CN103106385B - Face recognition method and apparatus - Google Patents

Face recognition method and apparatus Download PDF

Info

Publication number
CN103106385B
CN103106385B CN201110355062.0A CN201110355062A CN103106385B CN 103106385 B CN103106385 B CN 103106385B CN 201110355062 A CN201110355062 A CN 201110355062A CN 103106385 B CN103106385 B CN 103106385B
Authority
CN
China
Prior art keywords
component
information
processor
statistics
image
Prior art date
Application number
CN201110355062.0A
Other languages
Chinese (zh)
Other versions
CN103106385A (en
Inventor
古人豪
黄昱豪
陈昱翰
高铭璨
黄森煌
Original Assignee
原相科技股份有限公司
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 原相科技股份有限公司 filed Critical 原相科技股份有限公司
Priority to CN201110355062.0A priority Critical patent/CN103106385B/en
Publication of CN103106385A publication Critical patent/CN103106385A/en
Application granted granted Critical
Publication of CN103106385B publication Critical patent/CN103106385B/en

Links

Abstract

本发明公开了一种人脸影像辨识装置及方法。 The present invention discloses a facial image recognition apparatus and method. 该人脸影像辨识装置的处理器对复数张人脸影像的每一张计算一红色成分统计信息、一绿色成分统计信息及一蓝色成分统计信息。 The facial image recognition processor device, a green component and a blue component statistics statistics statistical information for each component of a complex calculation of a red image of human faces. 该处理器再利用一独立成分分析法对该等红色、该等绿色及该等蓝色成分统计信息中的至少二种颜色成分统计信息进行处理,以取得一第一成分信息及一第二成分信息。 The processor then using a separate red component analysis like this, at least two color components and statistics of such green blue component of such statistical information is processed to obtain information of a first component and a second component information. 该处理器再分别将该第一成分信息及该第二成分信息转换至一频率域,以取得一第一频域信息及一第二频域信息。 The processor then first information component and said second component information is converted to a frequency domain to obtain frequency-domain information of a first and a second respectively the frequency information. 该处理器再计算该第一频域信息于一预设频率范围的一能量值,并根据此能量值决定该等人脸影像撷取自一真人或一假人。 The processor then calculates the first frequency information on an energy value of a predetermined frequency range, and based on this energy value is determined from such a facial image capturing a live or dummy.

Description

人脸影像辨识方法及装置 Face recognition method and apparatus

技术领域 FIELD

[0001 ]本发明系关于一种人脸影像辨识方法及装置;详细而言,系关于一种利用独立成分分析法的人脸影像辨识方法及装置。 [0001] The present invention relates to a facial image recognition method and apparatus; Specifically, based on independent component analysis method utilizing a facial image recognition method and apparatus.

背景技术 Background technique

[0002]近年来,由于多媒体技术快速地发展,人脸辨识技术已被广泛地使用于不同的应用领域,例如电脑游戏、监控系统等等。 [0002] In recent years, due to the rapid development of multimedia technology, face recognition technology has been widely used in various applications, such as computer games, monitoring system and so on. 人脸辨识技术中,除了辨识影像内是否含有人脸之夕卜,更重要的则是判断影像中的人脸系撷取自“真人”或“假人”。 Face recognition technology, in addition to recognizing the image if it contains within Xi Bu face, the more important it is to determine the image of a human face lines retrieved from the "real" or "dummies."

[0003]习知技术系利用复数张人脸影像来判断其内的人脸为“真人”或“假人”。 [0003] The conventional technology takes advantage of the image to a plurality of human faces an inner face judge as "live" or "dummy." 具体而言,习知技术先判断这些人脸影像中是否含有人类的眼睛,再由这些人脸影像判断是否有“眨眼”的动作。 Specifically, these conventional techniques to determine whether the facial image contained in the human eye, and then from these facial images to determine whether there is "blinking" in action. 倘若判断的结果为“有眨眼的动作”,则表示影像中的人脸影像为真人的人脸影像;倘若判断的结果为“无眨眼的动作”,则表示影像中的人脸影像为假人的人脸影像。 If the determination result is "blinking operation has", it means that the facial image is a facial image of the real image; if the determination result is "no blink operation", the image of the facial image is represented as dummies the facial image.

[0004]由于习知技术需判断人脸影像中是否含有人类的眼睛,因此需要高解析度的影像方能获得较为正确的结果。 [0004] Since the conventional technology needs determines whether the facial image containing the human eye, a high-resolution image is required in order to obtain a more correct result. 当解析度低时,则判断错误的机率会大幅提高。 When the low-resolution, it is determined that the error probability will be greatly improved.

[0005]有鉴于此,本领域仍亟需一种能不受影像解析度影响而能判断人脸影像系撷取自真人或假人的技术。 [0005] Accordingly, the present art is still a need for an impact-resolution image can not be able to determine from the captured facial image based real or dummy technique.

发明内容 SUMMARY

[0006]为解决前述问题,本发明提供了一种人脸影像辨识方法及一种人脸影像辨识装置。 [0006] In order to solve the aforementioned problems, the present invention provides a facial image recognition method and a face image recognition apparatus.

[0007]本发明所提供的人脸影像辨识方法,其系适用于一电子装置。 [0007] The face image recognition method of the present invention is provided, which system is applicable to an electronic device. 该电子装置包含一处理器及一存储器单元。 The electronic device comprises a processor and a memory cell. 该存储器单元储存复数张人脸影像。 The memory cell storing a plurality of sheets of face images. 该人脸影像辨识方法包含下列步骤:(a)使该处理器对各该人脸影像计算一红色成分统计信息、一绿色成分统计信息及一蓝色成分统计信息;(b)使该处理器利用一独立成分分析法(Independent ComponentAnalysis)对该等红色成分统计信息、该等绿色成分统计信息及该等蓝色成分统计信息中的至少二种颜色成分统计信息进行处理,藉此取得一第一成分信息及一第二成分信息;(C)使该处理器分别将该第一成分信息及该第二成分信息转换至一频率域,藉此取得一第一频域信息及一第二频域信息;(d)使该处理器计算该第一频域信息于一预设频率范围的一能量值;以及(e)使该处理器将该能量值与一预设值进行比较,藉此决定该等人脸影像撷取自一真人或一假人。 The face image recognition method comprising the steps of: (a) The processor calculates a red component of the statistics for each facial image, a green component and a blue component statistics of the statistics; (b) The processor the use of an independent component analysis (independent ComponentAnalysis) such as the red component statistics, at least two kinds of color components of the green component of these statistics and the statistics of these statistics in the blue component for processing, thereby obtaining a first component information and a second information component; (C) the information processor of the first component and the second component information is converted to a frequency domain, thereby obtaining a first frequency information and the frequency domain, respectively, a second information; (d) the processor calculates the information in a first frequency domain energy value of a predetermined frequency range; and (e) the processor compares the energy value to a predetermined value, and decide such facial image captured from a real person or a dummy.

