of a circle; Calculate each preliminary screening finger tip out and put the first ratio of distance and the circumradius of barycenter, and described the first ratio and predetermined first threshold are compared, when the first ratio is less than first threshold, judge that corresponding preliminary screening finger tip point is out as erroneous judgement point filtering; When the first ratio is more than or equal to first threshold, judge that corresponding preliminary screening finger tip point is out as the finger tip point detecting; And output detections result, thereby improved the accuracy of detection that real-time finger tip detects, reduced finger tip false drop rate, its implementation is simple, by software, realizes, and cost is lower.
Accompanying drawing explanation
Fig. 1 is the process flow diagram of Fingertip Detection the first preferred embodiment based on centre of the palm range finding provided by the invention.
Fig. 2 a is the schematic diagram of original image after noise reduction process in Fingertip Detection the first preferred embodiment based on centre of the palm range finding provided by the invention.
Fig. 2 b is that Fig. 2 a is through the palm contour shape figure after unrestrained water completion method and connected domain judgement.
Fig. 3 a is the schematic diagram of Fig. 2 b after edge detection method and curvature judgement.
Fig. 3 b is the schematic diagram that Fig. 3 a removes the possible finger tip point after finger recess.
Fig. 3 c is Fig. 3 b preliminary screening finger tip point schematic diagram out after cluster.
Fig. 4 a is the finger tip point schematic diagram of the centre of the palm district inclusion erroneous judgement point after corrosion in Fingertip Detection the first preferred embodiment based on centre of the palm range finding provided by the invention.
Fig. 4 b is according to the finger tip point schematic diagram after the first ratio and first threshold filtering erroneous judgement point in Fingertip Detection the first preferred embodiment based on centre of the palm range finding provided by the invention.
Fig. 5 is the process flow diagram of Fingertip Detection the second preferred embodiment based on centre of the palm range finding provided by the invention.
Fig. 6 a is that in Fingertip Detection the second preferred embodiment based on centre of the palm range finding provided by the invention, to comprise erroneous judgement point be the finger tip point schematic diagram of half dactylogryposis cusp.
Fig. 6 b is that Fig. 6 a removes the finger tip point schematic diagram after erroneous judgement point.
Fig. 7 is the structured flowchart of finger tip pick-up unit the first preferred embodiment based on centre of the palm range finding provided by the invention.
Fig. 8 is the structured flowchart of finger tip pick-up unit the second preferred embodiment based on centre of the palm range finding provided by the invention.
Embodiment
The invention provides a kind of Fingertip Detection and device based on centre of the palm range finding, for making object of the present invention, technical scheme and advantage clearer, clear and definite, referring to accompanying drawing, developing simultaneously, the present invention is described in more detail for embodiment.Should be appreciated that specific embodiment described herein, only in order to explain the present invention, is not intended to limit the present invention.
Refer to Fig. 1, Fig. 1 is the process flow diagram of Fingertip Detection the first preferred embodiment based on centre of the palm range finding provided by the invention, comprises the following steps:
Step S100, the original image that comprises gesture gathering is carried out to noise reduction process, and by the image through noise reduction process by unrestrained water completion method judgement connected domain, find out gesture connected region;
The marginal point of step S200, traversal gesture connected region obtains marginal point sequence, the curvature value of every bit in edge calculation point sequence, and the curvature value of every bit and default curvature threshold are compared, find out the marginal point that curvature value is greater than curvature threshold, removing in the marginal point that curvature value is greater than curvature threshold is the point of finger recess again, and the marginal point that remaining curvature value is greater than to curvature threshold is labeled as possible finger tip point;
Step S300, the salient point sequence at each possible finger tip point place is independently divided into new point set, respectively new point set is carried out to cluster, from described point, concentrate and find out finger tip point, and be labeled as preliminary screening finger tip point out;
Step S400, gesture connected region is corroded, extract the barycenter of the palm area after corrosion, and the palm area after corrosion is done to circumscribed circle, obtain external radius of a circle;
Step S500, calculate the first ratio that each preliminary screening finger tip is out put distance and the circumradius of barycenter, and described the first ratio and predetermined first threshold are compared, when the first ratio is less than first threshold, judge that corresponding preliminary screening finger tip point is out as erroneous judgement point filtering; When the first ratio is more than or equal to first threshold, judge that corresponding preliminary screening finger tip point is out as the finger tip point detecting; And output detections result.
Below in conjunction with specific embodiment, above-mentioned steps is described in detail.
In step S100, the original image that first comprises gesture by camera collection, carries out smoothing processing to described original image, to remove unnecessary noise; Carry out again binary conversion treatment and image opening and closing operation, to eliminate cavity and further to eliminate noise, eliminate little noise and little cavity in the image after smoothing processing, as shown in Figure 2 a, Fig. 2 a is after noise reduction process, the image obtaining after binary conversion treatment and image opening and closing operation.
