CN107729883A - A kind of vein image area-of-interest exacting method - Google Patents
A kind of vein image area-of-interest exacting method Download PDFInfo
- Publication number
- CN107729883A CN107729883A CN201711175045.2A CN201711175045A CN107729883A CN 107729883 A CN107729883 A CN 107729883A CN 201711175045 A CN201711175045 A CN 201711175045A CN 107729883 A CN107729883 A CN 107729883A
- Authority
- CN
- China
- Prior art keywords
- hand
- interest
- area
- vein image
- center
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/66—Analysis of geometric attributes of image moments or centre of gravity
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/14—Vascular patterns
Abstract
The present invention discloses a kind of vein image area-of-interest exacting method, belongs to living things feature recognition field.The step of the present invention is using infrared camera collection vein image；Extraction hand profile simultaneously finds the maximum inscribed circle in profile (center of circle and radius), and vein image region corresponding to maximum inscribed circle is the area-of-interest of extraction；Summit between the maximum inscribed circle center of circle and wrist central point is found out according to point to straight line (line of the center of circle and wrist central point) distance, fits hand center line；The slope of center line is calculated, draws center line and vertical direction angle, according to angle, vein image is rotated, hand center line is in vertical direction；Now the vein image in maximum inscribed circle is exactly the region of interest area image for extracting and correcting.The present invention is for can not detect finger-joint point, or the curvature of palm edge area is excessive, is fitted inappropriate situation using straight line, can effectively extract area-of-interest.
Description
Technical field
The invention belongs to the region of interesting extraction part in biometrics identification technology field, a kind of hand of specific design
Portion's vein area-of-interest exacting method.
Background technology
During identity authentication is carried out using vein, it is necessary first to extract the area-of-interest of image, image does not wrap
The region of vein is included, reducing needs data volume to be processed.For the vein image gathered at different moments, it is emerging that the phase same feeling can be extracted
Interesting region can influence the performance of whole identifying system, and therefore, the extracting method of area-of-interest is extremely important.For offsetting,
Same people's difference of the changes such as rotation puts the vein image of posture, and traditional region of interesting extraction algorithm extracts interested
Otherness is obvious between area results, adds the complexity of follow-up recognizer.Specific the problem of existing, is as follows：
(1) because the fat or thin degree of people is different, in the case where clenching fist, the finger-joint point position of groups of people it is not bright
Aobvious concavo-convex change, it is difficult to detect, cannot use in such cases method based on curvature or relative wrist central point away from
From method carry out region of interesting extraction, while the method for relative wrist central point distance depends on the putting position of wrist；
(2) in the case where clenching fist, the palm sideline curvature of little finger side is bigger, uses the extracting method based on sideline
During the region of extraction, the straight line of fitting is difficult to describe palm sideline, and the sideline change ratio of the image fitting gathered at different moments
It is larger, cause the regional location deviation of extraction big；When extracting area-of-interest, it is necessary to distinguish right-hand man.
(3) when carrying out region-of-interest direction correction, generally use is point with extreme curvature, the extreme point of relative distance
The either sideline of little finger side, it is also bigger with two, front reason, the direction difference of the vein area-of-interest after correction.
The content of the invention
In view of the problem of to exist during upper vein region of interesting extraction, the present invention proposes that a kind of adaptability is stronger
Hand vein area-of-interest exacting method, specifically includes following steps：
A kind of vein image area-of-interest exacting method, it is characterised in that comprise the following steps：
(1) vein image binaryzation calculates with center-of-mass coordinate
Binary conversion treatment is carried out to vein image first, hand region is extracted from background, calculates the matter of hand region
The heart, obtain center-of-mass coordinate；
(2) hand contours extract
Hand contours extract is carried out to binary image, obtains the profile of whole hand region；
(3) the hand contoured interior maximum inscribed circle extraction based on Wei Nuotu
The Wei Nuotu of hand profile is generated, obtains Wei Nuotu summits, calculates distance of the summit with corresponding hand profile point,
Wherein, Wei Nuotu summits to the distance of corresponding hand profile point be the summit into the distance of all profile points recently.Time
All summits gone through in hand center-of-mass coordinate peripheral region, find and pushed up with corresponding hand profile point apart from farthest Wei Nuotu
Point, the summit are the center of circle of hand contoured interior maximum inscribed circle, and the distance of summit and profile point is the half of maximum inscribed circle
Footpath, vein image region corresponding to maximum inscribed circle are exactly the area-of-interest extracted；
(4) hand midline extraction
Connection wrist central point and incenter obtain straight line, obtain an approximate hand center line, find out position
In the summit that hand approximate centerline and position are among two points, by calculating the part summit to the distance of the straight line,
Wherein distance is taken to be less than the summit of some threshold value, these point fittings are in alignment, as hand center line；
(5) area-of-interest rotational correction
By the slope of hand center line, the angle of hand center line and vertical direction is calculated, rotating image makes hand center line position
In vertical direction, the as area-of-interest after rotational correction；
(6) size normalizes
To the region of interest area image extracted, size normalization is carried out, for different vein images, obtains size one
The area-of-interest of sample.
