WO2018040307A1 - 一种基于红外可见双目图像的活体检测方法及装置 - Google Patents

一种基于红外可见双目图像的活体检测方法及装置 Download PDF

Info

Publication number
WO2018040307A1
WO2018040307A1 PCT/CN2016/106673 CN2016106673W WO2018040307A1 WO 2018040307 A1 WO2018040307 A1 WO 2018040307A1 CN 2016106673 W CN2016106673 W CN 2016106673W WO 2018040307 A1 WO2018040307 A1 WO 2018040307A1
Authority
WO
WIPO (PCT)
Prior art keywords
identified
visible light
infrared
light imaging
camera
Prior art date
Application number
PCT/CN2016/106673
Other languages
English (en)
French (fr)
Inventor
陈远浩
Original Assignee
上海依图网络科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 上海依图网络科技有限公司 filed Critical 上海依图网络科技有限公司
Priority to SG11201802652YA priority Critical patent/SG11201802652YA/en
Publication of WO2018040307A1 publication Critical patent/WO2018040307A1/zh
Priority to PH12018500689A priority patent/PH12018500689A1/en

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/40Spoof detection, e.g. liveness detection
    • G06V40/45Detection of the body part being alive
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16Human faces, e.g. facial parts, sketches or expressions
    • G06V40/161Detection; Localisation; Normalisation
    • G06V40/165Detection; Localisation; Normalisation using facial parts and geometric relationships
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16Human faces, e.g. facial parts, sketches or expressions
    • G06V40/168Feature extraction; Face representation
    • G06V40/171Local features and components; Facial parts ; Occluding parts, e.g. glasses; Geometrical relationships
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16Human faces, e.g. facial parts, sketches or expressions
    • G06V40/172Classification, e.g. identification