By unrestrained water completion method, the image through noise reduction process is carried out to connected domain judgement again, calculate the area of each connected region.First by unrestrained water completion method, carry out connected domain judgement, calculate the area of different connected regions.Find out the connected region of area maximum again and be left gesture connected region, remove small size connected domain, the connected region of Retention area maximum is gesture connected region.Specifically, different connected regions can be labeled as respectively to 1 ~ N, wherein N is positive integer, then calculate the area of the connected region that each numerical value of 1 ~ N is corresponding, remove small size connected domain, the connected region of Retention area maximum is gesture connected region, as shown in Figure 2 b, and the gesture connected region image of Fig. 2 b for the image in Fig. 2 a is obtained by unrestrained water completion method and connected domain judgement afterwards.
In step S200, after obtaining gesture connected region, by utilizing curvature method to find finger tip point.First by edge detection method, detect the marginal point of gesture connected region, the marginal point of traversal gesture connected region obtains marginal point sequence M.Specifically, first use edge detection method to detect gesture connected region edge, detected gesture connected region edge is as shown in the closed curve in Fig. 3 a, and traversal gesture connected region marginal point obtains gesture marginal point sequence M, the i.e. set of gesture marginal point.
By calculating the size of gesture connected region area, estimate the distance between staff and camera again, according to the distance between staff and camera, a shaping is set and counts k, wherein, k is positive integer, for example, when hand and camera are when being 15cm left and right, setting k value is 50.To the arbitrfary point in gesture marginal point sequence M
ask curvature value: get in gesture marginal point sequence with
2 equidistant points
with
, as shown in Figure 3 a, compute vector
with vector
the cosine value of angle is
the approximate curvature value of point, curvature value formula is
.Specifically,
for any point in gesture marginal point sequence, i is integer,
with
for with i at a distance of be the point of k, for example, k is 50 o'clock, in marginal point sequence and
at a distance of 50 2 points, by included angle cosine formula, calculate the curvature value of this point.
Again the curvature value of every bit in marginal point sequence and default curvature threshold are compared, find out the marginal point that curvature value is greater than curvature threshold, as shown in line thicker on the closed curve in Fig. 3 a, the marginal point 101 and 102 in Fig. 3 a namely, due to the finger tip of finger with refer to that the curvature of contour curve at place, gap is higher, through threshold value relatively after, can screen the marginal point 101 that comprises finger fingertip and the marginal point 102 of finger recess.
Please continue to refer to Fig. 3 a, due to the larger marginal point of the curvature value of finding out, comprise groove of finger fingertip and finger etc.As shown in Figure 3 a, can be by judgement
whether the pixel value of line segment mid point is 0, if this pixel value is 0, is judged as finger recess filtering, has got rid of for the marginal point 102 of finger recess; Then the marginal point 101 that remaining curvature value is greater than to curvature threshold is labeled as possible finger tip point.Specifically, if curvature value is greater than the point of curvature threshold, be finger fingertip, the equidistant line segment mid point of 2 is inevitable in the closed curve shown in Fig. 3 a, namely in gesture connected region; If finger recess, 2 equidistant line segment mid points are not in gesture connected region.Can whether in gesture connected region, judge by Checking line mid point so, because this image has carried out binary conversion treatment, can be by judging that whether the pixel of this line segment mid point is 0 to judge.If this pixel value is 0, be judged as finger recess, these marginal points 102 are got rid of; Image after removal finger recess as shown in Figure 3 b.The marginal point 101 that remaining curvature value is greater than to curvature threshold is labeled as possible finger tip point, and the marginal point 101 in Fig. 3 b, that is to say after above-mentioned processing, and remaining marginal point is possible finger tip point, thereby has reduced False Rate.
In step S300, the salient point sequence at each finger tip point place is independently divided into new point set, these new point sets are carried out to cluster: the some P that gets each some set mean curvature value maximum
max, the method for employing ellipse fitting, the oval external matrix four limit mid points after digital simulation are from P
maxnearest point, this point is preliminary screening finger tip point out, as shown in Figure 3 c, preliminary screening is 5 finger tip points 103 out.These preliminary screening finger tip point 103 is out the possible finger tip point further screening after cluster from marginal point 102.
By above-mentioned steps, finger tip point can be detected preferably, but still have some misjudgment phenomenon, such as when clenching fist, because the reasons such as finger degree of holding is inadequate can exist the problem that some small salient points are finger tip by error detection, the situation such as for example clench fist.As shown in Fig. 4 a, in Fig. 4 a, dotted line represents to point the profile of normal extension, normally should refer to cusp 105, but owing to clenching fist, the finger tip screening is o'clock as shown in 104, this point is preliminary screening finger tip point out while clenching fist, and obviously, can cause erroneous judgement like this.The accuracy of detection, the reduction finger tip false drop rate that in order further to improve real-time finger tip, detect, the present invention also carries out following operation on the basis of above-mentioned steps: in step S400, first gesture connected region is repeatedly corroded, obtain one for the image of palm area, corrosion palm area is later extracted to barycenter O, this barycenter O point is the centre of the palm of palm area, thereby finds out palm centre of the palm point O.In order to reach the high efficiency of system, gesture connected region, by repeatedly corroding, can be removed to the thin pixel regions such as finger, reduce the impact of the thin pixel regions such as finger on palm centroid position.Illustrate, image A obtains C formula after being corroded by image B is
, wherein
it is generally acknowledged corrosion symbol during image is processed.
