CN107992866B - Based on in vivo method for detecting eye video stream of glistenings - Google Patents

Based on in vivo method for detecting eye video stream of glistenings Download PDF

Info

Publication number
CN107992866B
CN107992866B CN201810267873.7A CN201810267873A CN107992866B CN 107992866 B CN107992866 B CN 107992866B CN 201810267873 A CN201810267873 A CN 201810267873A CN 107992866 B CN107992866 B CN 107992866B
Authority
CN
China
Prior art keywords
iris
image
glistenings
video stream
point
Prior art date
Application number
CN201810267873.7A
Other languages
Chinese (zh)
Other versions
CN107992866A (en
Inventor
宫雅卓
钟千里
Original Assignee
上海聚虹光电科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority to CN201711126157 priority Critical
Application filed by 上海聚虹光电科技有限公司 filed Critical 上海聚虹光电科技有限公司
Publication of CN107992866A publication Critical patent/CN107992866A/en
Application granted granted Critical
Publication of CN107992866B publication Critical patent/CN107992866B/en

Links

Abstract

目前多数虹膜采集设备采集虹膜图像时均使用点光源照亮用户双眼,此类虹膜采集设备采集的虹膜图像,在用户眼部普遍存在光源的高亮反光点,本发明公开了种基于视频流眼部反光点的活体检测方法,在虹膜采集程序进行虹膜采集时,通过图像中反光点信息进行活体检测,所述反光点信息包括反光点平均灰度、尺寸和位置,本发明充分利用了虹膜采集时眼部反光点信息,可以有效区分纸质打印眼部图像与活体人眼,具有计算量小,检测速度快的优点。 At present, most devices use iris capture an iris image acquired when the user eyes illuminated by the point light source, such iris image capture device to capture the iris, the user eyes bright reflective light source common point, the present invention discloses a kind of eye based on the video stream vivo method for detecting unit reflex points, when the iris acquisition program iris capture, for in vivo detection by image reflecting point information, the reflecting point information includes glistenings average gray, size and location, the present invention makes full use of iris capture when the eye reflective point information, the printing paper can effectively distinguish eye image and the living human eye, with a small amount of calculation, test speed advantages.

Description

基于视频流眼部反光点的活体检测方法 Based on in vivo method for detecting eye video stream of glistenings

技术领域 FIELD

[0001] 本发明涉及图像处理技术领域,特别涉及一种基于视频流眼部反光点的活体检测方法。 [0001] The present invention relates to the technical field of image processing, and particularly to a method for in vivo detection of eye video stream based on glistening.

背景技术 Background technique

[0002] 目前主流虹膜识别算法,误识率普遍低于百万分之一,因为具有极高的准确性,虹膜识别技术常被应用到高安全级别的场景中,如:银行、医院、监狱;在这种情况下,识别非活体,防止他人盗取虹膜信息以假眼欺骗虹膜识别系统,就显得尤为重要。 [0002] The current mainstream iris recognition algorithm, the error rate is generally lower than one in a million, because of an extremely high accuracy of iris recognition technology is often applied to the high levels of security scenarios, such as: banks, hospitals, prisons ; in this case, to identify non-living, to prevent others from stealing the iris false information to deceive the eye iris recognition systems, it is particularly important.

[0003] 多数虹膜采集设备通过点光源照亮用户双眼,使用此类虹膜采集设备采集的虹膜图像,在用户眼部普遍存在光源的高亮反光点,现有的虹膜识别活体检测方法,未对反光点信息进行充分利用。 [0003] Most iris acquisition device users eyes illuminated by a point light source, using the iris image acquisition device such acquired iris, the user common eye point source bright reflective, conventional iris recognition method for detecting a living body, not on glistenings information fully utilized.

发明内容 SUMMARY

[0004] 本发明旨在提供一种计算量小,检测速度快的虹膜识别活体检测方法,利用反光点平均灰度信息进行活体检测,包括以下步骤。 [0004] The present invention aims to provide a small amount of calculation, fast detection speed iris recognition in vivo detection method using average gray reflective dot detecting living body information, comprising the following steps.