[0008]本发明所提供的人脸影像辨识装置,其系包含一存储器单元及一处理器。 [0008] The facial image recognition apparatus provided by the present invention, which system comprises a memory unit and a processor. 该存储器单元储存复数张人脸影像,而该处理器电性连接至该存储器单元。 The memory cell storing a plurality of faces in the image, and the processor is electrically connected to the memory cell. 该处理器对各该人脸影像计算一红色成分统计信息、一绿色成分统计信息及一蓝色成分统计信息。 The processor calculates a red component statistics for each of the facial image, a green component and a blue component statistics statistics. 该处理器再利用一独立成分分析法对该等红色成分统计信息、该等绿色成分统计信息及该等蓝色成分统计信息中的至少二种颜色成分统计信息进行处理,藉此取得一第一成分信息及一第二成分信息。 The processor then using statistical information of the red component a of the independent component analysis and the like, at least two color components of a green component of these statistics and statistical information in the statistical information of the blue component of such processed, thereby obtaining a first information component and a second component information. 该处理器更分别将该第一成分信息及该第二成分信息转换至一频率域,藉此取得一第一频域信息及一第二频域信息。 The processor further information on the respective first component and the second component information is converted to a frequency domain, thereby obtaining a first frequency information and a second frequency information. 该处理器更计算该第一频域信息于一预设频率范围的一能量值,且将该能量值与一预设值进行比较,藉此决定该等人脸影像撷取自一真人或一假人。 The processor further calculating a first frequency domain information to a predetermined range of frequencies of an energy value and the energy value is compared with a predetermined value, thereby determining on the facial image captured from a real or a dummy.

[0009]本发明所提供的人脸影像辨识方法及人脸影像辨识装置系利用独立成分分析法对复数张人脸影像的不同颜色成分的统计信息进行分析,藉此获取第一成分信息及第二成分信息。 [0009] The present invention provides a face recognition method and face image recognition device using independent component analysis based on the statistics of the different color component images of a plurality of human faces are analyzed, thereby obtaining information of a first and second component two component information. 之后再将第一成分信息及第二成分信息转换至频率域以分析该二成分信息于频率域上的能量值,再根据能量值判断该等人脸影像系撷取自真人或假人。 After then convert the first information and the second component to the component information in the frequency domain to the two-component analysis of the energy value of information on the frequency domain, and then determines the value of such a face image captured from the real or dummy lines of the energy. 简言之,本发明系对人脸影像中的颜色成分的统计信息进行分析再为后续的处理及判断。 Briefly, the present invention is based on the statistical information of the color component image of a human face for subsequent re-analysis and determination process. 从各张人脸影像的角度观之,由于各颜色成分系来自人脸影像,而非影像中的一小部份,故所涵盖的信息较为准确,不会因为影像的解析度不佳而影响到后续的处理及判断结果。 View from the perspective of each of the faces in the image, since each color component lines from facial images, rather than a small part of the image, so the information covered more accurate, not because of poor resolution images and affect and subsequent processing to the determination result.

[0010]为让本发明的上述目的、技术特征和优点能更明显易懂,下文系以较佳实施例配合所附图式进行详细说明。 [0010] In order to make the above-mentioned object, features and advantages of the present invention can be more fully understood, the preferred embodiments below based accompanied with figures are described in detail.

附图说明 BRIEF DESCRIPTION

[0011 ]图1系描绘第一实施例的人脸影像辨识装置的内部元件示意图; [0011] Figure 1 is depicting the internal components of the first embodiment of the human facial image recognition apparatus schematic;

[0012]图2A及图2B系描绘第三实施例的人脸影像辨识方法的流程图; [0012] FIGS. 2A and 2B is a flowchart of human recognition method of the third embodiment which is depicted in the face image;

[0013]图2C系描绘第三实施例的步骤21的细部流程图;以及 [0013] FIG. 2C detailed flowchart of step lines 21 depicting a third embodiment; and

[0014]图3系描绘第四实施例的人脸影像辨识方法的流程图。 [0014] The flowchart of FIG 3 based person identification method of the fourth embodiment of the facial image is depicted.

[0015]主要元件符号说明: [0015] Description of Symbols principal elements:

[0016] I人脸影像辨识装置 [0016] I facial image recognition apparatus

[0017] 11影像感测器 [0017] The image sensor 11

[0018] 13处理器 [0018] Processor 13

[0019] 15存储器单元 [0019] 15 memory unit

[0020] 17接收接口 [0020] The reception interface 17

具体实施方式 Detailed ways

[0021]以下将透过实施例来解释本发明所提供的人脸影像辨识装置及人脸影像辨识方法。 [0021] The following will explain the present invention to provide a face image recognition apparatus and face image recognition method through embodiments. 然而,本发明的实施例并非用以限制本发明需在如实施例所述的任何环境、应用或方式方能实施。 However, embodiments of the present invention is not intended to limit in any environment for an embodiment in order, applications, or as described in the embodiment of the present invention. 因此,关于实施例的说明仅为阐释本发明的目的,而非用以直接限制本发明。 Accordingly, the description of these embodiments is only for purpose of illustration of the present invention, but not to limit the present invention. 需说明者,以下实施例及图示中,与本发明非直接相关的元件已省略而未绘示。 It should be noted that in the following embodiments and the attached drawings, has been omitted from depiction not directly related to elements of the invention.

[0022]本发明的第一实施例为一人脸影像辨识装置I,其内部元件示意图如图1所示。 First Embodiment [0022] The present invention is a face image recognition apparatus I, its internal element is shown in Fig.1. 人脸影像辨识装置I包含一影像感测器11、一处理器13、一存储器单元15及一接收接口17,其中处理器13电性连接至影像感测器11、存储器单元15及接收接口17。 I face image recognition apparatus comprises an image sensor 11, a processor 13, a memory unit 15, and a receiving interface 17, processor 13 is electrically connected to the image sensor 11, a memory unit 15, and a reception interface 17 . 影像感测器11可为视讯摄影机(Webcam)或所属技术领域中具有通常知识者可轻易思及的其他具有影像撷取能力的装置。 The image sensor 11 may have other devices with image capture capability ordinary knowledge and can easily think of a video camera (Webcam) or techniques in the art. 处理器13可为本发明所属技术领域中具有通常知识者所熟知的各种处理器、中央处理装置(central processing unit)、微处理器或其它计算装置其中的任一种。 The processor 13 may be oriented in the art a variety of processors having ordinary skills in the art of the present invention, a central processing unit (central processing unit), a microprocessor, or any other computing apparatus wherein. 存储器单元15可为存储器、软碟、硬碟、光碟、随身碟、磁带、可由网路存取的资料库或所属技术领域中具有通常知识者可轻易思及具有相同功能的储存媒体。 The memory unit 15 may, or may skilled library access network having ordinary skill may readily think and storage media having the same functions as a memory, floppy disk, hard disk, CD, flash drive, a magnetic tape.

[0023]首先,影像感测器11于一时间区间内撷取复数张影像。 [0023] First, the image sensor 11 capture a plurality of files in the time interval. 针对影像感测器11所撷取的各张影像,处理器13会进行人脸侦测。 Zhang for each image captured by the image sensor 11, the processor 13 for face detection. 若侦测的结果为影像中含有人脸,则处理器13会视该张影像为人脸影像,并将的储存于存储器单元15;若侦测的结果为影像中不含有人脸,则处理器13会舍弃该张影像。 If the detection result of the image containing the human face, then the processor 13 will be a human face image depending on the image sheet, and is stored in the memory unit 15; if the detection result of the face image was free, the processor 13 will discard the images in succession. 处理器13对影像进行人脸侦测的技术为本发明所属技术领域的通常知识,且并非本发明的重点,因此不赘述其细节。 The video processor 13 for face detection technology generally knowledge of the art of the present invention and are not essential to the invention, and therefore do not repeat the details. 兹假设经影像感测器11及处理器13进行前述处理后,已储存了复数张人脸影像于存储器单元15中。 After hereby assumed by the image sensor 11 and a processor 13 for the processing of a plurality of sheets stored in the face image storage unit 15.

[0024]须说明者,依本发明所实施的其他人脸影像辨识装置可不配置影像感测器U。 [0024] It shall be, under this other facial image recognition apparatus of the embodiment of the present invention may configure the image sensor U. 于此情形下,使用者可先将复数张人脸影像储存于存储器单元即可。 Under this scenario, the user may first plurality of faces in the image can be stored in the memory unit.