After gesture connected region is repeatedly corroded, obtain one for the image of palm area, then corrosion pixel region is later extracted to barycenter O, ask the formula of barycenter O to be
,
, wherein, i, j is respectively x axial coordinate and the y axial coordinate of any one pixel in image; M, n are respectively the row and column of image,
for the x axial coordinate of barycenter,
y axial coordinate for barycenter.Palm area is made to circumscribed circle simultaneously, obtain the radius R of palm circumscribed circle, as shown in Fig. 4 a.
In step S500, the distance d of each preliminary screening finger tip point out and centre of the palm O point (being barycenter) in calculation procedure S300, as shown in Fig. 4 a, all possible finger tip points such as finger tip point 104 when being about to preliminary screening finger tip point out and comprising clenching fist as shown in Fig. 4 a and the distance of barycenter O.Calculate again each preliminary screening finger tip out and put the distance of barycenter and the first ratio of circumradius R, and described the first ratio and predetermined first threshold are compared, when the first ratio is less than first threshold, judge that corresponding preliminary screening finger tip point is out as erroneous judgement point filtering; When the first ratio is not less than first threshold, when the first ratio is more than or equal to first threshold, judge that corresponding preliminary screening finger tip point is out as the finger tip point detecting; And output detections result.While normally stretching out due to finger, finger tip point 105(is as shown in dotted portion in Fig. 4 a) to the distance in the centre of the palm and the ratio of palm circumradius, be probably between 1.5 ~ 2, if the finger tip of tentatively judging point is less than 1.5 apart from d with the ratio of circumradius R with the centre of the palm, judge that this finger tip point is as erroneous judgement point, if be greater than this threshold value, think that this finger tip point is non-erroneous judgement point.Detection effect after filtering as shown in Figure 4 b, according to first threshold and the first ratio of obtaining according to barycenter O, the finger tip point eliminating of the erroneous judgement that the finger tip point 104 in the time of can be by clenching fist in Fig. 4 a etc. situation forms, finally obtain further possibility finger tip point 106, thereby improved the accuracy of detection that real-time finger tip detects.When concrete application, the span of described first threshold is 1.5 ~ 2, specifically can set according to actual needs.
Fingertip Detection based on centre of the palm range finding provided by the invention, by at mark preliminary screening finger tip point out after cluster, gesture connected region is corroded, extract the barycenter of the palm area after corrosion, and the palm area after corrosion is done to circumscribed circle, obtain external radius of a circle; Calculate each preliminary screening finger tip out and put the first ratio of distance and the circumradius of barycenter, and described the first ratio and predetermined first threshold are compared, when the first ratio is less than first threshold, judge that corresponding preliminary screening finger tip point is out as erroneous judgement point filtering; When the first ratio is more than or equal to first threshold, judge that corresponding preliminary screening finger tip point is out as the finger tip point detecting; And output detections result, greatly reduced finger tip erroneous judgement situation while clenching fist, improved the accuracy of detection that real-time finger tip detects, reduced finger tip false drop rate, its implementation is simple, by software, realizes, and cost is lower.
The present invention is by above-mentioned the first preferred embodiment, greatly reduced finger tip erroneous judgement situation while clenching fist, but when having finger to stretch out, the size of palm area can be different and have a greater change along with finger extension degree, the corresponding external fenestra of palm have change, likely have half bending of the hand and refer to that by error detection be finger tip point, thereby reduce the precision that finger tip detects, as shown in Figure 6 a, due to the curved finger tip point 107 that screens of causing of finger half, be not real finger tip point actually.Further, the present invention also provides the second preferred embodiment to solve this problem, refer to Fig. 5, Fig. 5 is the process flow diagram of Fingertip Detection the second preferred embodiment based on centre of the palm range finding provided by the invention, different from the first preferred embodiment, the second preferred embodiment also comprises after described step S500:
S600, calculate the distance that each preliminary screening finger tip is out put barycenter, find out the maximum finger tip point of distance, calculating other finger tip except the finger tip point of ultimate range is put the second ratio of distance and the ultimate range of barycenter, described the second ratio and predetermined Second Threshold are compared, when the second ratio is less than Second Threshold, judge that corresponding preliminary screening finger tip point is out as erroneous judgement point filtering; When the second ratio is more than or equal to Second Threshold, judge that corresponding preliminary screening finger tip point is out as the finger tip point detecting; And output detections result.