During the region of interesting extraction, it is not necessary to right-hand man is distinguished, independent of other portions outside wrist
Position, wrist are used to determine hand center line.
After what the present invention can solve the problem that traditional ROI extraction algorithms occur in extraction process effectively can not extract or extract
The problem of deviation that makes correction for direction is bigger.When carrying out ROI extractions, adaptability of the invention is stronger.Can effectively it extract
ROI, and travel direction is corrected exactly, and not distinguish right-hand man.For finger-joint point, or palm side can not be detected
The curvature in line region is excessive, is fitted inappropriate situation using straight line, can effectively extract area-of-interest.
Brief description of the drawings
The embodiment of the present invention is described in further detail below in conjunction with the accompanying drawings；
Fig. 1 is vein image region of interesting extraction flow chart；
Fig. 2 is hand vein image；
Fig. 3 is binaryzation vein image；
Fig. 4 is hand outline drawing；
Fig. 5 is hand profile Nei Weinuotu vertex graphs；
Fig. 6 is the maximum inscribed circle in hand profile；
Fig. 7 is vein image maximum inscribed circle schematic diagram；
Fig. 8 is hand Graph；
Fig. 9 is rounded interested area；
Figure 10 is the rounded interested area after rotational correction；
Figure 11 is vein image and the area-of-interest of extraction under different height；
Figure 12 is the vein image that can't detect artis；
Figure 13 is the area-of-interest extracted from Figure 12；
Figure 14 is hand profile corresponding to Figure 12 and hand center line；
Figure 15 is the bigger vein image of little finger side line curvature；
Figure 16 is the area-of-interest extracted from Figure 15；
Figure 17 is hand profile corresponding to Figure 15 and hand center line.
Fig. 5 is hand profile Nei Weinuotu vertex graphs, and the scatterplot in figure is Wei Nuotu summits, and asterisk is hand center of mass point,
Square is the barycenter peripheral region chosen, and five-pointed star is the center of circle of maximum inscribed circle；Fig. 6 is the maximum of extraction in hand profile
Inscribed circle, the five-pointed star in centre position is the center of circle of inscribed circle, and middle circle is the center of mass point of hand region；Fig. 7 is extraction
The area-of-interest schematic diagram that maximum inscribed circle is corresponded on vein image；Scatterplot in Fig. 8 is the Wei Nuo near hand center line
Figure summit, middle straight line are wrist central point and the line of incenter, dotted line be using summit fit come hand
Center line；Fig. 9 is the rounded interested area of extraction, and Figure 10 is that Fig. 9 is revolved according to center line in Fig. 8 and vertical direction angle angle
Turn the area-of-interest after correction.