[0025]接着,处理器13对各张人脸影像计算一红色成分统计信息、一绿色成分统计信息及一蓝色成分统计信息。 [0025] Next, the processor 13 calculates a red component statistics for each of the faces in the image, a green component and a blue component statistics statistics. 兹以一范例说明红色成分统计信息、绿色成分统计信息及蓝色成分统计信息的计算方式。 Have a paradigm to explain the red component statistics, calculated statistics information of the green component and the blue component statistics. 举例而言,处理器13可先分离各张人脸影像的红色成分、绿色成分及蓝色成分,于处理完所有的人脸影像后,便得到复数张红色成分影像、复数张绿色成分影像及复数张蓝色成分影像。 For example, the processor 13 can first separate the sheets of red component, green component and blue component facial images, to after processing all the facial image, the red component will obtain a plurality of sheets of images, video and plural sheets of green component plural sheets of blue component images.

[0026]处理器13接着可对各该红色成分影像进行计算,以得一红色成分信息,当处理完所有的红色成分影像后,便获得复数个红色成分信息。 [0026] The processor 13 may then be calculated for each of the red component image, to obtain a red component information, when processed all of the red component image, then obtain a plurality of red component information. 类似的,处理器13可对各该绿色成分影像进行计算,以得一绿色成分信息,当处理完所有的绿色成分影像后,便获得复数个绿色成分信息。 Similarly, the processor 13 may be made to each of the green component image is calculated, to obtain a green component information, when processed all green component of the video, it is obtained a plurality of green component information. 类似的,处理器13可对各该蓝色成分影像进行计算,以得一蓝色成分信息,当处理完所有的蓝色成分影像后,便获得复数个蓝色成分信息。 Similarly, the processor 13 may be made to each of the blue component image is calculated, to give a blue color component information, when processed all of the blue component images, it is obtained a plurality of blue component information.

[0027]举例而言,各红色成分信息可为相对应的红色成分影像的平均亮度值、最大亮度值、最小亮度值或其他统计值;各绿色成分信息可为相对应的绿色成分影像的平均亮度值、最大亮度值、最小亮度值或其他统计值;各蓝色成分信息可为相对应的蓝色成分影像的平均亮度值、最大亮度值、最小亮度值或其他统计值。 [0027] For example, each of the red component information may be an average luminance value of the red component of the image corresponding to the maximum luminance value, minimum luminance value or other statistical value; each of the green component may be a green component image information corresponding to an average luminance value, maximum luminance value, minimum luminance value or other statistical value; each of the blue color component information may be an average luminance value of the blue component of the image corresponding to the maximum luminance value, minimum luminance value or other statistical value.

[0028]于取得该等红色成分信息、该等绿色成分信息及该等蓝色成分信息后,处理器13选取其中二种颜色成分信息进行后续处理,假设处理器13选取该等红色成分信息及该等绿色成分信息。 After [0028] To obtain such information the red component, the green component of such information and the blue component of such information, the processor 13 select one of two kinds of color component information for subsequent processing, the processor 13 is assumed to select the red component of such information and such green component information. 之后,处理器13利用独立成分分析法(Independent Component Analysis)对该等红色成分信息及该等绿色成分信息进行处理,处理之后便得到一第一成分信息及一第二成分信息。 Thereafter, the processor 13 using independent component analysis (Independent Component Analysis) to the Red component information and the like such information is processed green components, it will have a first component and a second component information after the information processing. 须说明者,独立成分分析法为一既有的演算法,当有二组输入资料时,独立成分分析法会产生二组输出资料,此为本发明所属技术领域中的通常知识者。 It shall be independent component analysis algorithm to an existing, when there are two set of input data, independent component analysis produces two sets of output data, the ordinary skills in the art of the present invention. 另外,第一成分信息及第二成分信息的“第一”及“第二”仅用来区分第一成分信息及第二成分信息为不同的成分信息而已。 Further, the first component and the second component information of the information as "first" and "second" are only used to distinguish a first information component and a second component different from component information is information only.

[0029]采用独立成分分析法的目的在于分离该等红色成分信息及该等绿色成分信息中的人类心跳的信号与其他信号。 [0029] The object of the independent component analysis that the component information is separated and the red and green components of such a human heartbeat information signal and other signals. 换言之,第一成分信息及第二成分信息其中之一可视为人类心跳的信号。 In other words, one of the first component and the second component information in which the information can be regarded as a signal of a human heartbeat. 人类心跳的信号在频率域上有特殊的特性,亦即,于特定频率范围内的信号能量较强。 Human heartbeat signals have special characteristics in the frequency domain, i.e., signal energy within a certain frequency range strong. 本实施例将此特定频率范围设为预设频率范围,以便进行后续的处理。 This particular embodiment of the present embodiment is set to a predetermined frequency range of the frequency range, for subsequent processing. 一般而言,大部分人类心跳的信号的频率较低,故此预设频率范围可设在低频之处。 In general, most of the lower frequency signal of the human heart, therefore preset frequency range may be provided at the low frequencies.

[0030]具体而言,处理器13接着会将第一成分信息及第二成分信息分别转换至一频率域,藉此取得一第一频域信息及一第二频域信息。 [0030] Specifically, the processor 13 will then first component information and the second information are respectively converted to a component in the frequency domain, whereby the domain information to obtain a first frequency and a second frequency information. 举例而言,处理器13可采用快速傅里叶转换(Fast Fourier Transform)以将第一成分信息及第二成分信息分别转换为第一频域信息及第二频域信息。 For example, processor 13 may employ Fast Fourier Transform (Fast Fourier Transform) component to the first information and the second component are the first information into frequency domain information and the second frequency information. 依本发明所实施的其他人脸影像辨识装置亦可采用其他的频域转换方式。 Under this other face image recognition apparatus of the embodiment of the present invention may also use other frequency domain conversion.

[0031]之后,处理器13计算第一频域信息于预设频率范围的一能量值,再将此能量值与一预设值进行比较。 After [0031], the processor 13 calculates the first frequency information in a predetermined frequency range energy values, then this energy value and a preset value. 倘若第一频域信息于预设频率范围的能量值大于预设值,则存储器单元15内所储存的人脸影像撷取自一真人。 If the face image information to a preset first frequency domain energy value of the frequency range is greater than the preset value, then the memory cell 15 is retrieved from a stored real. 倘若第一频域信息于预设频率范围的能量值小于预设值,则处理器13再计算第二频域信息于预设频率范围的一能量值,再将此能量值与预设值进行比较。 If the first frequency domain information to a preset frequency range energy values ​​less than the preset value, the processor 13 then calculates second frequency information in a predetermined frequency range energy values, then this energy value with a preset value Compare. 倘若第二频域信息于预设频率范围的能量值大于预设值,则存储器单元15内所储存的人脸影像亦可被认定为撷取自一真人。 If the face image information to a preset second frequency domain energy value of the frequency range is greater than the preset value, then the memory cell 15 may also be stored retrieved from an identified as live. 倘若第一频域信息于预设频率范围的能量值小于预设值,且第二频域信息于预设频率范围的能量值亦小于预设值,则存储器单元 If the first frequency domain information to a preset frequency range lower than the predetermined energy value, and the second frequency domain information to a preset frequency range is also lower than the predetermined energy value, then the memory cell

15内所储存的人脸影像会被认定为撷取自一假人。 15 facial images stored is identified as retrieved from a dummy. 简言之,只要第一频域信息或第二频域信息其中的一的能量值大于预设值,则存储器单元15内所储存的人脸影像会被认定为撷取自一真人。 Briefly face image, as long as the first or the second frequency domain information of a frequency domain information wherein an energy value greater than a preset value, stored in the memory unit 15 will be retrieved from an identified as live.