Specifically, as shown in Figure 6 a, first by the method in step S400, find namely barycenter O of centre of the palm O, then calculate the distance d that all finger tip points that screen and centre of the palm O are ordered
fi, the point of its middle distance maximum is regarded as to finger tip point, such as finger tip point 108 as shown in Figure 6 a, this ultimate range is d
f_max.Calculate again other finger tip except the finger tip point of ultimate range and put distance and the ultimate range d of barycenter
f_maxthe second ratio, for example finger tip point 107 and 109 as shown in Figure 6 a, compares described the second ratio and predetermined Second Threshold, when the second ratio is less than Second Threshold, judges that corresponding preliminary screening finger tip point out puts an also filtering as judging by accident; When the second ratio is not less than Second Threshold, when the second ratio is more than or equal to Second Threshold, judge that corresponding preliminary screening finger tip point is out as the finger tip point detecting; And output detections result.Because the joint in the middle of finger and the distance of finger tip are relatively stable and have certain length, point half finger tip while normally stretching out from the distance in the centre of the palm and finger when curved from the distance in the centre of the palm probably between 0.2 ~ 0.55, if except the distance d of ordering from centre of the palm O apart from other finger tip points the maximum finger tip point in the centre of the palm
fiwith d
f_maxratio be less than 0.6, this point is decided to be erroneous judgement point.Specifically, the span of described Second Threshold is 0.2 ~ 0.6, and numeral can be set as required herein.For example the somatic data of different application environment is different, as used crowd's body data inconsistent in Africa, America and Europe and Asia, cause these ratios also can be different, corresponding threshold value value is also different, these can be selected according to actual situation, such as different ethnic groups, different sexes and age bracket etc.Refer to Fig. 6 b, effect after erroneous judgement point is got rid of as shown in Figure 6 b, only have the finger tip point 109 of thumb and the finger tip point 108 of forefinger to screen, other 1 curved finger tip point is excluded, by erroneous judgement point in Fig. 6 a, be namely that half curved finger tip has been put 107 filterings, greatly improve the accuracy that finger tip detects, reduced False Rate.
Fingertip Detection the second embodiment based on centre of the palm range finding provided by the invention has not only removed the problem of error detection while clenching fist effectively, the problem of error detection while also effectively removing finger half dactylogryposis point further, its Fingertip Detection contains more gesture freedom and flexibility, promote the convenience of user interactions, and real-time.
Based on the above-mentioned Fingertip Detection based on centre of the palm range finding, the present invention also provides a kind of finger tip pick-up unit based on centre of the palm range finding, refer to Fig. 7, Fig. 7 is the structured flowchart of finger tip pick-up unit the first preferred embodiment based on centre of the palm range finding provided by the invention, as shown in Figure 7, described device comprises:
Gesture connected region generation module 10, for the original image that comprises gesture gathering is carried out to noise reduction process, and judges connected domain by the image through noise reduction process by unrestrained water completion method, finds out gesture connected region; Specifically as described in step S100.
Curvature value computing module 20, obtains marginal point sequence for traveling through the marginal point of gesture connected region, the curvature value of every bit in edge calculation point sequence; Specifically as described in step S200.
Curvature value contrast module 30, for comparing the curvature value of marginal point sequence every bit and default curvature threshold; Specifically as described in step S200.
Groove is removed module 40, for removing curvature value, is greater than the marginal point of curvature threshold for the point of finger recess; Specifically as described in step S200.
Barycenter and circumradius acquisition module 60, for gesture connected region is corroded, extract the barycenter of the palm area after corrosion, and the palm area after corrosion done to circumscribed circle, obtains external radius of a circle; Specifically as described in step S400.
The first ratio calculation module 70, puts the first ratio of distance and the circumradius of barycenter for calculating each preliminary screening finger tip out; Specifically as described in step S500.
The first ratio contrast module 80, for comparing described the first ratio and predetermined first threshold; Specifically as described in step S500.
Specifically, 10 pairs of original images that comprise gesture of described gesture connected region generation module first carry out smoothing processing, to remove unnecessary noise; Carry out again binary conversion treatment and image opening and closing operation, to eliminate cavity and further to eliminate noise, by unrestrained water completion method, carry out connected domain judgement again, calculate the area of each connected region, find out the connected region of area maximum and be left gesture connected region, remove small size connected domain, the connected region of Retention area maximum is gesture connected region.The marginal point of the gesture connected region of then, finding out in the described gesture connected region generation module 10 of described curvature value computing module 20 traversal obtains marginal point sequence, the curvature value of every bit in edge calculation point sequence.Described curvature value contrast module 30 compares the curvature value of the every bit calculating in curvature value computing module 20 and default curvature threshold, finds out the marginal point that curvature value is greater than curvature threshold.Through groove, remove module 40 again and remove in the marginal point that curvature values are greater than curvature threshold after the point for finger recess, the marginal point that remaining curvature value is greater than to curvature threshold is labeled as possible finger tip point.Then by cluster, screen module 50 the salient point sequence at each possible finger tip point place is independently divided into new point set, respectively new point set is carried out to cluster, from described point, concentrate and find out finger tip point, and be labeled as preliminary screening finger tip point out.By barycenter and 60 pairs of gesture connected regions of circumradius acquisition module, corrode again, extract the barycenter of the palm area after corrosion, and the palm area after corrosion is done to circumscribed circle, obtain external radius of a circle.Finally by the first ratio calculation module 70, calculate the first ratio that each preliminary screening finger tip is out put distance and the circumradius of barycenter, through the first ratio contrast module 80, described the first ratio and first threshold are compared, when the first ratio is less than first threshold, this preliminary screening finger tip point is out erroneous judgement point, judges that corresponding preliminary screening finger tip point is out as erroneous judgement point filtering; When the first ratio is more than or equal to first threshold, the finger tip point of this preliminary screening finger tip point out for detecting, judges that corresponding preliminary screening finger tip point is out as the finger tip point detecting; And output detections result, greatly reduced finger tip erroneous judgement situation while clenching fist, improved the accuracy of detection that real-time finger tip detects, reduced finger tip false drop rate, its implementation is simple, by software, realizes, and cost is lower.