Embodiment
1. carrying out binary conversion treatment using maximum variance between clusters hand vein image, binary image is obtained, by hand
Portion region separates with background image, as a result as shown in figure 3, using centroid method, asks for the barycenter (x of hand regioncentroid,
ycentroid)；
2. extract hand contour line from bianry image using hand contours extract algorithm, setting hand profile and base
Two intersecting point coordinates are respectively (x1, 0) and (the base point in left side in Fig. 4) and (xn, 0) and the base point of right side (in Fig. 4), n is in formula
The total length (pixel total number) of hand profile, according to formula xmid=(x1+xn)/2, calculate the center point coordinate of two intersection points
(xmid, 0), i.e. wrist center point coordinate.Hand profile is traveled through, finds out and is located at profile leftmost side point coordinates in hand profile
(xleft,yleft), rightmost side point coordinates (xright,yright) and uppermost point coordinate (xup,yup), hand region is located at (xleft,
0), (xright, 0), (xleft,yup), (xright,yup) in the rectangular extent that is surrounded of four summits；
3. draw the Wei Nuotu of hand contour line, wherein Wei Nuotu summits (x using MATLAB voronoi functionsv,yv)
To corresponding hand profile point (xvi,yvi) distance be less than summit to other profile points distance, wherein (xv,yv) it is Wei Nuotu positions
In any one summit of hand contoured interior.Summit and correspondence profile point (xvi,yvi) distance beTravel through center of mass point (xcentroid,ycentroid) (asterisk in Fig. 5 is matter for peripheral region
Heart position, wherein square be choose 60*60 pixels peripheral region, chosen according to actual conditions) in summit, find distance
Point (x maximum Dc,yc), with (xc,yc) it is the center of circle (five-pointed star in Fig. 5), the circle using distance D as radius, it is exactly hand profile
Interior maximum inscribed circle, and corresponding vein image part is exactly the rounded interested area extracted (such as circular institute in Fig. 7
Show), the five-pointed star of hand center is the maximum inscribed circle center of circle in Fig. 6, and neighbouring roundlet is the barycenter of hand region, profile
On five-pointed star be maximum inscribed circle and the tangent point of contact of profile；
4. connect wrist central point (xmid, 0) and incenter (xc,yc) (five-pointed star in Fig. 8), this was calculated
2 points of straight line formula L:Ax+By+C=0, whereinB=-1,Traversal positioned at the center of circle and
The Wei Nuotu summits in wrist central point centre position, pass through formulaCalculate summit (xi,yi), i=
1,2 ..., n (n is the Wei Nuotu summits total number in hand profile positioned at the center of circle and wrist central point centre position) arrives straight line L
Distance, find out it is all to straight line L distance be less than some threshold value (threshold value 10, chosen according to actual conditions) summits, utilize
These points go out hand center line L by least square fittingmid:Amidx+Bmidy+CmidScatterplot in=0, Fig. 8 is to straight line L
Distance is less than the summit of threshold value, and figure cathetus is the hand center line fitted；
5. according to the slope of hand center line, using θ=arctan (k),Calculate straight line and vertical direction
Angle (90 ° of-θ), if angle be less than 0 °, rounded interested area turns clockwise (θ -90 °), if angle be more than etc.
In 0 °, then rotate counterclockwise (90 ° of-θ), makes hand center line be in vertical direction；
6. the rounded interested area of pair extraction carries out dimension normalization, pass through circular radius R and scaling scale coefficientProduct normalized after radius be RdRounded interested area, wherein RdBy being manually set.
Experimentation：
1. in Figure 11 (a-c) to be the same hand put posture, apart from the case that the distance of camera is different in difference, adopt
The hand vein image of collection, wherein figure a is farthest apart from camera, figure c is nearest apart from camera；Figure (d-f) is from figure (a-c)
The area-of-interest extracted respectively；It can be seen that for different distance and the hand images for putting posture, this method can
Effectively to extract the area-of-interest of vein image；
2. Figure 12 is the vein image that can't detect artis, Figure 14 is corresponding hand profile and hand center line, from wheel
As can be seen that a finger summit can only well be detected in exterior feature, therefore image can not be extracted using the method for relative distance
Area-of-interest, Figure 13 be this method extraction area-of-interest；
3. Figure 15 is the bigger vein image of little finger side line curvature, Figure 17 is in corresponding hand profile and hand
Line, it can be seen that side curvature of a curve is very big from profile, it is difficult to be fitted a good straight line and carried as area-of-interest
The datum line taken；Figure 16 is the area-of-interest of this method extraction.