[0032]倘若经前述处理后,处理器13决定该等人脸影像系撷取自一真人,则人脸影像辨识装置I会继续进行登入程序。 [0032] If after the foregoing processing, the processor 13 determine any face image retrieved from a real system, the facial image recognition apparatus I will continue the login process. 具体而言,接收接口17会接收一登入信息,而处理器13则会进一步地处理此登入信息。 Specifically, a receiving interface 17 receives the login information, the processor 13 will further process this log-in information. 举例而言,假设登入信息包含使用者的帐号及密码,处理器13便会判断此帐号及密码是否正确。 For example, assuming login information includes user account and password, the processor 13 determines this would account and password are correct. 然而,倘若经前述处理后,处理器13决定该等人脸影像系撷取自一假人,则人脸影像辨识装置I不会进行登入程序。 However, if after the aforementioned processing, the processor 13 determine any face image retrieved from a dummy line, the facial image recognition apparatus does not perform I login process.

[0033]本发明的第二实施例亦为人脸影像辨识装置I,然其详细运作与第一实施例稍有不同。 [0033] The second embodiment of the present invention is also the face image recognition apparatus I, and then the detailed operation of the first embodiment is slightly different. 以下将仅说明第二实施例与第一实施例的相异处。 The following will only be explained the second embodiment is different from the first embodiment.

[0034]本实施例中,当处理器13取得该等红色成分信息、该等绿色成分信息及该等蓝色成分信息后,处理器13选取全部的颜色成分信息进行后续处理。 [0034] In this embodiment, when the processor 13 to acquire such information of the red component, the green component of such information and the blue component of such information, the processor 13 select all of the color component information for subsequent processing. 之后,处理器13利用独立成分分析法对该等红色成分信息、该等绿色成分信息及该等蓝色成分信息进行处理,处理之后便得到一第一成分信息、一第二成分信息及一第三成分信息。 Thereafter, the processor 13 uses the information of the red component of the independent component analysis and the like, such information green component and the blue component of such processing information, the information will be obtained after processing a first component, a second component and a second information three-component information. 须说明者,独立成分分析法为一既有的演算法,当有三组输入资料时,独立成分分析法会产生三组输出资料,此为本发明所属技术领域中的通常知识者。 It shall be independent component analysis algorithm to an existing, when there are three sets of input data, independent component analysis produces three output data, the ordinary skills in the art of the present invention. 另外,第一成分信息、第二成分信息及第三成分信息的“第一”、“第二”及“第三”仅用来区分第一成分信息、第二成分信息及第三成分信息为不同的成分信息而已。 Further, the first information component, the second component and the third component information of the information "first", "second" and "third," information is only used to distinguish a first component, a second component and the third component information is information different components of information only.

[0035]由于本实施的处理器13选取三种颜色成分信息进行后续处理,因此独立成分分析法会分离该等红色成分信息、该等绿色成分信息及该等蓝色成分信息中的人类心跳的信号、人类移动(或晃动)的信号与其他信号。 [0035] Since the present embodiment of the processor 13 to select the three color component information for subsequent processing, independent component analysis would thus separated and the information of the red component, the green component of such information and the information of the blue component of such a human heartbeat signal, a human moving (or vibrating) signal and other signals. 换言之,第一成分信息、第二成分信息及第三成分信息其中的一可视为人类心跳的信号。 In other words, information of a first component, a second component and a third component information is information which may be considered as a signal of a human heartbeat.

[0036]处理器13接着会将第一成分信息、第二成分信息及第三成分信息分别转换至一频率域,藉此取得一第一频域信息、一第二频域信息及一第三频域信息。 [0036] The processor 13 will then information of a first component, a second component and a third component message information are converted to a frequency domain, thereby obtaining a first frequency information, a second and a third frequency-domain information frequency-domain information. 之后,处理器13计算第一频域信息于预设频率范围的一能量值,第二频域信息于预设频率范围的一能量值,及第三频域信息于预设频率范围的一能量值。 Thereafter, the processor 13 calculates the first frequency information in a predetermined frequency range energy values, the second frequency domain information in a predetermined frequency range energy values, and the third frequency information in a predetermined frequency range of the energy value. 只要第一频域信息、第二频域信息或第三频域信息其中的一的能量值大于预设值,则存储器单元15内所储存的人脸影像会被认定为撷取自一真人。 As long as the first frequency information, the second information or the third frequency domain where the frequency domain information is greater than the energy value of a predetermined value, the facial image stored in the memory unit 15 is retrieved from an identified as live.

[0037]除了上述步骤,第二实施例亦能执行第一实施例所描述的所有操作及功能,所属技术领域具有通常知识者可直接了解第二实施例如何基于上述第一实施例以执行此等操作及功能,故不赘述。 [0037] In addition to the above-described steps, the second embodiment can also execute all the operations and functions of the first embodiment described, the art having ordinary knowledge can be readily appreciated how the second embodiment based on the first embodiment to perform this and other operations and functions, it is not repeated herein.

[0038]本发明的第三实施例为一人脸影像辨识方法,其流程图如图2A及图2B所示。 [0038] The third embodiment of the present invention is a face image recognition method, a flowchart in FIG. 2A and 2B. 此人脸影像辨识方法适用于一电子装置,例如第一实施例及第二实施例的人脸影像辨识装置I。 This facial image recognition method for use in an electronic device, examples and the first embodiment, for example, the second embodiment of the face image recognition apparatus I. 此电子装置包含一处理器及一存储器单元,且此存储器单元储存复数张人脸影像。 This electronic device includes a processor and a memory cell, and the memory cell storing a plurality of sheets of face images.

[0039]首先,人脸影像辨识方法执行步骤S21,以使处理器对各该人脸影像计算一红色成分统计信息、一绿色成分统计信息及一蓝色成分统计信息。 [0039] First, the facial image recognition method executes step S21, so that the processor calculates a red component statistics for each of the facial image, a green component and a blue component statistics statistics.

[0040]举例而言,步骤S21可藉由图2C所描绘的细部流程图来完成。 [0040] For example, step S21 may be accomplished by a detailed flow chart depicted in FIG. 2C. 于步骤S211中,人脸影像辨识方法使处理器分离各张人脸影像的红色成分、绿色成分及蓝色成分,藉此得到复数张红色成分影像、复数张绿色成分影像及复数张蓝色成分影像。 In step S211, the facial image recognition method of the red component processor human faces separated image, a green component and the blue component, red component thereby to obtain a plurality of sheets of images, a plurality of sheets of a plurality of sheets of a green component and blue component images image. 接着,步骤S212使处理器对各该红色成分影像计算,以得该等红色成分统计信息。 Next, step S212 causes the processor to calculate each of the red component image, in order to get the red component of such statistics. 步骤S213使处理器对各该绿色成分影像计算,以得该等绿色成分统计信息。 Step S213 causes the processor of the green component is calculated for each image, to obtain a green component of such statistics. 步骤S214使处理器对各该蓝色成分影像计算,以得该等蓝色成分统计信息。 Step S214 causes the processor to calculate each of the blue component image to obtain the blue component of such statistics. 须说明者,本实施例并未限制步骤S212、S213及S214的执行顺序。 It shall be appreciated that the present embodiment does not limit the order of execution of steps S212, S213 and S214 of. 再者,各该红色成分统计信息可为相对应的该红色成分影像的平均亮度值、最大亮度值、最小亮度值或其他统计值,各该绿色成分统计信息为相对应的该绿色成分影像的平均亮度值、最大亮度值、最小亮度值或其他统计值,且各该蓝色成分统计信息为相对应的该蓝色成分影像的平均亮度值、最大亮度值、最小亮度值或其他统计值。 Further, each of the red component may be statistical information corresponding to the average luminance value of the red component image, the maximum luminance value, minimum luminance value or other statistical values, each of the green component statistics for the green component of the image corresponding to the average luminance value, maximum luminance value, minimum luminance value or other statistical value, and each of the blue component statistics for the average luminance value of the blue component of the image corresponding to the maximum luminance value, minimum luminance value or other statistical value.