Specifically, described curvature computing module 20 comprises:
Marginal point sequence generating unit, for detect the marginal point of gesture connected region by edge detection method, the marginal point of traversal gesture connected region obtains marginal point sequence M;
Curvature value computing unit, for the size by gesture connected region, judge the distance between staff and camera, according to the distance between staff and camera, a shaping is set and counts k, wherein, k is positive integer, then calculates any point in gesture marginal point sequence M
curvature value: get in gesture marginal point sequence with
2 equidistant points
with
, compute vector
with vector
the cosine value of angle is
the approximate curvature value of point, curvature value formula is
.
When practical application, described groove is removed mould 40 by judgement
whether the pixel value of line segment mid point is 0; If this pixel value is 0, be judged as finger recess filtering.
Specifically, described barycenter and circumradius acquisition module 60 comprise:
Erosion unit, for gesture connected region is repeatedly corroded, obtains one for the image of palm area;
Centroid calculation unit, for corrosion palm area is later extracted to barycenter O, this barycenter O point is the centre of the palm of palm area, and the formula of barycenter O is
,
, wherein, i, j is respectively x axial coordinate and the y axial coordinate of any one pixel in image; M, n are respectively the row and column of image,
for the x axial coordinate of barycenter,
y axial coordinate for barycenter;
Circumradius computing unit, for to corrosion palm area is later made to circumscribed circle, obtains the radius R of palm circumscribed circle.
Refer to Fig. 8, the structured flowchart of finger tip pick-up unit the second preferred embodiment based on centre of the palm range finding provided by the invention, the problem of error detection while pointing half dactylogryposis point in order effectively to remove further.Described device also comprises on the basis of the first preferred embodiment:
The second ratio calculation module 90, puts the distance of barycenter for calculating each preliminary screening finger tip out, find out the maximum finger tip point of distance, calculates the second ratio that other finger tip except the finger tip point of ultimate range is put distance and the ultimate range of barycenter; Specifically as described in step S600.
The second ratio contrast module 100, for described the second ratio and predetermined Second Threshold are compared, specifically as described in step S600.
Specifically, described the second ratio calculation module 90 is calculated the distance that each preliminary screening finger tip is out put barycenter, find out the maximum finger tip point of distance, calculating other finger tip except the maximum finger tip point of distance is put the distance of barycenter and apart from maximum finger tip, is put the second ratio of the distance of barycenter, by described the second ratio contrast module 100, the second ratio and Second Threshold are contrasted again, when the second ratio is less than Second Threshold, this finger tip point is erroneous judgement point, judges that corresponding preliminary screening finger tip point is out as erroneous judgement point filtering; When the second ratio is more than or equal to Second Threshold, the finger tip point of this finger tip point for detecting, the finger tip screening point, namely judges that corresponding preliminary screening finger tip point is out as the finger tip point detecting; And output detections result.Like this, finger tip pick-up unit the second preferred embodiment based on centre of the palm range finding provided by the invention, further effectively got rid of non-finger tip point, not only greatly reduced finger tip erroneous judgement situation while clenching fist, further effectively removed the problem of pointing half dactylogryposis point error detection, contain more gesture freedom and flexibility, promote the convenience of user interactions, and real-time.
In sum, a kind of Fingertip Detection and device based on centre of the palm range finding provided by the invention, described method is by mark preliminary screening finger tip point out after cluster, gesture connected region is corroded, extract the barycenter of the palm area after corrosion, and the palm area after corrosion is done to circumscribed circle, obtain external radius of a circle; Calculate each preliminary screening finger tip out and put the first ratio of distance and the circumradius of barycenter, and described the first ratio and predetermined first threshold are compared, when the first ratio is less than first threshold, judge that corresponding preliminary screening finger tip point is out as erroneous judgement point filtering; When the first ratio is more than or equal to first threshold, judge that corresponding preliminary screening finger tip point is out as the finger tip point detecting; And output detections result, thereby improved the accuracy of detection that real-time finger tip detects, reduced finger tip false drop rate, its implementation is simple, by software, realizes, and cost is lower, has effectively removed the problem of error detection while clenching fist.
