ImageVerifierCode 换一换
格式:PDF , 页数:7 ,大小:223.98KB ,
资源ID:532352      下载积分:5000 积分
快捷下载
登录下载
邮箱/手机:
温馨提示:
快捷下载时,用户名和密码都是您填写的邮箱或者手机号,方便查询和重复下载(系统自动生成)。 如填写123,账号就是123,密码也是123。
特别说明:
请自助下载,系统不会自动发送文件的哦; 如果您已付费,想二次下载,请登录后访问:我的下载记录
支付方式: 支付宝扫码支付 微信扫码支付   
验证码:   换一换

加入VIP,免费下载
 

温馨提示:由于个人手机设置不同,如果发现不能下载,请复制以下地址【http://www.mydoc123.com/d-532352.html】到电脑端继续下载(重复下载不扣费)。

已注册用户请登录:
账号:
密码:
验证码:   换一换
  忘记密码?
三方登录: 微信登录  

下载须知

1: 本站所有资源如无特殊说明,都需要本地电脑安装OFFICE2007和PDF阅读器。
2: 试题试卷类文档,如果标题没有明确说明有答案则都视为没有答案,请知晓。
3: 文件的所有权益归上传用户所有。
4. 未经权益所有人同意不得将文件中的内容挪作商业或盈利用途。
5. 本站仅提供交流平台,并不能对任何下载内容负责。
6. 下载文件中如有侵权或不适当内容,请与我们联系,我们立即纠正。
7. 本站不保证下载资源的准确性、安全性和完整性, 同时也不承担用户因使用这些下载资源对自己和他人造成任何形式的伤害或损失。

版权提示 | 免责声明

本文(ASTM E3115-2017 Standard Guide for Capturing Facial Images for Use with Facial Recognition Systems《用于面部识别系统的捕获面部图像的标准指南》.pdf)为本站会员(testyield361)主动上传,麦多课文库仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对上载内容本身不做任何修改或编辑。 若此文所含内容侵犯了您的版权或隐私,请立即通知麦多课文库(发送邮件至master@mydoc123.com或直接QQ联系客服),我们立即给予删除!

ASTM E3115-2017 Standard Guide for Capturing Facial Images for Use with Facial Recognition Systems《用于面部识别系统的捕获面部图像的标准指南》.pdf

1、Designation: E3115 17Standard Guide forCapturing Facial Images for Use with Facial RecognitionSystems1This standard is issued under the fixed designation E3115; the number immediately following the designation indicates the year oforiginal adoption or, in the case of revision, the year of last revis

2、ion. A number in parentheses indicates the year of last reapproval. Asuperscript epsilon () indicates an editorial change since the last revision or reapproval.1. Scope1.1 This guide is intended for use by practitioners who arechoosing, setting up, and operating photographic equipmentdesigned to cap

3、ture facial images for use with an automatedFacial Recognition System or used for manual comparisons bya trained facial examiner. This guide provides an overview ofhow to achieve the specifications defined in Annex E ofANSI/NIST-ITL-1-2011, Update 2015, for capturing facialimages.1.2 Annex E of ANSI

4、/NIST-ITL-1-2011 defines a well-controlled capture environment and subject whereas this docu-ment will give guidance where tight controls in the captureenvironment and subject control cannot be achieved.1.3 This guide addresses equipment considerations for two-dimensional (2D) conventional images. I

5、t does not addressvideo, scanners, or three-dimensional (3D) capture.1.4 The values stated in SI units are to be regarded asstandard. The values given in parentheses are mathematicalconversions to non-SI units that are provided for informationonly and are not considered standard.1.5 This standard do

6、es not purport to address all of thesafety concerns, if any, associated with its use. It is theresponsibility of the user of this standard to establish appro-priate safety, health, and environmental practices and deter-mine the applicability of regulatory limitations prior to use.1.6 This internatio

7、nal standard was developed in accor-dance with internationally recognized principles on standard-ization established in the Decision on Principles for theDevelopment of International Standards, Guides and Recom-mendations issued by the World Trade Organization TechnicalBarriers to Trade (TBT) Commit