[0005] 在虹膜采集程序进行虹膜采集时,通过图像中反光点信息进行活体检测,所述反光点信息包括反光点平均灰度;虹膜采集设备对视频流图像中眼部区域进行测光,在预设帧以预设的低曝光补偿进行图像采集,获得曝光偏低的图像,根据低曝光图像反光点平均灰度是否达到预设值,判断当前识别对象是否为活体,若平均灰度高于预设值,则通过检测;所述低曝光补偿是指以眼部图像达到_2ev为标准进行曝光,在图像像素灰度区间为0到255的情况下,所述反光点平均灰度的预设值为200。 [0005] In an iris for iris capture collection procedures, by reflecting the living body information detecting point in the image, the reflecting point information includes the average gray glistenings; iris image capture device in a video stream metering eye region, in preset frames to a preset low exposure compensation for image acquisition, image exposure is low, the low-exposure image in accordance with the average gray reflective dot reaches a preset value, it is determined whether the current identification object is a living body, if the average is higher than the gradation predetermined value, by detecting; in the case of the low-exposure image compensation refers to the eye to reach the standard _2ev exposure, in the image pixel grayscale range of 0 to 255, the average gray pre glistenings set a value of 200.

[0006] 所述反光点信息还包括反光点的大小,若图像中存在低于预设尺寸的反光点,则检测通过。 [0006] The size of the reflecting point information further comprises reflective dots, if present in the image is lower than a preset size glistenings is detected by. 所述反光点信息还包括反光点位置,所述反光点位置是指反光点与瞳孔和虹膜的相对位置,若视频流各帧之间反光点位置发生变化,则通过检测。 The reflector further includes reflecting point position information, said reflecting means reflecting point position relative positions of the pupil and the iris, reflecting changes in position between frames if the video stream, by detecting.

[0007] 本发明基于视频流眼部反光点的活体检测方法具有所需计算量小,检测速度快的优点,对于目前最常见的盗取虹膜图像进行纸质打印欺骗虹膜识别系统,具有良好的检测效果。 [0007] The present invention has the advantage of a small amount of calculation needed, test speed detection method based on a video stream in vivo ocular reflex points, for the most common steal iris image printing paper spoofing iris recognition system, having good detection.

附图说明 BRIEF DESCRIPTION

[0008] 图1是基于视频流眼部反光点的活体检测流程示意图。 [0008] FIG. 1 is a live video stream ocular reflex points detected based on a schematic flow diagram.

具体实施方式 Detailed ways

[0009] 本发明提出了一种基于视频流眼部反光点的活体检测方法,下面结合附图对本发明作进一步详细描述: [0009] The present invention provides a method for detecting a living body based on video stream ocular reflex points DRAWINGS The present invention will be further described in detail:

[0010] S101:人眼区域是否存在反光点: [0010] S101: eye area exists glistenings:

[0011] 目前多数虹膜采集设备采集虹膜图像时均使用点光源照亮用户双眼,所述反光点是指,使用此类虹膜采集设备采集的虹膜图像,在用户眼部的巩膜、虹膜和瞳孔图像中普遍存在的光源的高亮反光点,所述反光点具有以下特征: [0011] At present, most devices use iris capture an iris image acquired when the user eyes illuminated by the point light source, the reflecting point is, the iris image acquisition device such acquired iris, sclera in the user's eye, the iris and pupil image highlight glistenings prevalent source, the reflecting point has the following characteristics:

[0012] 1、图像中反光点与周围区域对比度高于对比度预设值;优选的本实施例中反光点与周围区域对比度的计算公式为C= | Smax-Smin | /(Smax+Smin) *100%,其中C是反光点与周围区域对比度,Smax是反光点内亮度最高的像素点的灰度值,Smin是反光点周围预设区域内亮度最低的像素点的灰度值; [0012] 1, the reflective dot image contrast higher than the surrounding area contrast preset value; embodiment glistenings contrast with surrounding areas is calculated according to the preferred embodiment is a C = | Smax-Smin | / (Smax + Smin) * 100%, where C is the area of ​​reflective contrast with the surrounding points, the gradation value Smax of the highest point of the reflective pixel luminance, Smin of a reflective point around the preset gradation value in the region of the lowest luminance of the pixels;