Claims (2)
1. a kind of vein image area-of-interest exacting method, it is characterised in that comprise the following steps：
(1) vein image binaryzation calculates with center-of-mass coordinate
Binary conversion treatment is carried out to vein image first, hand region is extracted from background, the barycenter of hand region is calculated, obtains
To center-of-mass coordinate；
(2) hand contours extract
Hand contours extract is carried out to binary image, obtains the profile of whole hand region；
(3) the hand contoured interior maximum inscribed circle extraction based on Wei Nuotu
The Wei Nuotu of hand profile is generated, obtains Wei Nuotu summits, calculates distance of the summit with corresponding hand profile point, its
In, Wei Nuotu summits to the distance of corresponding hand profile point are that the summit is nearest into the distance of all profile points.Traversal
All summits in hand center-of-mass coordinate peripheral region, find with corresponding hand profile point apart from farthest Wei Nuotu summits,
The summit is the center of circle of hand contoured interior maximum inscribed circle, and the distance of summit and profile point is the radius of maximum inscribed circle,
Vein image region corresponding to maximum inscribed circle is exactly the area-of-interest extracted；
(4) hand midline extraction；
Connection wrist central point and incenter obtain straight line, obtain an approximate hand center line, find out positioned at hand
The summit that portion's approximate centerline and position are among two points, by calculating the part summit to the distance of the straight line, takes it
Middle distance is less than the summit of some threshold value, and these point fittings are in alignment, as hand center line；
(5) area-of-interest rotational correction；
By the slope of hand center line, the angle of hand center line and vertical direction is calculated, it is perpendicular that rotation image is located at hand center line
Nogata is to the as area-of-interest after rotational correction；
(6) size normalizes
To the region of interest area image extracted, size normalization is carried out, for different vein images, it is the same to obtain size
Area-of-interest.
A kind of 2. vein image area-of-interest exacting method, it is characterised in that：
During the region of interesting extraction, it is not necessary to right-hand man is distinguished, independent of other positions outside wrist, hand
Wrist is used to determine hand center line.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711175045.2A CN107729883A (en) | 2017-11-22 | 2017-11-22 | A kind of vein image area-of-interest exacting method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711175045.2A CN107729883A (en) | 2017-11-22 | 2017-11-22 | A kind of vein image area-of-interest exacting method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107729883A true CN107729883A (en) | 2018-02-23 |
Family
ID=61218056
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711175045.2A Pending CN107729883A (en) | 2017-11-22 | 2017-11-22 | A kind of vein image area-of-interest exacting method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107729883A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109190517A (en) * | 2018-08-14 | 2019-01-11 | 北京凌云天润智能科技有限公司 | A kind of finger venous image edge extracting and direction antidote |
| CN109345735A (en) * | 2018-10-12 | 2019-02-15 | 南京理工大学 | A kind of self-service machine commodity recognition method and system |
| CN109727289A (en) * | 2019-01-18 | 2019-05-07 | 珠海市万瑙特健康科技有限公司 | Location determining method, device and the computer equipment of pulse condition perception point |
| CN109784083A (en) * | 2019-02-22 | 2019-05-21 | 吉林大学 | The bionical encryption system merged based on grip information with hand back vein information |
| CN110032936A (en) * | 2019-03-08 | 2019-07-19 | 吉林大学 | The maximum round method for extracting region of printenv hand back vein |
| CN110909631A (en) * | 2019-11-07 | 2020-03-24 | 黑龙江大学 | Finger vein image ROI extraction and enhancement method |
| WO2020135230A1 (en) * | 2018-12-29 | 2020-07-02 | 北京金山安全软件有限公司 | Inscribed circle determination method and device |
| CN114454266A (en) * | 2022-02-14 | 2022-05-10 | 三菱电机自动化(中国)有限公司 | Log cutting device, method and computer readable medium |
| CN115359249A (en) * | 2022-10-21 | 2022-11-18 | 山东圣点世纪科技有限公司 | Palm image ROI region extraction method and system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103400134A (en) * | 2013-06-28 | 2013-11-20 | 哈尔滨工程大学 | Non-contact method for extracting region of interest of finger vein sample |