Definitions

  • the present invention relates to the field of video security, and in particular to a living body detection method and apparatus based on an infrared visible binocular image.
  • the visual information-based living body detection in fixed places has a wide range of applications in many scenarios, such as community access control, bank self-service withdrawals, and automatic processing by government departments. These scenarios use face recognition to verify the identity of the person being treated, but require live detection to prevent attacks.
  • Common attack methods are: photos, videos, hole masks, 3d masks, etc. There is currently no mainstream reliable solution. Common practices include: detecting blinks to prevent photo attacks; detecting multiple faces and three-dimensional information from multiple cameras in different locations to prevent simple photo-video attacks; preventing simple video attacks through texture information on the screen; The temperature sensor protects against simple video attacks.
  • the object of the present invention is to provide a living body detecting method and apparatus based on infrared visible binocular images in order to overcome the defects of the prior art described above.
  • a method of living body detection comprising:
  • S1 collecting infrared light imaging of the object to be identified through an infrared camera, and collecting visible light imaging of the object to be identified through the visible light camera;
  • step S2 if there is no image in the infrared light imaging, it is determined that the object to be identified is not a living body, otherwise, step S3 is performed;
  • S3 extracting features of the eye and face of the object to be identified in visible light imaging and infrared light imaging, and base The extracted feature determines whether the object to be identified is a living body.
  • the step S3 specifically includes the following steps:
  • step S31 determining whether there is a face in visible light imaging and infrared light imaging, or if yes, proceeding to step S32, if not, repeating step S31;
  • S32 Extracting features of the face of the object to be identified in visible light imaging and infrared light imaging. If the tip of the nose is reflected in the infrared light image and the cheeks are dark, proceed to S33.
  • S33 Extracting features of the eye of the object to be identified in visible light imaging and infrared light imaging. If the pupil is reflected in the infrared light imaging and the white of the eye is gray, it is determined that the object to be identified is a living body, and vice versa.
  • step S3 After being identified as a living body in step S3, the steps are further performed:
  • step S6 according to the data obtained in steps S4 and S5, combined with the position and angle relationship between the infrared camera and the visible light camera to determine whether the object to be recognized captured by the infrared camera and the visible light camera is the same object, and if so, maintain The determination that the object is a living body is determined, and if not, it is determined that the object to be identified is not a living body.
  • An angle formed between the front side of the object to be identified and the infrared camera is an angle formed directly in front of the object to be recognized and directly in front of the infrared camera, and an angle formed between the front side of the object to be recognized and the infrared camera is to be Identify the angle between the front of the object and the front of the visible light camera.
  • a device comprising:
  • a visible light camera for collecting visible light imaging of an object to be identified
  • An infrared camera for collecting infrared light imaging of an object to be identified
  • the computer is respectively connected to the visible light camera and the infrared camera for determining whether the object to be identified is a living body according to visible light imaging and infrared light imaging.
  • the computer includes:
  • a pre-identification module configured to determine whether the image is non-living based on whether there is an image in the infrared light imaging
  • the feature recognition module is connected to the pre-identification module, and is configured to extract features of the eye and the face of the object to be recognized in the visible light imaging and the infrared light imaging after the verification of the pre-identification module, and determine whether the object to be identified is a living body based on the extracted features. .
  • the feature recognition module includes:
  • the first identifying unit is configured to extract features of the face of the object to be recognized in the visible light imaging and the infrared light imaging. If the nose is reflective and the cheek is dark in the infrared imaging, the object to be identified is determined to be a living body, and the opposite is a non-living body.
  • the second identifying unit is configured to extract features of the eye to be recognized in the visible light imaging and the infrared light imaging. If the pupil is reflected in the infrared light imaging and the gray of the eye is gray, the object to be identified is determined to be a living body, and vice versa.
  • the computer also includes:
  • the distance recognition module is connected to the feature recognition module, and is configured to respectively determine a distance between the object to be identified and the infrared camera and the visible light camera after the feature recognition module passes the verification, and between the infrared camera and the visible light camera respectively in front of the object to be identified
  • the angle formed combined with the position and angle relationship between the infrared camera and the visible light camera, determines whether the object to be recognized captured by the infrared camera and the visible light camera is the same object, and if so, the determination to maintain the object to be identified as a living object If no, it is determined that the object to be identified is not a living body.
  • the present invention has the following advantages:
  • Figure 2 is a schematic structural view of the device of the present invention.
  • a method of living body detection includes:
  • S1 collecting infrared light imaging of the object to be identified through an infrared camera, and collecting visible light imaging of the object to be identified through the visible light camera;
  • step S2 if there is no image in the infrared light imaging, it is determined that the object to be identified is not a living body, otherwise, step S3 is performed;
  • S3 extracting features of the eye and the face of the object to be recognized in the visible light imaging and the infrared light imaging, and determining whether the object to be identified is a living body based on the extracted features.
  • the shooting scene of the photo would be a point source or a parallel source, and there would be only one feature.
  • Under the illumination of the infrared fill light there are obvious traces of the point light on the face (such as the tip of the nose is reflected closer to the screen, the cheeks are darker, etc.), and the visible light picture does not have these features.
  • the implementation of these classifications uses the leading deep learning techniques of academia at home and abroad. Combining our years of practical experience in the visual industry, we finally got a classifier with good performance, fast speed and small size.
  • the real human eyes under infrared light will have pupil reflexes, gray eyes and other features, and the pupils under visible light will not have these features.
  • the use of visible light infrared light to distinguish the characteristics of the human eye can further enhance the anti-attack capability.
  • Step S3 specifically includes the steps of:
  • step S31 determining whether there is a face in both visible light imaging and infrared light imaging, if yes, step S32 is performed, and if no, step S32 is performed;
  • S32 Extracting features of the face to be recognized and features of the eye in visible light imaging and infrared light imaging.
  • step S3 After being identified as a living body in step S3, the steps are further performed:
  • step S6 according to the data obtained in steps S4 and S5, combined with the position and angle relationship between the infrared camera and the visible light camera to determine whether the object to be recognized captured by the infrared camera and the visible light camera is the same object, and if so, maintain The determination that the object is a living body is determined, and if not, it is determined that the object to be identified is not a living body.
  • the output is judged by the living body, and the best quality face is selected.
  • the angle formed between the front of the object to be recognized and the infrared camera is an angle formed directly in front of the object to be recognized and directly in front of the infrared camera, and an angle formed between the front of the object to be recognized and the infrared camera is directly in front of the object to be identified.
  • the angle formed directly in front of the visible light camera is an angle formed directly in front of the visible light camera.
  • a device as shown in Figure 2, includes:
  • a visible light camera 1 for collecting visible light imaging of an object to be identified
  • Infrared camera 2 for collecting infrared light imaging of an object to be identified
  • the computer 3 is respectively connected to the visible light camera 1 and the infrared camera 2 for determining whether the object to be identified is a living body according to visible light imaging and infrared light imaging.
  • Computer 3 includes:
  • a pre-identification module configured to determine whether the image is non-living based on whether there is an image in the infrared light imaging
  • the feature recognition module is connected to the pre-identification module, and is configured to extract features of the eye and the face of the object to be recognized in the visible light imaging and the infrared light imaging after the verification of the pre-identification module, and determine whether the object to be identified is a living body based on the extracted features. .
  • the feature recognition module includes:
  • the first identifying unit is configured to extract features of the face of the object to be recognized in the visible light imaging and the infrared light imaging. If the nose is reflective and the cheek is dark in the infrared imaging, the object to be identified is determined to be a living body, and the opposite is a non-living body.
  • the second identifying unit is configured to extract features of the eye to be recognized in the visible light imaging and the infrared light imaging. If the pupil is reflected in the infrared light imaging and the gray of the eye is gray, the object to be identified is determined to be a living body, and vice versa.
  • the computer also includes:
  • the distance recognition module is connected to the feature recognition module, and is configured to respectively determine a distance between the object to be identified and the infrared camera and the visible light camera after the feature recognition module passes the verification, and between the infrared camera and the visible light camera respectively in front of the object to be identified
  • the angle formed combined with the position and angle relationship between the infrared camera and the visible light camera, determines whether the object to be recognized captured by the infrared camera and the visible light camera is the same object, and if so, the determination to maintain the object to be identified as a living object If no, it is determined that the object to be identified is not a living body.
  • the whole system can guarantee 99.9% correct rate in the case of excellent illumination, 0% pass rate of video attack, and other attack mode less than 0.1% pass rate.
  • the speed can be judged within half a second.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Human Computer Interaction (AREA)
  • Health & Medical Sciences (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • General Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Geometry (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
  • Endoscopes (AREA)