[0013] 2、单个反光点尺寸小于尺寸预设值; [0013] 2, a single reflective dot size smaller than a preset value;

[00M] 3、图像中反光点平均灰度高于灰度预设值,在图像像素灰度区间为〇到255的情况下,所述反光点平均灰度的灰度预设值为200;所述反光点平均灰度是指图像中反光点内所有像素的平均灰度。 [00M] 3, the image gray reflective dot is higher than the average gray value preset in the image pixel gray square section of the case 255 to the next, the reflecting point average gray gradation default value 200; the average gray reflective point is the average of all pixels within the grayscale image reflecting points.

[0015]若采集的虹膜图像不存在反光点,则对此帧图像不处理。 [0015] When the reflective iris image acquisition do not exist, then the image of this frame is not processed.

[0016] S102:反光点数量是否符合预设值: [0016] S102: the number of reflex points meets a preset value:

[0017]反光点是由于虹膜采集设备的点光源在用户眼部表面反射而产生,所以针对不同的虹膜采集设备,所配备的点光源数量不同,反光点的数量也会不同,虹膜采集设备点光源的位置、安装角度、光源功率和使用距离也会影响所采集的虹膜图像中反光点的数量,因此针对不同的虹膜采集设备,可以在活体检测算法设置相应的预设值,优选的,本实施例中反光点数量的预设值是一只眼中反光点数量小于五个,进一步的也可以设置一个或几个确定的值,比如一只眼中反光点数量是两个、三个或五个,反光点数量符合预设值。 [0017] The reflective iris capture point a point light source device generates the ocular surface reflection due to the user, so that for different iris acquisition device, the point light sources with different numbers, the number of points will be different reflective iris acquisition device point the position of the light source, mounting angle, use of light power and distance also affect the number of acquired iris image glistenings, so for different iris acquisition device, may set the corresponding preset value, preferably in vivo detection algorithm, this default number of glistenings eyes embodiment is a reflective dot number less than five, may be further provided to determine one or more values, such as the number of the eyes of a reflective dot is two, three, or five , reflecting the number of points in line with the preset value.

[0018] S103:低曝光采集图像,反光点平均灰度是否大于预设值: [0018] S103: low-exposure image acquisition, reflective dot is greater than a predetermined average gray value:

[0019] 优选的S102检测通过后的三帧图片,均以眼部区域-2ev的曝光标准进行采集,判断低曝光图像中是否存在高于灰度阈值的反光点,若是,则检测通过。 After [0019] Preferred S102 is detected by three images, are exposed standard eye region -2ev acquisition, determines whether there is a threshold gray reflective point above the low-exposure image, if it is detected by. 在图像像素灰度区间为0到255的情况下,优选的预设灰度阈值为200。 Case where the image pixel grayscale range of 0 to 255, preferably 200 to a predetermined gradation threshold value. 此判断方法可以有效区分纸质反光点和活体反光点,对于纸质打印的虹膜图像,实际采集时纸张表面并不存在反光点,所述纸质反光点是指虹膜图像本身有反光点并被打印到纸面,因此所述纸质反光点会在曝光降低时,灰度明显降低;而活体反光点是虹膜采集设备点光源在用户眼部的反光,在一定范围内曝光降低,反光点平均灰度不会明显降低,可接受的不会使活体人眼反光点平均灰度明显降低的曝光降低范围受到虹膜采集设备点光源的数量和位置、安装角度、光源功率以及使用距离等因素的影响,通常情况下眼部区域的曝光标准降低为-2ev时,活体反光点平均灰度仍大于200。 This determination method can effectively distinguish the reflective paper reflective dot point and in vivo, for the iris image is printed paper, the paper surface does not exist during the actual acquisition reflex points, the paper point is reflective iris image itself and glistenings when the printing paper, so the paper can reduce the exposure of glistenings gradation decreased; glistenings the living body is a point light source reflective iris acquisition device in the user's eyes, to reduce the exposure within a certain range, reflecting point average gradation not significantly reduce, not acceptable that the living human eye reflex point average gray exposure significantly reduced by reducing the range of the number and location of point light sources iris acquisition device, the mounting angle factors, and the use of the power source distance when the exposure of the standard eye region is normally reduced -2EV, reflective dot living body 200 is still greater than the average gray.