[0041]于步骤S21后,人脸影像辨识方法执行步骤S22,使处理器利用独立成分分析法对该等红色成分统计信息、该等绿色成分统计信息及该等蓝色成分统计信息中的至少二种颜色成分统计信息进行处理,藉此取得一第一成分信息及一第二成分信息。 [0041] after step S21, the facial image recognition method performs the step S22, the processor utilization statistics for the red component of the independent component analysis and other statistical information such green component and blue component in these statistical information at least two kinds of color components of statistical information processing, thereby obtaining a first component and a second component information information.

[0042]接着,步骤S23使处理器分别将该第一成分信息及该第二成分信息转换至一频率域,藉此取得一第一频域信息及一第二频域信息。 [0042] Next, in step S23 that the information processor of the first component and the second component information is converted to a frequency domain, thereby obtaining a first information and a second frequency-domain information of the frequency domain, respectively. 举例而言,步骤S23可使处理器采用傅里叶转换法。 For example, step S23 using the processor can Fourier transform method.

[0043]人脸影像辨识方法之后执行步骤S24,使处理器计算第一频域信息于一预设频率范围的一能量值。 [0043] After the step S24 is executed human face image recognition method of the processor to calculate a first frequency domain information in a predetermined frequency range of energy values. 接着执行步骤25,使处理器判断第一频域信息于预设频率范围的能量值是否大于一预设值。 Then, step 25, the processor determines that the information value of the first frequency domain energy in a predetermined frequency range is greater than a predetermined value. 若步骤S25的判断结果为是,人脸影像辨识方法便执行步骤S29,使处理器判断该等人脸影像系撷取自真人。 If the result of step S25 is yes, facial recognition method will perform the step S29, the processor determines those facial image capture system from a real person. 若步骤S25的结果为否,则执行步骤S26使处理器计算第二频域信息于一预设频率范围的一能量值。 If the result is NO in step S25, step S26 is executed cause a processor to calculate second frequency information on an energy value of a predetermined frequency range. 接着步骤S27使处理器判断第二频域信息于预设频率范围的能量值是否大于预设值。 Then the processor determines in step S27 that the second frequency-domain information to a preset frequency range energy values ​​is greater than a predetermined value. 若步骤S27的判断结果为是,人脸影像辨识方法便执行步骤S29,使处理器判断该等人脸影像系撷取自真人。 If the result of step S27 is yes, facial recognition method will perform the step S29, the processor determines those facial image capture system from a real person. 若步骤S25的判断结果为否,人脸影像辨识方法便执行步骤S28,使处理器判断该等人脸影像系撷取自假人。 If the result of step S25 is NO, facial recognition method will perform the step S28, the processor determines those facial image capture system from dummy.

[0044] 须说明者,步骤S24至步骤S29亦可改由其他步骤来加以实现。 [0044] It shall be appreciated, step S24 to step S29 is also changed to be achieved by the other steps. 举例而言,可先计算第一频域信息于一预设频率范围的一能量值,并计算计算第二频域信息于一预设频率范围的一能量值。 For example, the first information to calculate a first frequency domain a predetermined frequency range energy values, and the calculation of the second frequency information on an energy value of a predetermined frequency range. 接着选取此二能量值较大者进行后续的比对。 Then select these two energy values ​​greater subsequent comparison. 倘若此二能量值较大者大于预设值,则处理器便能判断该等人脸影像系撷取自真人。 If these two energy values ​​greater than the preset value, the processor determines those able facial image captured from the real system. 倘若此二能量值较大者小于预设值,则处理器便能判断该等人脸影像系撷取自假人。 If these two energy values ​​greater than the preset value, the processor determines those able facial image captured from the dummy line.

[0045]倘若前述步骤S27的判断结果为是且人脸影像辨识方法因此执行步骤S29,则人脸影像辨识方法可继续进行登入程序。 [0045] If the determination result of step S27 is YES and the facial image recognition method thus performing step S29, the image recognition method of the facial login process can proceed. 具体而言,人脸影像辨识方法可进一步执行步骤S291,使电子装置的一接收接口接收一登入信息;接着,人脸影像辨识方法执行步骤S293,使处理器处理此登入信息。 Specifically, the facial image recognition method may further perform the step S291, so that a receiving electronic device interface receives a signed message; Next, face image recognition method proceeds to step S293, cause the processor to process this log-in information. 举例而言,假设登入信息包含使用者的帐号及密码,则步骤S293便使处理器判断此帐号及密码是否正确。 For example, assuming the login information includes user ID and password, step S293 it is determined that the processor of this ID and password are correct. 须说明者,倘若前述步骤S27的判断结果为否且人脸影像辨识方法因此执行步骤S28,则人脸影像辨识方法不会进行登入程序。 It shall be appreciated, if the determination result is NO in step S27 and the facial image recognition method thus performing step S28, the recognition method of the facial image is not the login process.

[0046]除了上述步骤,第三实施例亦能执行第一实施例所描述的所有操作及功能,所属技术领域具有通常知识者可直接了解第三实施例如何基于上述第一实施例以执行此等操作及功能,故不赘述。 [0046] In addition to the above steps, the third embodiment can also execute all the operations and functions of the first embodiment described, the art having ordinary knowledge can be readily appreciated how the third embodiment based on the first embodiment to perform this and other operations and functions, it is not repeated herein.

[0047]本发明的第四实施例为一人脸影像辨识方法,其流程图如图3所示。 Fourth Example [0047] The present invention is a face image recognition method, a flowchart shown in FIG. 此人脸影像辨识方法适用于一电子装置,例如第一实施例及第二实施例的人脸影像辨识装置I。 This facial image recognition method for use in an electronic device, examples and the first embodiment, for example, the second embodiment of the face image recognition apparatus I. 此电子装置包含一处理器及一存储器单元,且此存储器单元储存复数张人脸影像。 This electronic device includes a processor and a memory cell, and the memory cell storing a plurality of sheets of face images. 第四实施例的详细流程与第三实施例稍有不同。 Detailed flow of the fourth embodiment is slightly different from the third embodiment. 以下将仅详述第四实施例与第三实施例的相异处。 The following detailed description will only be at a different embodiment of the third embodiment according to the fourth embodiment.

[0048]与第三实施例相同,第四实施例的人脸影像辨识方法先执行步骤S21,使处理器对各该人脸影像计算一红色成分统计信息、一绿色成分统计信息及一蓝色成分统计信息。 [0048] The same as the third embodiment, the fourth embodiment of the people face to the image recognition method of performing step S21, the processor calculates a red component so that the statistics for each of the facial image, a green component and a blue Statistics component statistics. 步骤321亦可藉由图2(:所绘示的步骤3211、3212、3213及3214来完成。 Step 321 in FIG. 2 by also (: step depicted 3211,3212,3213 and 3214 to complete.