Abstract

本发明涉及一种基于红外可见双目图像的活体检测方法及装置,其中方法包括:S1:通过红外线摄像头采集待识别对象的红外光成像,并通过可见光摄像头采集待识别对象的可见光成像;S2:若红外光成像不存在图像,判定待识别对象为非活体,反之,则执行步骤S3;S3:提取可见光成像和红外光成像中待识别对象眼部和脸部的特征,并基于提取的特征判断待识别对象是否为活体。与现有技术相比,本发明先通过红外摄像头能否采集图像进行预识别,这样可以防住全部的视频攻击、大部分的照片、孔洞攻击、3d面具攻击,再通过对可见光成像和红外光成像之间的特征对比,最终保证了极大的非活体攻击。

Description

一种基于红外可见双目图像的活体检测方法及装置 技术领域
本发明涉及一种视频安防领域,尤其是涉及一种基于红外可见双目图像的活体检测方法及装置。
背景技术
固定场所的基于视觉信息的活体检测,在很多场景有着广泛的应用,比如小区门禁、银行自助取款、政府部门自动办理手续。这些场景使用人脸识别核实办理人的身份,但是需要活体检测手段来防止攻击。常见的攻击方法有:照片、视频、孔洞面具、3d面具等,目前尚无主流的可靠的解决方案。常见的做法有:通过检测眨眼来防住照片攻击;通过不同位置的多个摄像头检测人脸三维信息来防住简单的照片视频攻击;通过屏幕的纹理信息防住简单的视频攻击;通过红外感温传感器防住简单的视频攻击等。
上面提到的诸多做法均存在如下问题:首先,没有一套系统能够防住所有攻击。这样会给攻击者留有可趁之机;其次,性能很难保证,很多场景对安全和用户体验均有较高要求,比如银行的自助取款。
发明内容
本发明的目的就是为了克服上述现有技术存在的缺陷而提供一种基于红外可见双目图像的活体检测方法及装置。
本发明的目的可以通过以下技术方案来实现:
一种活体检测的方法,包括:
S1:通过红外线摄像头采集待识别对象的红外光成像,并通过可见光摄像头采集待识别对象的可见光成像;
S2:若红外光成像不存在图像,判定待识别对象为非活体,反之,则执行步骤S3;
S3:提取可见光成像和红外光成像中待识别对象眼部和脸部的特征,并基 于提取的特征判断待识别对象是否为活体。
所述步骤S3具体包括步骤:
S31:判断可见光成像和红外光成像是否存在脸部,或为是,则执行步骤S32,若为否,则重复步骤S31;
S32:提取可见光成像和红外光成像中待识别对象脸部的特征。若红外光成像中鼻尖反光且脸颊偏暗,则进行S33。
S33:提取可见光成像和红外光成像中待识别对象眼部的特征。若红外光成像中瞳孔反光且眼白泛灰,则判定待识别对象为活体,反之为非活体。
当步骤S3中识别为活体后,还执行步骤:
S4:分别确定待识别对象与红外线摄像头和可见光摄像头之间的距离;
S5:确定待识别对象正前方分别与红外线摄像头和可见光摄像头之间所成的角度;
S6:根据步骤S4和S5中得到的数据,结合红外线摄像头和可见光摄像头之间的位置和角度关系确定红外线摄像头和可见光摄像头所捕捉到的待识别对象是否为同一物体,若为是,则维持待识别对象为活体的判定,若为否,则判定待识别对象为非活体。
所述待识别对象正前方与红外线摄像头之间所成的角度为待识别对象正前方与红外线摄像头正前方所成的角度,所述待识别对象正前方与红外线摄像头之间所成的角度为待识别对象正前方与可见光摄像头正前方所成的角度。
一种装置,包括:
可见光摄像头,用于采集待识别对象的可见光成像;
红外线摄像头,用于采集待识别对象的红外光成像;
计算机,分别与可见光摄像头和红外线摄像头连接,用于根据可见光成像和红外光成像判定待识别对象是否为活体。
所述计算机包括:
预识别模块,用于基于红外光成像中是否存在图像判定是否为非活体;
特征识别模块,与预识别模块连接,用于在预识别模块验证通过后提取可见光成像和红外光成像中待识别对象眼部和脸部的特征,并基于提取的特征判断待识别对象是否为活体。
所述特征识别模块包括:
第一识别单元,用于提取可见光成像和红外光成像中待识别对象脸部的特征,若红外光成像中鼻尖反光且脸颊偏暗,则判定待识别对象为活体,反之为非活体
第二识别单元,用于提取可见光成像和红外光成像中待识别对象眼部的特征,若红外光成像中瞳孔反光且眼白泛灰,则判定待识别对象为活体,反之为非活体。
所述计算机还包括:
距离识别模块,与特征识别模块连接,用于在特征识别模块验证通过后分别确定待识别对象与红外线摄像头和可见光摄像头之间的距离,以及待识别对象正前方分别与红外线摄像头和可见光摄像头之间所成的角度,并结合红外线摄像头和可见光摄像头之间的位置和角度关系确定红外线摄像头和可见光摄像头所捕捉到的待识别对象是否为同一物体,若为是,则维持待识别对象为活体的判定,若为否,则判定待识别对象为非活体。
与现有技术相比,本发明具有以下优点:
1)先通过红外摄像头能否采集图像进行预识别,这样可以防住全部的视频攻击、大部分的照片、孔洞攻击、3d面具攻击,再通过对可见光成像和红外光成像之间的特征对比,最终保证了极大的非活体攻击。