[0020] S104:反光点位置是否发生变化: [0020] S104: reflecting point position is changed:

[0021]虹膜采集程序判断视频流连续几帧图像之间反光点位置是否发生变化,若是,则检测通过;在虹膜采集时,由于活体人眼总会下意识的略微转动,因此反光点和虹膜与瞳孔的相对位置就会发生变化,此判断方法可以有效区分活体人眼与静态图像;优选的,在本实施例中虹膜采集程序获取连续的20帧图像,并判断所述20帧图像的各帧之间反光点位置是否发生变化。 [0021] The reflective iris capture program determines whether a position change between several successive images of the video stream of frames, if, by the detection; iris capture at slightly rotated due to the living human eye always subconsciously, and the iris and therefore glistenings the relative position of the pupil will change, this determination method can effectively distinguish living human eye still images; preferably, in the present embodiment, successive acquired iris image acquisition program 20, and determines the image of each frame 20 a change point position between the reflective occurs.

[0022]在本发明的另一实施例中,所述反光点位置变化,通过控制虹膜采集设备红外灯来产生,优选的虹膜采集设备正面平均分布一排6颗红外灯,正常采集虹膜图像时,6颗红外灯全部亮起。 When the embodiment, the reflecting point position variation [0022] In another embodiment of the present invention, be produced by controlling the iris acquisition device infrared lamp, preferably a front iris acquisition device 6 evenly distributed a row of infrared light, the iris image acquisition normal , six infrared LEDs light up. 活体检测时,采集每一帧虹膜图像期间随机点亮一颗红外灯,进而虹膜采集程序连续获取几帧图像,活体检测算法分析各帧之间反光点和虹膜与瞳孔的相对位置变化情况是否和红外灯点亮顺序相对应,若是,则检测通过。 Detecting living body, each frame randomly collected infrared light is lit during an iris image, an iris acquisition program and thus acquires several successive frame images, in vivo detection algorithm and analysis glistenings relative positional change between the iris and the pupil, and whether the frame where infrared light corresponding to the lighting sequence, if it is detected by. 优选的,在本实施例中虹膜采集程序获取连续的6帧图像,并判断所述6帧图像的各帧之间反光点位置是否发生变化。 Preferably, in this embodiment the iris acquisition program acquiring successive images embodiment 6, and determines whether the change in position of reflective dot between the six image frames. 接下来进行正常采集,对采集到的图像进行人眼定位,图像分割,编码提取,完成虹膜注册或虹膜识别。 Next, the normal collection of images collected eye location, image segmentation, coding extraction, to complete the registration iris or iris recognition.

[0023] 进一步的,本发明还可以检测反光点尺寸是否达到阈值: [0023] Further, the present invention can also detect glistenings size has reached the threshold value:

[0024]虹膜采集程序以正常曝光进行图片采集,若图像中存在低于预设尺寸的反光点, 则检测通过。 [0024] iris image acquisition program acquired at the normal exposure, if present in the image is lower than a preset size glistenings is detected by. 优选的预设反光点尺寸最大为2〇个像素点,所述反光点预设尺寸由使用虹膜采集设备采集活体人眼图像实测确定,受到使用距离、镜头焦距、图像传感器分辨率的影响,对于不同设备本发明基于视频流反光点的活体检测方法的反光点预设尺寸可能具有较大差别。 The preferred default maximum size glistenings 2〇 pixel points, the reflective dot size determined by the predetermined iris collecting devices living eye images found by the distance, affect the focal length of the lens, the image sensor resolution, for the present invention is based on a different detection method reflective glistenings video stream of a living body having a predetermined size may be quite different.