8、tee.2. Referenced Documents2.1 ASTM Standards:2E2916 Terminology for Digital and Multimedia EvidenceExamination2.2 Other Biometric Standards:ANSI/NIST-ITL-1-2011, Update 2015 Data Format for theInterchange of Fingerprint, Facial and Other BiometricInformation3NIST Special Publication 500-280v2 Mobil

9、e ID Device BestPractice Recommendation, Version 2.033. Terminology3.1 Definitions:3.1.1 For terms relating to digital and multimedia evidence,refer to Terminology E2916.3.2 Definitions of Terms Specific to This Standard:3.2.1 fish eye effect, na type of distortion, where centralobjects of the image

10、 erroneously appear closer than those at theedge typically resulting in what appear to be unusually largenoses in the image.3.3 Acronyms:3.3.1 2Dtwo-dimensional3.3.2 FRfacial recognition4. Significance and Use4.1 The key factors that determine image quality for highlycontrolled facial images (for ex

11、ample: passports, policemugshots, drive motor vehicle, etc.) are well understood withrespect to their use in automated FR systems and are a criticalfactor which directly affects the searching accuracy of the FRsystem.1This guide is under the jurisdiction of ASTM Committee E30 on ForensicSciences and

12、 is the direct responsibility of Subcommittee E30.12 on Digital andMultimedia Evidence.Current edition approved Dec. 1, 2017. Published January 2018. DOI: 10.1520/E3115-17.2For referenced ASTM standards, visit the ASTM website, www.astm.org, orcontact ASTM Customer Service at serviceastm.org. For An

13、nual Book of ASTMStandards volume information, refer to the standards Document Summary page onthe ASTM website.3Available from National Institute of Standards and Technology (NIST), 100Bureau Dr., Stop 1070, Gaithersburg, MD 20899-1070, http:/www.nist.gov.Copyright ASTM International, 100 Barr Harbo

14、r Drive, PO Box C700, West Conshohocken, PA 19428-2959. United StatesThis international standard was developed in accordance with internationally recognized principles on standardization established in the Decision on Principles for theDevelopment of International Standards, Guides and Recommendatio

15、ns issued by the World Trade Organization Technical Barriers to Trade (TBT) Committee.14.2 Image quality also plays a major role in determining theextent to which a trained facial examiner is able to reach aconclusion as to whether two images containing faces arelikely to be of the same person or no

16、t. This applies to eitherreviewing a FR system candidate result set or 1:1 imagecomparisons.4.3 This guide provides guidance for the capture of facialimages under controlled, semi-controlled, and uncontrolledscenarios for constraints which can be categorized into thefollowing types of requirements:4

17、.3.1 Scene, refers to the content, subject and background inthe image,4.3.2 Photographic, refers to lighting, focus and otherconstraints required for image capture, and4.3.3 Digital, refers to the conversion of the captured imageinto a digital record.5. Scenarios5.1 Controlled Acquisition:5.1.1 This

18、 scenario is when all constraints can be controlledincluding equipment, the photographic environment (cameraposition, lighting, distance, background, and resolution), andthe pose and positioning of the subject. The subject can bedirected or positioned, the environment can be arranged toprovide ideal

19、 illumination, and suitable equipment is available.5.1.2 Referenced document ANSI/NIST-ITL-1-2011, Up-date 2015, Annex E, provides specifications on all aspects ofcontrolled acquisition scenarios.5.1.3 Annex A1 in this guide presents illustrative examplesfor controlled capture scenarios.5.1.4 Fig. 1

20、 provides an example of an optimal setting for acontrolled capture environment. The selection and placementof the lighting addresses the following variables:5.1.4.1 There is sufficient and uniform lighting to capture allvisible skin characteristics (that is, blemishes, moles, marks,etc.).5.1.4.2 To

21、avoid hot spots and shadows on the subjectsface.5.1.4.3 The 0.3 m to 0.6 m (1 ft to 2 ft) distance (with orwithout a backlight) between the background and the subject isintended to reduce background shadows.5.1.5 Fig. 2 provides examples of optimal facial imagescaptured. Frontal, right, left, and th