2)通过鼻尖反光、脸颊偏暗、瞳孔反光、眼白泛灰等特征进行活体识别,整个系统能够保证在光照优良的情况下99.9%的通过率,视频攻击0%的通过率,其他攻击方式小于0.1%的通过率。速度上能够做到半秒内完成判断。
3)使用红外摄像头和可见红外双目摄像头的分类:能够真正做到防住所有攻击,保证通用性。
4)使用学术界最新的分类技术:性能、速度更有保证,更好的用户体验。
附图说明
图1为本发明方法的主要步骤流程示意图;
图2为本发明装置的结构示意图;
其中:1、可见光摄像头,2、红外线摄像头,3、计算机。
具体实施方式
下面结合附图和具体实施例对本发明进行详细说明。本实施例以本发明技术方案为前提进行实施,给出了详细的实施方式和具体的操作过程,但本发明的保护范围不限于下述的实施例。
发明人经过自己的反复探索发现:大部分纸质材料、全部的视频材料在红外光摄像头下均是不可显示的,而真人在红外光摄像头会正常出现。
一种活体检测的方法,如图1所示,包括:
S1:通过红外线摄像头采集待识别对象的红外光成像,并通过可见光摄像头采集待识别对象的可见光成像;
S2:若红外光成像不存在图像,判定待识别对象为非活体,反之,则执行步骤S3;
S3:提取可见光成像和红外光成像中待识别对象眼部和脸部的特征,并基于提取的特征判断待识别对象是否为活体。
发明人还发现,对于照片攻击,照片的拍摄场景会是点光源或平行光源,只会有一种特征。而真人在红外补光灯的照射下,脸部会有明显的点光源的痕迹(比如鼻尖因更靠近屏幕而反光、脸颊偏暗等),同时可见光的图片不会有这些特征。以此方法能够防住剩下的照片、孔洞攻击,同时几乎不会影响真人的通过率。这些分类的实现使用了国内外学术界领先的深度学习的技术。结合上我们在视觉行业的多年实战经验,最终得到了一个个性能好、速度快、尺寸小的分类器。另外,对于攻击时没有露出真实人眼的场景(比如照片攻击、3d面具攻击),红外光下的真实人眼会有瞳孔反光、眼白泛灰等特征,而可见光下瞳孔不会有这些特征。利用可见光红外光下人眼特征的不同,能够进一步提升防攻击能力。
步骤S3具体包括步骤:
S31:判断可见光成像和红外光成像是否都存在脸部,若为是,则执行步骤S32,若为否,则执行步骤S32;
S32:提取可见光成像和红外光成像中待识别对象脸部的特征和眼部的特征。
S34:若红外光成像中脸部特征的鼻尖反光且脸颊偏暗,眼部特征的瞳孔反光且眼白泛灰,而可见光成像中没有上述特征,则判定待识别对象为活体,反之为非活体。
当步骤S3中识别为活体后,还执行步骤:
S4:分别确定待识别对象与红外线摄像头和可见光摄像头之间的距离;
S5:确定待识别对象正前方分别与红外线摄像头和可见光摄像头之间所成的角度;
S6:根据步骤S4和S5中得到的数据,结合红外线摄像头和可见光摄像头之间的位置和角度关系确定红外线摄像头和可见光摄像头所捕捉到的待识别对象是否为同一物体,若为是,则维持待识别对象为活体的判定,若为否,则判定待识别对象为非活体。
最终输出活体判断,挑选出的质量最好的人脸。
待识别对象正前方与红外线摄像头之间所成的角度为待识别对象正前方与红外线摄像头正前方所成的角度,待识别对象正前方与红外线摄像头之间所成的角度为待识别对象正前方与可见光摄像头正前方所成的角度。
一种装置,如图2所示,包括:
可见光摄像头1,用于采集待识别对象的可见光成像;
红外线摄像头2,用于采集待识别对象的红外光成像;
计算机3,分别与可见光摄像头1和红外线摄像头2连接,用于根据可见光成像和红外光成像判定待识别对象是否为活体。
计算机3包括:
预识别模块,用于基于红外光成像中是否存在图像判定是否为非活体;
特征识别模块,与预识别模块连接,用于在预识别模块验证通过后提取可见光成像和红外光成像中待识别对象眼部和脸部的特征,并基于提取的特征判断待识别对象是否为活体。
特征识别模块包括:
第一识别单元,用于提取可见光成像和红外光成像中待识别对象脸部的特征,若红外光成像中鼻尖反光且脸颊偏暗,则判定待识别对象为活体,反之为非活体
第二识别单元,用于提取可见光成像和红外光成像中待识别对象眼部的特征,若红外光成像中瞳孔反光且眼白泛灰,则判定待识别对象为活体,反之为非活体。
计算机还包括:
距离识别模块,与特征识别模块连接,用于在特征识别模块验证通过后分别确定待识别对象与红外线摄像头和可见光摄像头之间的距离,以及待识别对象正前方分别与红外线摄像头和可见光摄像头之间所成的角度,并结合红外线摄像头和可见光摄像头之间的位置和角度关系确定红外线摄像头和可见光摄像头所捕捉到的待识别对象是否为同一物体,若为是,则维持待识别对象为活体的判定,若为否,则判定待识别对象为非活体。
整个系统能够保证在光照优良的情况下99.9%的正确率,视频攻击0%的通过率,其他攻击方式小于0.1%的通过率。速度上能够做到半秒内完成判断。