[0025]在实施例中,为了保证活体检测的准确度,除了利用反光点平均灰度信息进行活体检测,还通过虹膜图像中反光点的大小、位置等信息做进一步的活体检测,在其他实施例中,也可仅通过反光点平均灰度信息的对比进行活体检测;或者在反光点的大小、位置等信息中选择一种或几种,并连同反光点的灰度信息进行活体检测。 [0025] In an embodiment, in order to ensure the accuracy of the living body is detected, in addition to using the reflective point average gray information vivo testing, but also for further living body is detected by the information of the size of the iris image glistenings, location, etc., implemented in other embodiment, may also be reflective point by comparing the average gray living body information detection only; or select one or more points of information reflecting the size, position, etc., together with reflective gray scale information detecting point is performed in vivo. 应当理解,基于改变本实施例检测步骤的排列顺序,或删减本实施例检测步骤,所获得的所有其他实施例,都属于本发明保护的范围。 It should be understood that the order in the present embodiment based on the change detecting step embodiment, the present embodiment the detection step, or deletion embodiment, all other embodiments obtained by, fall within the scope of the present invention.

Claims (4)

  1. . 一种^于视频流眼部反光点的活体检测方法,在虹膜采集程序进行虹膜采集时,通过米集的虹图像中的反光点信息进行活体检测,其特征在于所述反光点信息包括反光点平均灰度f膜采集设备对视频流图像眼部区域进行测光,在预设帧以预设的低曝光补偿进行图像采集,获得曝光偏低的图像,根据低曝光图像反光点平均灰度是否达到预设值,判断当前识别对象是否是活体,若平均灰度高于预设值,则判断当前识别对象是活体,通过活体检测;若平均灰度等于或低于预设值,则判断当前识别对象不是活体,不通过活体检测。 When A ^ in vivo method for detecting eye video stream glistenings in iris capture iris capture procedures, for in vivo detection by image glistenings rainbow meter set in the information, wherein said information comprises a reflective glistenings f-point average gray film collecting device to a video stream photometric eye region image, an image acquisition frames at a predetermined preset low exposure compensation, to obtain an exposure image is low, the low-exposure image in accordance with the average gray glistenings reaches a preset value, it is determined whether the current object is a living body identification, if the average gray scale is higher than a preset value, it is determined that the current object is a living body identified by the living body is detected; if the average gradation value is equal to or lower than a preset, it is determined identifying the current object is not a living body, it is not detected by the living body.
  2. 2. 根据权利要求1所述的基于视频流眼部反光点的活体检测方法,其特征在于所述低曝光补偿是指以眼部图像达到_2ev为标准进行曝光。 According to claim vivo detection method based on a video stream ocular reflex points, wherein said one of the low-exposure image compensation means to achieve _2ev eye exposure standard.
  3. 3. 根据权利要求1或2所述的基于视频流眼部反光点的活体检测方法,其特征在于以所述低曝光补偿拍摄的图像,在图像像素灰度区间为〇到255的情况下,所述反光点平均灰度的预设值为200。 The method for detecting live video stream based on the eye of the reflecting point 1 or claim 2, wherein said low exposure compensation image captured in the image pixel gray square to the case where the interval 255, the reflective preset point average gray value of 200.
  4. 4. 根据权利要求1所述的基于视频流眼部反光点的活体检测方法,其特征在于所述反光点信息还包括反光点位置,所述反光点位置是指反光点与瞳孔和虹膜的相对位置,对视频流预设数量的连续帧的反光点位置进行判断,若各1)1贞之间反光点位置发生变化,则通过活体检测。 The detection method based on a video stream in vivo ocular reflex points, wherein the reflecting point 1, further comprising a reflective point position information, the position of the reflecting point is the relative pupil and iris glistenings claims position, the reflecting position of the video stream of a predetermined number of consecutive frames determines if) reflecting point position is changed between 1 and 1 Chen, detected through the living body.
CN201810267873.7A 2017-11-15 2018-03-29 Based on in vivo method for detecting eye video stream of glistenings CN107992866B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201711126157 2017-11-15

Publications (2)

Publication Number Publication Date
CN107992866A CN107992866A (en) 2018-05-04
CN107992866B true CN107992866B (en) 2018-06-29

Family

ID=62040457

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810267873.7A CN107992866B (en) 2017-11-15 2018-03-29 Based on in vivo method for detecting eye video stream of glistenings

Country Status (1)

Country Link
CN (1) CN107992866B (en)