| CN105474234A (en) * | 2015-11-24 | 2016-04-06 | 厦门中控生物识别信息技术有限公司 | Method and apparatus for palm vein recognition |
-
2017
- 2017-11-22 CN CN201711175045.2A patent/CN107729883A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103400134A (en) * | 2013-06-28 | 2013-11-20 | 哈尔滨工程大学 | Non-contact method for extracting region of interest of finger vein sample |
| CN105474234A (en) * | 2015-11-24 | 2016-04-06 | 厦门中控生物识别信息技术有限公司 | Method and apparatus for palm vein recognition |
Non-Patent Citations (2)
| Title |
|---|
| RANDA BOUKHRIS TRABELSI: ""Hand vein recognition system with circular difference and statistical directional patterns basedon an artificial neural network"", 《MULTIMEDIA TOOLS AND APPLICATIONS》 * |
| 王贺: ""基于特征融合的手背静脉识别算法研究"", 《中国优秀硕士论文全文数据库 信息科技辑》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109190517A (en) * | 2018-08-14 | 2019-01-11 | 北京凌云天润智能科技有限公司 | A kind of finger venous image edge extracting and direction antidote |
| CN109190517B (en) * | 2018-08-14 | 2022-05-10 | 北京凌云天润智能科技有限公司 | Finger vein image edge extraction and direction correction method |
| CN109345735A (en) * | 2018-10-12 | 2019-02-15 | 南京理工大学 | A kind of self-service machine commodity recognition method and system |
| WO2020135230A1 (en) * | 2018-12-29 | 2020-07-02 | 北京金山安全软件有限公司 | Inscribed circle determination method and device |
| CN109727289A (en) * | 2019-01-18 | 2019-05-07 | 珠海市万瑙特健康科技有限公司 | Location determining method, device and the computer equipment of pulse condition perception point |
| CN109784083A (en) * | 2019-02-22 | 2019-05-21 | 吉林大学 | The bionical encryption system merged based on grip information with hand back vein information |
| CN110032936A (en) * | 2019-03-08 | 2019-07-19 | 吉林大学 | The maximum round method for extracting region of printenv hand back vein |
| CN110032936B (en) * | 2019-03-08 | 2022-08-09 | 吉林大学 | Method for extracting maximum circular area of non-parameter hand back vein |
| CN110909631A (en) * | 2019-11-07 | 2020-03-24 | 黑龙江大学 | Finger vein image ROI extraction and enhancement method |
| CN114454266A (en) * | 2022-02-14 | 2022-05-10 | 三菱电机自动化(中国)有限公司 | Log cutting device, method and computer readable medium |
| CN114454266B (en) * | 2022-02-14 | 2023-09-08 | 三菱电机自动化(中国)有限公司 | Log cutting device, method and computer readable medium |
| CN115359249A (en) * | 2022-10-21 | 2022-11-18 | 山东圣点世纪科技有限公司 | Palm image ROI region extraction method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107729883A (en) | A kind of vein image area-of-interest exacting method | |
| WO2017059591A1 (en) | Finger vein identification method and device | |
| CN102368290B (en) | Hand gesture identification method based on finger advanced characteristic | |
| WO2021027364A1 (en) | Finger vein recognition-based identity authentication method and apparatus | |
| CN104899600B (en) | A kind of hand-characteristic point detecting method based on depth map | |
| CN106096569B (en) | A kind of finger vein identification method | |
| Lu et al. | Finger vein identification using polydirectional local line binary pattern | |
| WO2015067084A1 (en) | Human eye positioning method and apparatus | |
| CN101470800A (en) | Hand shape recognition method | |
| Zheng et al. | Application of projective invariants in hand geometry biometrics | |
| CN110909631B (en) | Finger vein image ROI extraction and enhancement method | |
| CN106446773A (en) | Automatic robust three-dimensional face detection method | |
| Yang et al. | Superpixel based finger vein ROI extraction with sensor interoperability | |
| CN102968624A (en) | Method for positioning human eyes in human face image | |
| CN104573634A (en) | Three-dimensional face recognition method | |
| CN105160305B (en) | A kind of multi-modal Feature fusion of finger | |
| CN106886741A (en) | A kind of gesture identification method of base finger identification | |
| CN106503619B (en) | Gesture recognition method based on BP neural network | |
| Wang et al. | Region of interest extraction for finger vein images with less information losses | |
| Sun et al. | A DSP-based finger vein authentication system | |
| CN110032936A (en) | The maximum round method for extracting region of printenv hand back vein | |
| CN106970701A (en) | A kind of gesture changes recognition methods | |
| CN115359249B (en) | Palm image ROI region extraction method and system | |
| Mokni et al. | Pre-processing and extraction of the ROIs steps for palmprints recognition system | |
| Zhang et al. | Improving feature based dorsal hand vein recognition through random keypoint generation and fine-grained matching |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180223 |
|
| RJ01 | Rejection of invention patent application after publication |