Claims (8)

  1. 一种活体检测的方法,其特征在于,包括:
    S1:通过红外线摄像头采集待识别对象的红外光成像,并通过可见光摄像头采集待识别对象的可见光成像;
    S2:若红外光成像不存在图像,判定待识别对象为非活体,反之,则执行步骤S3;
    S3:提取可见光成像和红外光成像中待识别对象眼部和脸部的特征,并基于提取的特征判断待识别对象是否为活体。
  2. 根据权利要求1所述的一种活体检测的方法,其特征在于,所述步骤S3具体包括步骤:
    S31:判断可见光成像和红外光成像是否都存在脸部,或为是,则执行步骤S32,若为否,则重复步骤S31;
    S32:提取可见光成像和红外光成像中待识别对象脸部的特征。若红外光成像中鼻尖反光且脸颊偏暗,则进行S33。
    S33:提取可见光成像和红外光成像中待识别对象眼部的特征。若红外光成像中瞳孔反光且眼白泛灰,则判定待识别对象为活体,反之为非活体。
  3. 根据权利要求2所述的一种活体检测的方法,其特征在于,当步骤S3中识别为活体后,还执行步骤:
    S4:分别确定待识别对象与红外线摄像头和可见光摄像头之间的距离;
    S5:确定待识别对象正前方分别与红外线摄像头和可见光摄像头之间所成的角度;
    S6:根据步骤S4和S5中得到的数据,结合红外线摄像头和可见光摄像头之间的位置和角度关系确定红外线摄像头和可见光摄像头所捕捉到的待识别对象是否为同一物体,若为是,则维持待识别对象为活体的判定,若为否,则判定待识别对象为非活体。
  4. 根据权利要求3所述的一种活体检测的方法,其特征在于,所述待识别 对象正前方与红外线摄像头之间所成的角度为待识别对象正前方与红外线摄像头正前方所成的角度,所述待识别对象正前方与红外线摄像头之间所成的角度为待识别对象正前方与可见光摄像头正前方所成的角度。
  5. 一种实现权利要求1-4中任一所述方法的装置,其特征在于,包括:
    可见光摄像头,用于采集待识别对象的可见光成像;
    红外线摄像头,用于采集待识别对象的红外光成像;
    计算机,分别与可见光摄像头和红外线摄像头连接,用于根据可见光成像和红外光成像判定待识别对象是否为活体。
  6. 根据权利要求5所述的装置,其特征在于,所述计算机包括:
    预识别模块,用于基于红外光成像中是否存在图像判定是否为非活体;
    特征识别模块,与预识别模块连接,用于在预识别模块验证通过后提取可见光成像和红外光成像中待识别对象眼部和脸部的特征,并基于提取的特征判断待识别对象是否为活体。
  7. 根据权利要求5所述的装置,其特征在于,所述特征识别模块包括:
    第一识别单元,用于提取可见光成像和红外光成像中待识别对象脸部的特征,若红外光成像中鼻尖反光且脸颊偏暗,则判定待识别对象为活体,反之为非活体
    第二识别单元,用于提取可见光成像和红外光成像中待识别对象眼部的特征,若红外光成像中瞳孔反光且眼白泛灰,则判定待识别对象为活体,反之为非活体。
  8. 根据权利要求5所述的装置,其特征在于,所述计算机还包括:
    距离识别模块,与特征识别模块连接,用于在特征识别模块验证通过后分别确定待识别对象与红外线摄像头和可见光摄像头之间的距离,以及待识别对象正前方分别与红外线摄像头和可见光摄像头之间所成的角度,并结合红外线摄像头和可见光摄像头之间的位置和角度关系确定红外线摄像头和可见光摄像头所捕捉到的待识别对象是否为同一物体,若为是,则维持待识别对象为活体的判定,若为否,则判定待识别对象为非活体。
PCT/CN2016/106673 2016-08-31 2016-11-21 一种基于红外可见双目图像的活体检测方法及装置 WO2018040307A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
SG11201802652YA SG11201802652YA (en) 2016-08-31 2016-11-21 A living body detection method and device based on infrared and visible binocular images
PH12018500689A PH12018500689A1 (en) 2016-08-31 2018-03-27 A living body detection method and device based on infrared and visible binocular images