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104537292A (en) * 2012-08-10 2015-04-22 眼验有限责任公司 Method and system for spoof detection for biometric authentication
CN104781830A (en) * 2012-11-19 2015-07-15 虹膜技术公司 Method and apparatus for identifying living eye
CN104834908A (en) * 2015-05-07 2015-08-12 惠州Tcl移动通信有限公司 Image exposure method for mobile terminal based on eye pattern recognition and exposure system
CN105354545A (en) * 2015-10-28 2016-02-24 广东欧珀移动通信有限公司 Iris information acquisition method and acquisition system
US9594969B1 (en) * 2015-11-19 2017-03-14 Intel Corporation Iris recognition including liveness testing
CN107111704A (en) * 2014-10-30 2017-08-29 达美生物识别科技有限公司 Systems and methods for spoof detection in iris based biometric systems
CN107423699A (en) * 2017-07-14 2017-12-01 广东欧珀移动通信有限公司 In-vivo detection method and related product

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8260008B2 (en) * 2005-11-11 2012-09-04 Eyelock, Inc. Methods for performing biometric recognition of a human eye and corroboration of same
KR101286454B1 (en) * 2012-02-29 2013-07-16 주식회사 슈프리마 Fake face identification apparatus and method using characteristic of eye image
US9767358B2 (en) * 2014-10-22 2017-09-19 Veridium Ip Limited Systems and methods for performing iris identification and verification using mobile devices

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104537292A (en) * 2012-08-10 2015-04-22 眼验有限责任公司 Method and system for spoof detection for biometric authentication
CN104781830A (en) * 2012-11-19 2015-07-15 虹膜技术公司 Method and apparatus for identifying living eye
CN107111704A (en) * 2014-10-30 2017-08-29 达美生物识别科技有限公司 Systems and methods for spoof detection in iris based biometric systems
CN104834908A (en) * 2015-05-07 2015-08-12 惠州Tcl移动通信有限公司 Image exposure method for mobile terminal based on eye pattern recognition and exposure system
CN105354545A (en) * 2015-10-28 2016-02-24 广东欧珀移动通信有限公司 Iris information acquisition method and acquisition system
US9594969B1 (en) * 2015-11-19 2017-03-14 Intel Corporation Iris recognition including liveness testing
CN107423699A (en) * 2017-07-14 2017-12-01 广东欧珀移动通信有限公司 In-vivo detection method and related product

Also Published As

Publication number Publication date
CN107992866A (en) 2018-05-04

Similar Documents

Publication Publication Date Title
US8542879B1 (en) Facial recognition
EP2680192B1 (en) Facial recognition
CN104537292B (en) Spoofing detection for biometric authentication method and system
CN101142584B (en) Method for facial features detection
JP5530503B2 (en) A method and apparatus for line-of-sight measurement
CN1261905C (en) Eye position detection method and apparatus thereof
US20080112599A1 (en) method of detecting redeye in a digital image
US9280706B2 (en) Efficient method and system for the acquisition of scene imagery and iris imagery using a single sensor
JP5567853B2 (en) Image recognition apparatus and method
JP5297486B2 (en) It detects eyes and gaze direction, tracking device
KR101159830B1 (en) Red eye false positive filtering using face location and orientation
US8743274B2 (en) In-camera based method of detecting defect eye with high accuracy
US7146027B2 (en) Iris recognition method and system using the same
US8798334B2 (en) Methods for performing biometric recognition of a human eye and corroboration of same
US7948524B2 (en) Image processor and face detector using the same
KR100888554B1 (en) Recognition system
US9002073B2 (en) Mobile identity platform
US9940516B2 (en) Mobile identity platform
US9036871B2 (en) Mobility identity platform
US7253739B2 (en) System and method for determining eye closure state
US7801335B2 (en) Apparatus and methods for detecting the presence of a human eye
JP4078334B2 (en) Image processing apparatus and image processing method
US7784943B2 (en) Eyelid detecting apparatus, eyelid detecting method and program thereof
JP4307496B2 (en) Facial part detection device and program
JP5139533B2 (en) Detection of red-eye defects in digital images

Legal Events

Date Code Title Description
PB01
SE01
GR01