22、ree-quarter profile imagesshould be captured as these provide additional value forexaminers undertaking forensic comparison. The subjectshead is positioned in the image as shown.5.1.6 Automated image quality software is available fromvarious commercial suppliers that can help to determine if thefron

23、tal image is captured as per the recommendations.However, in all cases, a manual check is advised as thesoftware may not take into account all of the constraints.5.2 Semi-Controlled Acquisition:5.2.1 This scenario refers to when some, but not all,constraints can be controlled. Examples of this inclu

24、de pass-port photos not taken in a controlled environment, law enforce-ment mobile capture, crime scenes, access control, walk-through automated gates.5.2.2 In a semi-controlled environment, the following con-straints should be considered as a priority:5.2.2.1 Illumination,5.2.2.2 A single full fron

25、tal face with open eyes and neutralexpression.5.2.2.3 Elimination of convex lens distortion (subjectshould be more than1m(3ft)from the camera),5.2.2.4 Elimination of occlusions, andFIG. 1 Example of a Controlled Acquisition EnvironmentE3115 1725.2.2.5 Where available, multiple images should be cap-t

26、ured to compensate for shortcomings in individual images.5.3 Uncontrolled Acquisition:5.3.1 This scenario refers to when neither the environmentnor the subject can be controlled (for example, surveillance,hand held camera, body cameras, cell phones, etc.). It ischaracterized by high variations in qu

27、ality and content andtypically requires human review and specialized tools toidentify and extract usable facial content. (See NIST SpecialPublication 500-280v2.)5.3.2 In these situations, the person capturing the imageshould improvise to get the best image possible (see Fig. 3):5.3.2.1 When appropri

28、ate, use a camera flash or additionallights source(s) to improve illumination,5.3.2.2 Position the camera to get the best pose possible orprovide a visual attractor to influence the pose,5.3.2.3 A single full frontal face with open eyes,5.3.2.4 Elimination of convex lens distortion (subjectshould be

29、 more than1m(3ft)from the camera), and5.3.2.5 In these uncontrolled scenarios, multiple imagesshould be captured to compensate for shortcomings in indi-vidual images.6. Keywords6.1 capture equipment; facial image; facial recognitionFIG. 2 Example Poses from a Controlled Acquisition EnvironmentE3115

30、173NOTE 1Tennessee Bureau of Investigation, “Mug Shot Best Practices Guide,” available from https:/www.tn.gov/content/dam/tn/tbi/documents/MugShot Best Practices Poster.pdf.FIG. 3 Examples of Improper Poses from a Controlled Acquisition EnvironmentE3115 174ANNEX(Mandatory Information)A1. TABLESA1.1

31、See Tables A1.1-A1.3.TABLE A1.1 Description of a Controlled Acquisition EnvironmentNo. Item Description Comments1 Illumination Lighting shall uniformly illuminate the subject and thebackground.Use of an optional third light as a backlight (to eliminate shadowson the background) generally requires ab

32、out 1 m (3 ft) ofadditional floor space behind the subject. This additional spacemay not be available in all environments.Hot spots or reflections shall be minimized. Use of flash is discouraged. However, if a flash is used, wherepossible, either diffuse the light or bounce the light off aneutral-co

33、lored ceiling to avoid harsh, uneven lighting.Consideration should be made to the color of the ceiling toavoid casting unnatural color on the subject.2 Camera position The camera shall be at eye level and shall be mounted in ahorizontal position with no tilt.Possible solutions are to place subject i

34、n an adjustable seatedposition or to mount the camera on a pole and adjust itsheight.The camera height shall be positioned at the same height as thesubjects eyes and positioned approximately2m(6.5ft)fromthe subject.3 Background Background should be uniform light gray. 18 % grey is recommended as a b

35、ackground color. 18 %photographic grey cards and backgrounds are widely availablefrom photographic suppliers.The level of gray is less important than ensuring that it is aneutral gray and that the subjects head is clearlydistinguishable from the background with no shadows or otherobjects visible in