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201610790411.4 2016-08-31
CN201610790411.4A CN106372601B (zh) 2016-08-31 2016-08-31 一种基于红外可见双目图像的活体检测方法及装置

Publications (1)

Publication Number Publication Date
WO2018040307A1 true WO2018040307A1 (zh) 2018-03-08

Family

ID=57899631

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/106673 WO2018040307A1 (zh) 2016-08-31 2016-11-21 一种基于红外可见双目图像的活体检测方法及装置

Country Status (4)

Country Link
CN (1) CN106372601B (zh)
PH (1) PH12018500689A1 (zh)
SG (1) SG11201802652YA (zh)
WO (1) WO2018040307A1 (zh)

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110008878A (zh) * 2019-03-27 2019-07-12 中控智慧科技股份有限公司 一种人脸检测的防假方法及具备防假功能的人脸识别装置
CN110127468A (zh) * 2019-04-12 2019-08-16 深圳壹账通智能科技有限公司 电梯控制方法、装置、计算机可读存储介质和计算机设备
CN111191519A (zh) * 2019-12-09 2020-05-22 同济大学 一种用于移动供电装置用户接入的活体检测方法
CN111209870A (zh) * 2020-01-09 2020-05-29 杭州涂鸦信息技术有限公司 一种双目活体摄像头快速配准方法及其系统和装置
CN111414831A (zh) * 2020-03-13 2020-07-14 深圳市商汤科技有限公司 监测方法及系统、电子设备和存储介质
CN111814659A (zh) * 2020-07-07 2020-10-23 杭州海康威视数字技术股份有限公司 一种活体检测方法、和系统
CN113496209A (zh) * 2021-06-09 2021-10-12 湖南中惠旅智能科技有限责任公司 基于人脸识别的景区大门闸机的数据处理方法及系统
CN114827435A (zh) * 2021-01-28 2022-07-29 深圳绿米联创科技有限公司 免IR-Cut的视频流处理方法及装置、智能门锁及介质
CN115082972A (zh) * 2022-07-27 2022-09-20 山东圣点世纪科技有限公司 一种基于纹理rgb图像和静脉灰度图像的活体检测方法
CN117994865A (zh) * 2024-04-01 2024-05-07 杭州海康威视数字技术股份有限公司 一种双目面部匹配方法、装置、电子设备及存储介质