[0049]第四实施例接着执行步骤S32,使处理器利用独立成分分析法对该等红色成分统计信息、该等绿色成分统计信息及该等蓝色成分统计信息中的全部颜色成分统计信息进行处理,藉此取得一第一成分信息、一第二成分信息及一第三成分信息。 [0049] Next a fourth embodiment of performing step S32, the processor makes use of statistical information of the red component of the independent component analysis and the like, all the color component green component of these statistics and statistical information in the statistical information such blue component is process, thereby obtaining information of a first component, a second component and a third component message information. 之后,步骤S33使处理器分别将该第一成分信息、该第二成分信息及该第三成分信息转换至一频率域,藉此取得一第一频域信息、一第二频域信息及一第三频域信息。 After the step S33, respectively, cause the processor to the first component information, the second information and the component information is converted to a third component in the frequency domain, thereby obtaining a first frequency information, a frequency-domain information, and a second The third frequency-domain information.

[0050]接着,执行步骤S34、S35及S36,使处理器计算第一频域信息、第二频域信息及第三频域信息各自于一预设频率范围的一能量值。 [0050] Next, step S34, S35 and S36, so that the processor calculates the first frequency information, the second frequency domain information and frequency domain information of each a third energy value in a predetermined frequency range. 须说明者,第四实施例并未限制步骤S34、S35及S36的执行顺序。 It shall be appreciated, the fourth embodiment does not limit the steps S34, S35 and S36 is sequentially executed. 于步骤S37,处理器选取步骤S34、步骤S35及步骤S36的三个能量值的较大者以便进行后续的处理。 In step S37, the processor selecting step S34, step S35, and the larger step S36 is three energy values ​​for subsequent processing.

[0051]第四实施例接着执行步骤S38,判断步骤S37所选取的能量值是否大于预设值。 [0051] Next a fourth embodiment of performing step S38, step S37 determines the selected energy value is greater than a preset value. 若步骤S37的判断结果为是,则执行步骤S39,使处理器判断该等人脸影像系撷取自真人。 If the determination result of step S37 is YES, the executed step S39, the processor determines that the facial image of such line retrieved from reality. 若步骤S37的判断结果为否,人脸影像辨识方法便执行步骤S40,使处理器判断该等人脸影像系撷取自假人。 If the result of the step S37 is NO, facial recognition method will perform the step S40, the processor determines those facial image capture system from dummy.

[0052] 类似的,第四实施例的步骤S34至S40亦可改由其他步骤来加以实现。 Similar [0052], a fourth embodiment of steps S34 to S40 can also be replaced by other steps to be implemented. 举例而言,第四实施例可先计算 For example, the fourth embodiment may first calculate

[0053]第一频域信息、第二频域信息及第三频域信息其中的一于预设频率范围的能量值,并判断该能量值是否大于一预设值。 [0053] The first frequency domain information and frequency domain information of the second and the third frequency information to which a predetermined frequency range energy values, and determines whether the energy value is greater than a predetermined value. 若是,则该等人脸影像系撷取自真人,便可停止运作;若否,则再选取其他的频域信息计算能量值及后续的判断。 If so, the facial image of such lines retrieved from the real, can stop working; if not, then selecting a different frequency information calculated energy value and the subsequent judgment. 后续的处理依此类推。 Subsequent processing and so on. 倘若所有的频域信息的能量值皆不大于预设值,则该等人脸影像系撷取自假人。 If all the energy values ​​in the frequency domain information is neither greater than the preset value, then such facial image capture system from dummy.

[0054]除了上述步骤,第四实施例亦能执行第二实施例所描述的所有操作及功能,所属技术领域具有通常知识者可直接了解第四实施例如何基于上述第二实施例以执行此等操作及功能,故不赘述。 [0054] In addition to the above steps, the fourth embodiment can also execute all the operations and functions described in the second embodiment, generally having knowledge of the art will be readily understood how the fourth embodiment based on the second embodiment to perform this and other operations and functions, it is not repeated herein.

[0055]另外,第三实施例及第四实施例所描述的人脸影像辨识方法可由一电脑程式产品执行,当电子装置或人脸辨识装置载入该电脑程式产品,并执行该电脑程式产品所包含的复数个指令后,即可完成第三实施及第四实施例例所述的投影方法。 [0055] Further, the third embodiment and the fourth embodiment described embodiments human face recognition method performed by a computer program product, when the electronic apparatus or face recognition means loading the computer program product, the computer program product and executed after a plurality of instructions included in the third embodiment, and a projection method of the fourth embodiment of the Example to complete. 前述的电脑程式产品可储存于电脑可读取记录媒体中,例如唯读存储器(read only memory;ROM)、快闪存储器、软碟、硬碟、光碟、随身碟、磁带、可由网路存取的资料库或熟习此项技艺者所习知且具有相同功能的任何其它储存媒体中。 The computer program product may be stored in a computer-readable recording medium, such as read-only memory (read only memory; ROM), flash memory, floppy disk, hard disk, CD, flash drive, magnetic tape, network access repository skilled in the art and well known to those having the same function or any other storage media.

[0056]由上述第一至第四实施例的说明可知,本发明所提供的人脸影像辨识方法及人脸影像辨识装置系利用独立成分分析法对复数张人脸影像的不同颜色成分的统计信息进行分析,藉此获取至少第一成分信息及第二成分信息。 [0056] apparent from the embodiment described above the first to fourth embodiments, the present invention provides a face recognition method and face image recognition device using independent component analysis based on the statistics of different color components of a plurality of images of human faces information is analyzed, thereby obtaining at least a first component and a second information component information. 第一成分信息及第二成分信息其中的一可视为人类心跳的信号。 The first information and the second component wherein the component information may be considered as a human heartbeat signal. 本发明之后再将第一成分信息及第二成分信息转换至频率域以分析该二成分信息于频率域上的能量值,再根据能量值判断该等人脸影像系撷取自真人或假人。 After the first component of the present invention is then converted component information and the second information to the frequency domain analysis of the energy value of the two-component information on the frequency domain, and then determine those based facial image captured from the real or dummy according to energy values .

[0057]简言之,本发明系对人脸影像中的颜色成分的统计信息进行分析再为后续的处理及判断。 [0057] Briefly, the present invention is based on the statistical information of the color component image of a human face for subsequent re-analysis and determination process. 从各张人脸影像的角度观之,由于各颜色成分系来自整张影像,而非影像中的一小部份,故所涵盖的信息较为准确,不会因为影像的解析度不佳而影响到后续的处理及判断结果。 View from the perspective of each of the faces in the image, since each color component system from the entire image, rather than a small part of the image, so the information covered more accurate, not because of poor resolution images and affect and subsequent processing to the determination result.

[0058]上述的实施例仅用来举例本发明的实施态样,以及阐述本发明的技术特征,并非用来限制本发明的保护范畴。 [0058] The above examples serve embodiment aspect of the present invention, and set forth the features of the invention, not to limit the scope of the invention only. 任何熟悉此技术者可轻易完成的改变或均等性的安排均属于本发明所主张的范围,本发明的权利保护范围应以本发明权利要求范围为准。 Skilled in this field may proceed with a variety of modifications of the arrangement according to the present invention belong to the scope of the claimed, the scope of rights of the present invention should be subject to the scope of the claimed invention.