36、the frame.The background shall completely fill the image frame behind thesubject.TABLE A1.2 Optimal Subject AppearanceNo. Item Description Comments1 Facing Frontal:Subject shall face directly towards the camera with no morethan 5 variance from frontal in roll, pitch, and yaw.Visual cues that can ass

37、ist an operator in determining “forwardorientation” are:Profile:When there is a requirement for profile shots, the whole bodyshall be turned 45 and 90 right and left.(1) Both eyes are level on an imaginary horizontal line (zero rollangle),(2) Both ears are equally visible if unobstructed by hair) an

38、dnose is forward (zero yaw angle), and(3) The chin is neither elevated or dropped (zero pitch angle).2 Head coverings Head coverings, including hats and scarves, shall not be worn,unless for religious reasons or medical purposes.The full face shall be displayed and the ears should bedisplayed.3 Hair

39、 When capturing both frontal and profile images, the subjectshair should be moved back.The full face shall be displayed and the ears should bedisplayed.4 Glasses If the subject wears glasses, at least one frontal image shouldbe captured with glasses and one image without glasses.If the subject is we

40、aring glasses, glare from the lenses shall beavoided, for example, by adjusting the lighting.Eye patches are allowed only for medical reasons.5 Expression Subject shall have a neutral expression. It is the operator and subjects responsibility to ensure there areno smiles or grimaces.6 Mouth Mouth sh

41、all be closed.E3115 175TABLE A1.2 ContinuedNo. Item Description Comments7 Shoulder position Shoulder position shall be square and forward facing for frontalimages. Shoulder position shall be perpendicular for profileimages.8 Accessories Accessories should be removed.9 Make-up andcleanlinessThe subje

42、cts face shall be presented without heavy makeup ordirt.Use of a disposable wipe cloth is recommended to removeexcessive makeup or surface dirt.This may not be practical in noncriminal image capture.10 Face count Only one face per image is allowed. Take measures to ensure that other subjects are not

43、 captured,such as a baby in arms.11 Medical conditions If bruising, injuries, bandages, or medical conditions arepresent, these should be presented as is.The presence of injuries or bandages on the face will limit thevalue of the images for use with automated FR systems.Depending on the use case, it

44、 may be beneficial to capture anadditional image at a later date when the injuries are notvisible.TABLE A1.3 Optimum Parameters for Controlled AcquisitionNo. Item Description Comments1 Camera type Use a high-quality digital camera such as a digital single-lensreflex (DSLR) or a point and shoot camer

45、a with manualexposure override features.Digital cameras that lack manual exposure override are lesssuitable.The use of webcams and cell phone cameras is notrecommended.2 Camera resolution ANSI/NIST-ITL-1-2011, Update 2015, provides subject acquisi-tion profiles (SAP) that describe a set of character

46、istics ex-pected of the facial image with stringency of the requirementsincreasing with SAP level.Higher SAP levels result in larger pixel image sizes (height,width), on-disk file storage, and will require a correspondingincrease in the camera resolution.The highest level of SAP 50 and 51 are recomm

47、ended for foren-sic image analysis.3 ISO rating If the ISO setting on the camera is adjustable, it should be set tothe lowest value that also meets shutter and aperture require-ments (in accordance with Items 7 and 8 in this table).Operators should be familiar with the cameras automated fea-tures th

48、at may help simplify the adjustment of ISO, shutterspeed, and f-stop to compensate for low light conditions.The setting should be set so that there is no visible noise in theimage.4 Focus The operator can use auto or manual focus. Do not use fixed focus (for example, cell phones or disposablecameras

49、).Good results will normally be obtained by focusing on the eyearea.5 Face detectioncapability (facepriority auto fo-cus)Optional. If the camera has face detection capability, its use is recom-mended.If there are multiple subjects, it shall be verified that the correctface has been selected and is in focus.6 White balance The white balance setting on the camera should be adjustable tocapture natural colors.The white balance should be calibrated using 18 % gray back-ground or gray card. Calibration should be done as part of aregular preventive maintenance program

copyright@ 2008-2019 麦多课文库(www.mydoc123.com)网站版权所有
备案/许可证编号:苏ICP备17064731号-1