Families Citing this family (22)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107169405B (zh) * 2017-03-17 2020-07-03 上海云从企业发展有限公司 基于双目摄像机活体识别的方法及装置
CN107314504A (zh) * 2017-06-15 2017-11-03 美的集团武汉制冷设备有限公司 空调系统、空调器和进行人体热影像异常报警的方法
CN107633198A (zh) * 2017-07-25 2018-01-26 百度在线网络技术(北京)有限公司 活体检测方法、装置、设备及存储介质
CN110008783A (zh) 2018-01-04 2019-07-12 杭州海康威视数字技术股份有限公司 基于神经网络模型的人脸活体检测方法、装置及电子设备
KR102466997B1 (ko) * 2018-01-22 2022-11-14 삼성전자주식회사 라이브니스 검사 방법 및 장치
CN108197615A (zh) * 2018-03-07 2018-06-22 北京上古视觉科技有限公司 一种多媒体展示装置及系统
US11410458B2 (en) 2018-04-12 2022-08-09 Guangdong Oppo Mobile Telecommunications Corp., Ltd. Face identification method and apparatus, mobile terminal and storage medium
WO2019200575A1 (zh) * 2018-04-18 2019-10-24 深圳阜时科技有限公司 身份鉴权方法、身份鉴权装置、和电子设备
CN108764071B (zh) * 2018-05-11 2021-11-12 四川大学 一种基于红外和可见光图像的真实人脸检测方法及装置
CN108694389A (zh) * 2018-05-15 2018-10-23 Oppo广东移动通信有限公司 基于前置双摄像头的安全验证方法及电子设备
CN108776786A (zh) * 2018-06-04 2018-11-09 北京京东金融科技控股有限公司 用于生成用户真伪识别模型的方法和装置
CN109002786B (zh) * 2018-07-06 2022-02-18 深圳市富士智能系统有限公司 人脸检测方法、设备及计算机可读存储介质
CN109035516A (zh) * 2018-07-25 2018-12-18 深圳市飞瑞斯科技有限公司 控制智能锁的方法、装置、设备及存储介质
CN109147116A (zh) * 2018-07-25 2019-01-04 深圳市飞瑞斯科技有限公司 智能锁及控制智能锁开启的方法
CN109672858A (zh) * 2018-11-23 2019-04-23 深圳奥比中光科技有限公司 3d人脸识别监控系统
CN109543635A (zh) * 2018-11-29 2019-03-29 北京旷视科技有限公司 活体检测方法、装置、系统、解锁方法、终端及存储介质
CN112906440A (zh) * 2019-12-04 2021-06-04 深圳君正时代集成电路有限公司 一种活体识别的防破解方法
CN111126216A (zh) * 2019-12-13 2020-05-08 支付宝(杭州)信息技术有限公司 风险检测方法、装置及设备
CN110979158A (zh) * 2019-12-23 2020-04-10 北京海纳川汽车部件股份有限公司 车辆及其车灯的控制方法与装置
CN111160299A (zh) * 2019-12-31 2020-05-15 上海依图网络科技有限公司 活体识别方法及装置
CN112906619A (zh) * 2021-03-09 2021-06-04 福建库克智能科技有限公司 活体检测方法、装置、电子设备及存储介质
CN113128481A (zh) * 2021-05-19 2021-07-16 济南博观智能科技有限公司 一种人脸活体检测方法、装置、设备及存储介质

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101303724A (zh) * 2007-05-10 2008-11-12 中国银联股份有限公司 一种认证授权方法及系统
CN101404060A (zh) * 2008-11-10 2009-04-08 北京航空航天大学 一种基于可见光与近红外Gabor信息融合的人脸识别方法
CN102024145A (zh) * 2010-12-01 2011-04-20 五邑大学 一种伪装人脸分层识别方法及系统
JP2011243862A (ja) * 2010-05-20 2011-12-01 Sony Corp 撮像デバイス及び撮像装置
CN104766072A (zh) * 2015-04-29 2015-07-08 深圳市保千里电子有限公司 一种活体人脸识别的装置及其使用方法
CN105912908A (zh) * 2016-04-14 2016-08-31 苏州优化智能科技有限公司 基于红外的真人活体身份验证方法

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101964056B (zh) * 2010-10-26 2012-06-27 徐勇 一种具有活体检测功能的双模态人脸认证方法和系统
CN102708383B (zh) * 2012-05-21 2014-11-26 广州像素数据技术开发有限公司 一种多模态比对功能的活体人脸检测系统与方法
TW201401186A (zh) * 2012-06-25 2014-01-01 Psp Security Co Ltd 人臉判斷系統以及方法
CN105376469B (zh) * 2014-11-03 2018-11-13 苏州思源科安信息技术有限公司 一种用于生物特征识别移动终端的驱动自动聚焦方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101303724A (zh) * 2007-05-10 2008-11-12 中国银联股份有限公司 一种认证授权方法及系统
CN101404060A (zh) * 2008-11-10 2009-04-08 北京航空航天大学 一种基于可见光与近红外Gabor信息融合的人脸识别方法
JP2011243862A (ja) * 2010-05-20 2011-12-01 Sony Corp 撮像デバイス及び撮像装置
CN102024145A (zh) * 2010-12-01 2011-04-20 五邑大学 一种伪装人脸分层识别方法及系统
CN104766072A (zh) * 2015-04-29 2015-07-08 深圳市保千里电子有限公司 一种活体人脸识别的装置及其使用方法
CN105912908A (zh) * 2016-04-14 2016-08-31 苏州优化智能科技有限公司 基于红外的真人活体身份验证方法