Claims (17)

1.一种人脸影像辨识方法,适用于一电子装置,所述的电子装置包含一处理器及一存储器单元,所述的存储器单元储存复数张人脸影像,其特征在于,所述的人脸影像辨识方法包含下列步骤: (a)使所述的处理器对各所述的人脸影像计算一红色成分统计信息、一绿色成分统计信息及一蓝色成分统计信息; (b)使所述的处理器利用一独立成分分析法对所述的红色成分统计信息、所述的绿色成分统计信息及所述的蓝色成分统计信息中的至少二种颜色成分统计信息进行处理,藉此取得一第一成分信息及一第二成分信息; (C)使所述的处理器分别将所述的第一成分信息及所述的第二成分信息转换至一频率域,藉此取得一第一频域信息及一第二频域信息; (d)使所述的处理器计算所述的第一频域信息于一预设频率范围的一能量值;以及(e)使所述的处理器将所述的能量 A face image recognition method for use in an electronic device, said electronic device comprising a processor and a memory cell, said memory cell storing a plurality of faces in images, wherein said human face recognition method comprising the steps of: (a) causing the processor calculates a red component of the statistical information for each of said facial image, a green component and a blue component statistics of the statistics; (b) so that the said statistical information processor using a method of independent component analysis of the red component, at least two color components of blue component statistics of the statistics and the statistical information of the green component in the processing, thereby to obtain a first information component and a second component information; (C) causing the processor to the respective said first information component and said second component information is converted to a frequency domain, thereby obtaining a first frequency domain information and frequency domain information of a second; (d) causing the processor calculating the said first frequency information on an energy value of a predetermined frequency range; and (e) cause said processor the energy 值与一预设值进行比较,藉此决定所述的人脸影像撷取自一真人或一假人。 Face image value is compared with a predetermined value, whereby said determined retrieved from a real or a dummy.
2.如权利要求1所述的人脸影像辨识方法,其特征在于,所述的步骤(a)包含下列步骤: (al)使所述的处理器分离各所述的人脸影像的红色成分、绿色成分及蓝色成分,藉此得到复数张红色成分影像、复数张绿色成分影像及复数张蓝色成分影像; (a2)使所述的处理器对各所述的红色成分影像计算,以得所述的红色成分统计信息; (a3)使所述的处理器对各所述的绿色成分影像计算,以得所述的绿色成分统计信息;以及(a4)使所述的处理器对各所述的蓝色成分影像计算,以得所述的蓝色成分统计信息。 Red component (Al) the processor cause separation of each of the facial image: 2. The face image recognition method according to claim 1, wherein said step (a) comprises the steps of , green component and blue component, red component thereby to obtain a plurality of sheets of images, a plurality of sheets of a plurality of sheets of a green component image and a blue component images; (A2) cause the processor to calculate for each image of the red component to to obtain statistical information of the red component; (A3) causes the processor to calculate the green component of each of said image, to obtain statistical information of the green component; and (a4) causing the processor to each of the a blue component image is calculated according to the statistics obtained blue component.
3.如权利要求2所述的人脸影像辨识方法,其特征在于,所述的各红色成分统计信息为相对应的所述的红色成分影像的一平均亮度值,各所述的绿色成分统计信息为相对应的所述的绿色成分影像的一平均亮度值,且各所述的蓝色成分统计信息为相对应的所述的蓝色成分影像的一平均亮度值。 3. The face image recognition method as claimed in claim 2, wherein each of the red component according to the statistics of the average luminance value of the red component of an image corresponding to said each of the green component Statistics an average luminance value of the green component of the image corresponding to the information, and each of the blue component statistics of an average brightness of a blue component of the image corresponding to said.
4.如权利要求1所述的人脸影像辨识方法,其特征在于,所述的步骤(b)系使所述的处理器利用所述的独立成分分析法对所述的红色成分统计信息、所述的绿色成分统计信息及所述的蓝色成分统计信息进行处理,以取得所述的第一成分信息、所述的第二成分信息及一第三成分信息,所述的步骤(c)更使所述的处理器将所述的第三成分信息转换至所述的频率域,藉此取得一第三频域信息。 4. The face image recognition method according to claim 1, wherein said step (b) enable the processor system using the statistical information in the independent component analysis of the red component, the green component of the statistical information and the statistical information of the blue component is processed to obtain information of said first component, said second component and a third component message information, said step (c) said processor further cause the third component information is converted to the frequency domain, thereby obtaining a third frequency information.
5.如权利要求1所述的人脸影像辨识方法,其特征在于,所述的步骤(c)系使所述的处理器分别对所述的第一成分信息及所述的第二成分信息进行一快速傅里叶转换,藉此取得所述的第一频域信息及所述的第二频域信息。 5. A face image recognition method according to claim 1, wherein said step (c) causing the processor based on information of the respective information according to a first component and said second component performing a fast Fourier transform, thereby obtaining the second frequency information of the first information and the frequency domain is.
6.如权利要求1所述的人脸影像辨识方法,其特征在于,当所述的能量值大于所述的预设值时,所述的步骤(e)决定所述的人脸影像系撷取自所述的真人。 6. The face image recognition method according to claim 1, wherein, when the energy value is greater than the predetermined value, said step (e) determining the face images based captures from the real.
7.如权利要求1所述的人脸影像辨识方法,其特征在于,当所述的能量值小于所述的预设值时,所述的步骤(e)决定所述的人脸影像系撷取自所述的假人。 7. The face image recognition method according to claim 1, wherein, when the energy value is less than the predetermined value according to said step (e) determining the face images based captures from the dummy.
8.如权利要求1所述的人脸影像辨识方法,其特征在于,所述的步骤(e)系决定所述的人脸影像撷取自一真人,且所述的人脸影像辨识方法更包含以下步骤: 使所述的电子装置的一接收接口接收一登入信息;及使所述的处理器处理所述的登入信息。 8. The face image recognition method according to claim 1, wherein said step (e) the system determines the facial image retrieved from a real, and the face image recognition method further comprising the steps of: said electronic device is a receiver interface to receive a signed message; and a processor cause said processing of said information sign.
9.一种人脸影像辨识装置,其特征在于,所述的人脸影像辨识装置包含: 一存储器单元,储存复数张人脸影像; 一处理器,电性连接至所述的存储器单元,所述的处理器包括: 第一模块,用于对各所述的人脸影像计算一红色成分统计信息、一绿色成分统计信息及一蓝色成分统计信息, 第二模块,用于利用一独立成分分析法对所述的红色成分统计信息、所述的绿色成分统计信息及所述的蓝色成分统计信息中的至少二种颜色成分统计信息进行处理,藉此取得一第一成分信息及一第二成分信息, 第三模块,用于分别将所述的第一成分信息及所述的第二成分信息转换至一频率域,藉此取得一第一频域信息及一第二频域信息, 第四模块,用于计算所述的第一频域信息于一预设频率范围的一能量值,以及第五模块,用于将所述的能量值与一预设值进行比较,藉 A face image recognition apparatus, wherein the face image recognition apparatus comprises: a memory means storing a plurality of faces in the image; a processor, electrically connected to the memory cell, the said processor comprising: a first module for calculating a statistical information for each of the red component of the face image, a green component and a blue component statistics of the statistics, a second module for utilizing a separate component analysis of statistical information of the red component, the green component of the statistics and at least two color components of blue component of the statistical information in the statistical information is processed, thereby obtaining a first component and a second information two-component information, the third module, for respectively a first information component and said second component according to a frequency domain conversion information, thereby obtaining a first frequency information and a second frequency information, a fourth module configured to calculate first frequency information according to an energy value of a predetermined frequency range, and a fifth module for the energy value with a preset value, by 此决定所述的人脸影像撷取自一真人或一假人。 The decision of the facial image captured from a real person or a dummy.
10.如权利要求9所述的人脸影像辨识装置,其特征在于,所述的第一模块分离各所述的人脸影像的红色成分、绿色成分及蓝色成分,藉此得到复数张红色成分影像、复数张绿色成分影像及复数张蓝色成分影像,所述的第一模块对各所述的红色成分影像计算,以得所述的红色成分统计信息,所述的第一模块对各所述的绿色成分影像计算,以得所述的绿色成分统计信息,且所述的第一模块对各所述的蓝色成分影像计算,以得所述的蓝色成分统计信息。 10. The face image recognition apparatus as claimed in claim 9, wherein the red component of the separation of the first module of the facial image, a green component and a blue component, thereby to obtain a plurality of sheets red component of the video, a plurality of sheets of a plurality of sheets of a green component and blue component video images, the first image calculation module for each of the red component, the red component to obtain statistical information, each of said first module a green component of the image is calculated, to obtain statistical information of the green component and the blue component of the first image of each module of the calculations to obtain the blue component of the statistics.
11.如权利要求10所述的人脸影像辨识装置,其特征在于,所述的各红色成分统计信息为相对应的所述的红色成分影像的一平均亮度值,各所述的绿色成分统计信息为相对应的所述的绿色成分影像的一平均亮度值,且各所述的蓝色成分统计信息为相对应的所述的蓝色成分影像的一平均亮度值。 11. The face image recognition apparatus as claimed in claim 10, wherein each of the red component according to the statistics of the average luminance value of the red component of an image corresponding to said each of the green component Statistics an average luminance value of the green component of the image corresponding to the information, and each of the blue component statistics of an average brightness of a blue component of the image corresponding to said.
12.如权利要求9所述的人脸影像辨识装置,其特征在于,所述的第二模块利用所述的独立成分分析法对所述的红色成分统计信息、所述的绿色成分统计信息及所述的蓝色成分统计信息进行处理,以取得所述的第一成分信息、所述的第二成分信息及一第三成分信息,所述的第二模块更将所述的第三成分信息转换至所述的频率域,藉此取得一第三频域信息。 12. The face image recognition apparatus as claimed in claim 9, wherein said second module using the independent component analysis according to statistical information of the red component, the green component and statistics the blue component statistical information is processed to obtain information of said first component, said second component and a third component message information, said second module further information the third component the conversion to the frequency domain, thereby obtaining a third frequency information.
13.如权利要求9所述的人脸影像辨识装置,其特征在于,所述的第三模块分别对所述的第一成分信息及所述的第二成分信息进行一快速傅里叶转换,藉此取得所述的第一频域信息及所述的第二频域信息。 13. The face image recognition apparatus as claimed in claim 9, wherein said third module, respectively the second component of the information and the first component is a Fast Fourier Transform, thereby obtaining the second frequency information of the first information and the frequency domain is.
14.如权利要求9所述的人脸影像辨识装置,其特征在于,当所述的能量值大于所述的预设值时,所述的第五模块决定所述的人脸影像系撷取自所述的真人。 14. A face image recognition apparatus as claimed in claim 9, wherein, when the energy value is greater than a pre-determined value, said fifth module determines the facial image capturing system from the reality.
15.如权利要求9所述的人脸影像辨识装置,其特征在于,当所述的能量值小于所述的预设值时,所述的第五模块决定所述的人脸影像系撷取自所述的假人。 15. A face image recognition apparatus as claimed in claim 9, wherein, when the energy value is less than the predetermined value according to said fifth module determines the facial image capturing system from the dummy.
16.如权利要求9所述的人脸影像辨识装置,其特征在于,所述的人脸影像识别装置更包含: 一影像感测器,用以撷取复数张影像;以及第六模块,用于对各所述的影像进行人脸侦测以得所述的人脸影像。 16. The face image recognition apparatus as claimed in claim 9, wherein the facial image recognition apparatus further comprises: an image sensor for capturing a plurality of sheets of images; and a sixth module, with for each image in the face detection is performed to obtain the face images.
17.如权利要求9所述的人脸影像辨识装置,其特征在于,所述的人脸影像识别装置更包含: 一接收接口,电性连接至所述的处理器,且于所述的第五模块决定所述的人脸影像系撷取自一真人后,接收一登入信息;以及第七模块,用于处理所述的登入信息。 17. The face image recognition apparatus as claimed in claim 9, wherein the facial image recognition apparatus further comprises: a receiving interface, electrically coupled to said processor, according to the first and five module determines the facial image capturing since a real line, a receiving login information; and a seventh module for processing the login information.
CN201110355062.0A 2011-11-10 2011-11-10 Face recognition method and apparatus CN103106385B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201110355062.0A CN103106385B (en) 2011-11-10 2011-11-10 Face recognition method and apparatus