Further, by calculating each preliminary screening finger tip out, put the distance of barycenter, find out the maximum finger tip point of distance, calculating other finger tip except the finger tip point of ultimate range is put the second ratio of distance and the ultimate range of barycenter, described the second ratio and predetermined Second Threshold are compared, when the second ratio is less than Second Threshold, judge that corresponding preliminary screening finger tip point is out as erroneous judgement point filtering; When the second ratio is more than or equal to Second Threshold, judge that corresponding preliminary screening finger tip point is out as the finger tip point detecting; And output detections result, not only effectively removed the problem of error detection while clenching fist, the problem of error detection while also effectively removing finger half dactylogryposis point further, its Fingertip Detection contains more gesture freedom and flexibility, promote the convenience of user interactions, and real-time.
Should be understood that, application of the present invention is not limited to above-mentioned giving an example, and for those of ordinary skills, can be improved according to the above description or convert, and all these improvement and conversion all should belong to the protection domain of claims of the present invention.
Claims (10)
1. the Fingertip Detection based on centre of the palm range finding, is characterized in that, comprises the following steps:
A, the original image that comprises gesture gathering is carried out to noise reduction process, and by the image through noise reduction process by unrestrained water completion method judgement connected domain, find out gesture connected region;
The marginal point of B, traversal gesture connected region obtains marginal point sequence, the curvature value of every bit in edge calculation point sequence, and the curvature value of every bit and default curvature threshold are compared, find out the marginal point that curvature value is greater than curvature threshold, removing in the marginal point that curvature value is greater than curvature threshold is the point of finger recess again, and the marginal point that remaining curvature value is greater than to curvature threshold is labeled as possible finger tip point;
C, the salient point sequence at each possible finger tip point place is independently divided into new point set, respectively new point set is carried out to cluster, from described point, concentrate and find out finger tip point, and be labeled as preliminary screening finger tip point out;
D, gesture connected region is corroded, extract the barycenter of the palm area after corrosion, and the palm area after corrosion is done to circumscribed circle, obtain external radius of a circle;
E, calculate the first ratio that each preliminary screening finger tip is out put distance and the circumradius of barycenter, and described the first ratio and predetermined first threshold are compared, when the first ratio is less than first threshold, judge that corresponding preliminary screening finger tip point is out as erroneous judgement point filtering; When the first ratio is more than or equal to first threshold, judge that corresponding preliminary screening finger tip point is out as the finger tip point detecting; And output detections result.
2. the Fingertip Detection based on centre of the palm range finding according to claim 1, is characterized in that, after described step e, also comprises:
F, calculate the distance that each preliminary screening finger tip is out put barycenter, find out the maximum finger tip point of distance, calculating other finger tip except the finger tip point of ultimate range is put the second ratio of distance and the ultimate range of barycenter, described the second ratio and predetermined Second Threshold are compared, when the second ratio is less than Second Threshold, judge that corresponding preliminary screening finger tip point is out as erroneous judgement point filtering; When the second ratio is more than or equal to Second Threshold, judge that corresponding preliminary screening finger tip point is out as the finger tip point detecting; And output detections result.
3. the Fingertip Detection based on centre of the palm range finding according to claim 1, is characterized in that, described steps A specifically comprises:
A1, the original image that comprises gesture by camera collection, carry out smoothing processing to described original image, to remove unnecessary noise; Carry out again binary conversion treatment and image opening and closing operation, to eliminate cavity and further to eliminate noise;
A2, by unrestrained water completion method, the image through noise reduction process is carried out to connected domain judgement, calculate the area of each connected region;
A3, find out the connected region of area maximum and be left gesture connected region, remove small size connected domain.
4. the Fingertip Detection based on centre of the palm range finding according to claim 1, is characterized in that, described step B specifically comprises:
B1, by edge detection method, detect the marginal point of gesture connected region, the marginal point of traversal gesture connected region obtains marginal point sequence M;
B2, the size by gesture connected region judge the distance between staff and camera, according to the distance between staff and camera, a shaping are set and count k, and wherein, k is positive integer, then calculates any point in gesture marginal point sequence M
curvature value: get in gesture marginal point sequence with
2 equidistant points
with
, compute vector
with vector
the cosine value of angle is
the approximate curvature value of point, curvature value formula is
;
B3, the curvature value of every bit in marginal point sequence and default curvature threshold are compared, find out the marginal point that curvature value is greater than curvature threshold;
B4, judgement
whether the pixel value of line segment mid point is 0; If this pixel value is 0, be judged as finger recess filtering; The marginal point that remaining curvature value is greater than to curvature threshold is labeled as possible finger tip point.
5. the Fingertip Detection based on centre of the palm range finding according to claim 1, is characterized in that, described step C specifically comprises:
C1, the salient point sequence at each possible finger tip point place is independently divided into new point set, respectively new point set is carried out to cluster;
C2, get the some P of each some set mean curvature value maximum
max, the method for employing ellipse fitting, the oval external matrix four limit mid points after digital simulation are from P
maxnearest point, is labeled as preliminary screening finger tip point out by this point.