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110008878A (zh) * 2019-03-27 2019-07-12 中控智慧科技股份有限公司 一种人脸检测的防假方法及具备防假功能的人脸识别装置
CN110127468B (zh) * 2019-04-12 2023-02-07 深圳壹账通智能科技有限公司 电梯控制方法、装置、计算机可读存储介质和计算机设备
CN110127468A (zh) * 2019-04-12 2019-08-16 深圳壹账通智能科技有限公司 电梯控制方法、装置、计算机可读存储介质和计算机设备
CN111191519A (zh) * 2019-12-09 2020-05-22 同济大学 一种用于移动供电装置用户接入的活体检测方法
CN111191519B (zh) * 2019-12-09 2023-11-24 同济大学 一种用于移动供电装置用户接入的活体检测方法
CN111209870A (zh) * 2020-01-09 2020-05-29 杭州涂鸦信息技术有限公司 一种双目活体摄像头快速配准方法及其系统和装置
CN111414831A (zh) * 2020-03-13 2020-07-14 深圳市商汤科技有限公司 监测方法及系统、电子设备和存储介质
CN111814659B (zh) * 2020-07-07 2024-03-29 杭州海康威视数字技术股份有限公司 一种活体检测方法、和系统
CN111814659A (zh) * 2020-07-07 2020-10-23 杭州海康威视数字技术股份有限公司 一种活体检测方法、和系统
CN114827435A (zh) * 2021-01-28 2022-07-29 深圳绿米联创科技有限公司 免IR-Cut的视频流处理方法及装置、智能门锁及介质
CN114827435B (zh) * 2021-01-28 2024-05-07 深圳绿米联创科技有限公司 免IR-Cut的视频流处理方法及装置、智能门锁及介质
CN113496209A (zh) * 2021-06-09 2021-10-12 湖南中惠旅智能科技有限责任公司 基于人脸识别的景区大门闸机的数据处理方法及系统
CN113496209B (zh) * 2021-06-09 2024-03-29 湖南中惠旅智能科技有限责任公司 基于人脸识别的景区大门闸机的数据处理方法及系统
CN115082972A (zh) * 2022-07-27 2022-09-20 山东圣点世纪科技有限公司 一种基于纹理rgb图像和静脉灰度图像的活体检测方法
CN115082972B (zh) * 2022-07-27 2022-11-22 山东圣点世纪科技有限公司 一种基于rgb图像和静脉灰度图像的活体检测方法
CN117994865A (zh) * 2024-04-01 2024-05-07 杭州海康威视数字技术股份有限公司 一种双目面部匹配方法、装置、电子设备及存储介质

Also Published As

Publication number Publication date
CN106372601B (zh) 2020-12-22
PH12018500689A1 (en) 2018-10-15
CN106372601A (zh) 2017-02-01
SG11201802652YA (en) 2018-04-27

Similar Documents

Publication Publication Date Title
WO2018040307A1 (zh) 一种基于红外可见双目图像的活体检测方法及装置
US10762334B2 (en) System and method for entity recognition
CN106778518B (zh) 一种人脸活体检测方法及装置
CN108764071B (zh) 一种基于红外和可见光图像的真实人脸检测方法及装置
CN104834901B (zh) 一种基于双目立体视觉的人脸检测方法、装置及系统
US20170337441A1 (en) Systems And Methods Of Biometric Analysis With A Spectral Discriminator
JP5076563B2 (ja) 顔照合装置
KR101286454B1 (ko) 눈 영상의 특성을 이용한 모조얼굴 식별장치 및 방법
WO2019056988A1 (zh) 人脸识别方法及装置、计算机设备
WO2016084072A1 (en) Anti-spoofing system and methods useful in conjunction therewith
BR112017004427B1 (pt) método para análise de vivacidade
KR102317180B1 (ko) 3차원 깊이정보 및 적외선정보에 기반하여 생체여부의 확인을 행하는 얼굴인식 장치 및 방법
CN111382592B (zh) 活体检测方法和设备
CN108140114A (zh) 虹膜识别
CN109977846B (zh) 一种基于近红外单目摄像的活体检测方法及系统
US20210256244A1 (en) Method for authentication or identification of an individual
KR101919090B1 (ko) 3차원 깊이정보 및 적외선정보에 기반하여 생체여부의 확인을 행하는 얼굴인식 장치 및 방법
CN111178233A (zh) 基于活体认证的身份认证方法及装置
CN112633217A (zh) 基于三维眼球模型计算视线方向的人脸识别活体检测方法
Tian et al. Face anti-spoofing by learning polarization cues in a real-world scenario
Peng et al. Face liveness detection for combating the spoofing attack in face recognition
CN108363944A (zh) 人脸识别终端双摄防伪方法、装置及系统
JP7264308B2 (ja) 二次元顔画像の2つ以上の入力に基づいて三次元顔モデルを適応的に構築するためのシステムおよび方法
CN108140113A (zh) 虹膜识别
KR20090132838A (ko) 사진 촬영 장치

Legal Events

Date Code Title Description
WWE Wipo information: entry into national phase

Ref document number: 12018500689

Country of ref document: PH

WWE Wipo information: entry into national phase

Ref document number: 11201802652Y

Country of ref document: SG

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16914869

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16914869

Country of ref document: EP

Kind code of ref document: A1