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201110355062.0A CN103106385B (en) 2011-11-10 2011-11-10 Face recognition method and apparatus

Publications (2)

Publication Number Publication Date
CN103106385A CN103106385A (en) 2013-05-15
CN103106385B true CN103106385B (en) 2016-06-01

Family

ID=48314235

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201110355062.0A CN103106385B (en) 2011-11-10 2011-11-10 Face recognition method and apparatus

Country Status (1)

Country Link
CN (1) CN103106385B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107205104A (en) * 2017-04-24 2017-09-26 奇鋐科技股份有限公司 Face identification module

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101216887A (en) * 2008-01-04 2008-07-09 浙江大学 An automatic computer authentication method for photographic faces and living faces
CN101999900A (en) * 2009-08-28 2011-04-06 南京壹进制信息技术有限公司 Living body detecting method and system applied to human face recognition

Family Cites Families (1)

* 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

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101216887A (en) * 2008-01-04 2008-07-09 浙江大学 An automatic computer authentication method for photographic faces and living faces
CN101999900A (en) * 2009-08-28 2011-04-06 南京壹进制信息技术有限公司 Living body detecting method and system applied to human face recognition

Also Published As

Publication number Publication date
CN103106385A (en) 2013-05-15

Similar Documents

Publication Publication Date Title
Swaminathan et al. Digital image forensics via intrinsic fingerprints
Cheng et al. Efficient salient region detection with soft image abstraction
Zhang et al. The unreasonable effectiveness of deep features as a perceptual metric
US9294475B2 (en) System and method for generating a biometric identifier
Fang et al. Bottom-up saliency detection model based on human visual sensitivity and amplitude spectrum
US7283645B2 (en) Object recognition using binary image quantization and Hough kernels
Zhang et al. A face antispoofing database with diverse attacks
Gu et al. Saliency-guided quality assessment of screen content images
Vig et al. Space-variant descriptor sampling for action recognition based on saliency and eye movements
Tao et al. Manifold ranking-based matrix factorization for saliency detection
Sun et al. Photo assessment based on computational visual attention model
CN102200830A (en) Non-contact control system and control method based on static gesture recognition
CN104166835A (en) Method and device for identifying living user
CN104395856B (en) The computer-implemented method and system for identifying the gesture
CN102096822A (en) Color discrimination method and device
US8013890B2 (en) Image processing apparatus and image processing method for recognizing an object with color
JP5772821B2 (en) Face feature point position correcting device, the face feature point position correcting method and face feature point position correcting program
Patel et al. Live face video vs. spoof face video: Use of moiré patterns to detect replay video attacks
Boulkenafet et al. OULU-NPU: A mobile face presentation attack database with real-world variations
CN103975330B (en) Directed on mobile platforms awareness certification
US9600744B2 (en) Adaptive interest rate control for visual search
JP2013164834A (en) Image processing device, method thereof, and program
US9721387B2 (en) Systems and methods for implementing augmented reality
US8965067B2 (en) Face data acquirer, end user video conference device, server, method, computer program and computer program product for extracting face data
CN102842128A (en) Method for object detection and apparatus using the same

Legal Events

Date Code Title Description
C06 Publication
C10 Entry into substantive examination
C14 Grant of patent or utility model