6. the Fingertip Detection based on centre of the palm range finding according to claim 1, is characterized in that, described step D specifically comprises:
D1, gesture connected region is repeatedly corroded, obtain one for the image of palm area, corrosion palm area is later extracted to barycenter O, this barycenter O point is the centre of the palm of palm area, and the formula of barycenter O is
,
, wherein, i, j is respectively x axial coordinate and the y axial coordinate of any one pixel in image; M, n are respectively the row and column of image,
for the x axial coordinate of barycenter,
y axial coordinate for barycenter;
D2, palm area is made to circumscribed circle, obtain the radius R of palm circumscribed circle.
7. the finger tip pick-up unit based on centre of the palm range finding, is characterized in that, comprising:
Gesture connected region generation module, for the original image that comprises gesture gathering is carried out to noise reduction process, and judges connected domain by the image through noise reduction process by unrestrained water completion method, finds out gesture connected region;
Curvature value computing module, obtains marginal point sequence for traveling through the marginal point of gesture connected region, the curvature value of every bit in edge calculation point sequence;
Curvature value contrast module, for comparing the curvature value of marginal point sequence every bit and default curvature threshold;
Groove is removed module, for removing curvature value, is greater than the marginal point of curvature threshold for the point of finger recess;
Cluster screening module, for the salient point sequence at each possible finger tip point place is independently divided into new point set, carries out cluster to new point set respectively, concentrates and finds out finger tip point, and be labeled as preliminary screening finger tip point out from described point;
Barycenter and circumradius acquisition module, for gesture connected region is corroded, extract the barycenter of the palm area after corrosion, and the palm area after corrosion done to circumscribed circle, obtains external radius of a circle;
The first ratio calculation module, puts the first ratio of distance and the circumradius of barycenter for calculating each preliminary screening finger tip out;
The first ratio contrast module, for comparing described the first ratio and predetermined first threshold.
8. the finger tip pick-up unit based on centre of the palm range finding according to claim 7, is characterized in that, also comprises:
The second ratio calculation module, puts the distance of barycenter for calculating each preliminary screening finger tip out, find out the maximum finger tip point of distance, calculates the second ratio that other finger tip except the finger tip point of ultimate range is put distance and the ultimate range of barycenter;
The second ratio contrast module, for comparing described the second ratio and predetermined Second Threshold.
9. the finger tip pick-up unit based on centre of the palm range finding according to claim 7, is characterized in that, described curvature value computing module comprises:
Marginal point sequence generating unit, for detect the marginal point of gesture connected region by edge detection method, the marginal point of traversal gesture connected region obtains marginal point sequence M;
Curvature value computing unit, for the size by gesture connected region, judge the distance between staff and camera, according to the distance between staff and camera, a shaping is set and counts k, wherein, k is positive integer, then calculates any point in gesture marginal point sequence M
curvature value: get in gesture marginal point sequence with
2 equidistant points
with
, compute vector
with vector
the cosine value of angle is
the approximate curvature value of point, curvature value formula is
.
10. the finger tip pick-up unit based on centre of the palm range finding according to claim 7, is characterized in that, described barycenter and circumradius acquisition module comprise:
Erosion unit, for gesture connected region is repeatedly corroded, obtains one for the image of palm area;
Centroid calculation unit, for corrosion palm area is later extracted to barycenter O, this barycenter O point is the centre of the palm of palm area, and the formula of barycenter O is
,
, wherein, i, j is respectively x axial coordinate and the y axial coordinate of any one pixel in image; M, n are respectively the row and column of image,
for the x axial coordinate of barycenter,
y axial coordinate for barycenter;
Circumradius computing unit, for to corrosion palm area is later made to circumscribed circle, obtains the radius R of palm circumscribed circle.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310312446.3A CN103544469B (en) | 2013-07-24 | 2013-07-24 | Fingertip detection method and device based on palm ranging |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310312446.3A CN103544469B (en) | 2013-07-24 | 2013-07-24 | Fingertip detection method and device based on palm ranging |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103544469A true CN103544469A (en) | 2014-01-29 |
| CN103544469B CN103544469B (en) | 2017-05-10 |
Family
ID=49967904
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310312446.3A Expired - Fee Related CN103544469B (en) | 2013-07-24 | 2013-07-24 | Fingertip detection method and device based on palm ranging |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103544469B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104908048A (en) * | 2014-03-14 | 2015-09-16 | 本田技研工业株式会社 | Motion estimation device, robot, and motion estimation method |
| CN109344689A (en) * | 2018-08-07 | 2019-02-15 | 西安理工大学 | A kind of sign language gesture identification method based on Kinect |
| CN109754379A (en) * | 2018-12-29 | 2019-05-14 | 北京金山安全软件有限公司 | Image processing method and device |
| CN111046796A (en) * | 2019-12-12 | 2020-04-21 | 哈尔滨拓博科技有限公司 | Low-cost space gesture control method and system based on double-camera depth information |
| CN111523387A (en) * | 2020-03-24 | 2020-08-11 | 杭州易现先进科技有限公司 | Method and device for detecting hand key points and computer device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101763500A (en) * | 2008-12-24 | 2010-06-30 | 中国科学院半导体研究所 | Method applied to palm shape extraction and feature positioning in high-freedom degree palm image |
| CN102063618A (en) * | 2011-01-13 | 2011-05-18 | 中科芯集成电路股份有限公司 | Dynamic gesture identification method in interactive system |
| CN102622601A (en) * | 2012-03-12 | 2012-08-01 | 李博男 | Fingertip detection method |
| US20130057469A1 (en) * | 2010-05-11 | 2013-03-07 | Nippon Systemware Co Ltd | Gesture recognition device, method, program, and computer-readable medium upon which program is stored |
-
2013
- 2013-07-24 CN CN201310312446.3A patent/CN103544469B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101763500A (en) * | 2008-12-24 | 2010-06-30 | 中国科学院半导体研究所 | Method applied to palm shape extraction and feature positioning in high-freedom degree palm image |
| US20130057469A1 (en) * | 2010-05-11 | 2013-03-07 | Nippon Systemware Co Ltd | Gesture recognition device, method, program, and computer-readable medium upon which program is stored |
| CN102063618A (en) * | 2011-01-13 | 2011-05-18 | 中科芯集成电路股份有限公司 | Dynamic gesture identification method in interactive system |
| CN102622601A (en) * | 2012-03-12 | 2012-08-01 | 李博男 | Fingertip detection method |
Non-Patent Citations (2)
| Title |
|---|
| DUNG DUC NGUYEN 等: ""Fingertip Detection with Morphology and Geometric Calculation"", 《THE 2009 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS》 * |
| 冯志权等: ""基于特征点分布分析的手势特征检测方法"", 《计算机集成制造系统》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104908048A (en) * | 2014-03-14 | 2015-09-16 | 本田技研工业株式会社 | Motion estimation device, robot, and motion estimation method |
| US9684824B2 (en) | 2014-03-14 | 2017-06-20 | Honda Motor Co., Ltd | Motion estimation device, robot, and method to estimate reaction of a person |
| CN109344689A (en) * | 2018-08-07 | 2019-02-15 | 西安理工大学 | A kind of sign language gesture identification method based on Kinect |
| CN109754379A (en) * | 2018-12-29 | 2019-05-14 | 北京金山安全软件有限公司 | Image processing method and device |
| CN111046796A (en) * | 2019-12-12 | 2020-04-21 | 哈尔滨拓博科技有限公司 | Low-cost space gesture control method and system based on double-camera depth information |
| CN111523387A (en) * | 2020-03-24 | 2020-08-11 | 杭州易现先进科技有限公司 | Method and device for detecting hand key points and computer device |
| CN111523387B (en) * | 2020-03-24 | 2024-04-19 | 杭州易现先进科技有限公司 | Method and device for detecting key points of hands and computer device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103544469B (en) | 2017-05-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104899600B (en) | A kind of hand-characteristic point detecting method based on depth map | |
| CN103984928B (en) | Finger gesture recognition methods based on depth image | |
| KR102118408B1 (en) | Method of performing a touch action in a touch sensitive device | |
| CN108960163B (en) | Gesture recognition method, device, equipment and storage medium | |
| EP3537375B1 (en) | Image segmentation methods, image segmentation system and device comprising same, and storage medium | |
| CN103544469A (en) | Fingertip detection method and device based on palm ranging | |
| CN109829368B (en) | Palm feature recognition method and device, computer equipment and storage medium | |
| CN102368290A (en) | Hand gesture identification method based on finger advanced characteristic | |
| CN110420025A (en) | Surface electromyogram signal processing method, device and wearable device | |
| CN110837792B (en) | Three-dimensional gesture recognition method and device | |
| CN111652017B (en) | Dynamic gesture recognition method and system | |
| CN102236412A (en) | Three-dimensional gesture recognition system and vision-based gesture recognition method | |
| TWI475422B (en) | Method for recognizing gesture and electronic device | |
| CN103544472A (en) | Processing method and processing device based on gesture images | |
| KR101559502B1 (en) | Method and recording medium for contactless input interface with real-time hand pose recognition | |
| CN103870071B (en) | One kind touches source discrimination and system | |
| CN107357414B (en) | Click action recognition method and device | |
| CN111754391A (en) | Face correcting method, face correcting equipment and computer readable storage medium | |
| Bei et al. | Sitting posture detection using adaptively fused 3D features | |
| CN112115853A (en) | Gesture recognition method and device, computer storage medium and electronic equipment | |
| CN103186241B (en) | A kind of interactive desktop contact right-hand man's recognition methods | |
| CN111160173B (en) | Gesture recognition method based on robot and robot | |
| CN106971132A (en) | One kind scanning gesture simultaneously knows method for distinguishing | |
| JP4991278B2 (en) | Image processing device | |
| CN113936305A (en) | Middle finger position confirmation method and feature extraction method in palm recognition process |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170510 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |