VeriLook 5.3/MegaMatcher 4.4 Algorithm Demo

Transcription

1 VeriLook 5.3/MegaMatcher 4.4 Algorithm Demo User's guide

2 User's guide version: Publish date: 1/29/2013

3 Table of Contents 1 Introduction System Requirements 1 2 IP Cameras Configuration 2 3 Face Image Constraints 5 4 Liveness detection 6 5 Matching Threshold and Score 7 6 Application Windows Main Window Options Dialog Menu Commands Simple Usage Examples Linux Mac OS X 13 Index a iii

4 1 Introduction VeriLook 5.3/MegaMatcher 4.4 Algorithm Demo application is designed with aim to demonstrate the capabilities of Neurotechnology faces recognition engine. The program is a Windows 2000/XP/Vista/7 compatible GUI application. 1 Evaluation software supports image acquisition from the external video source (such as Web or IP cameras ( see page 2)). It can also read face images from still images (.bmp,.tif,.png,.jpg,.gif files). The application has 3 operation modes: 1. Enrollment. Software processes the face image, extracts features and writes them to the database. 2. Face enrollment features generalization. This mode generates the generalized face features collection from a number of the face templates of the same person. Each face image is processed and features are extracted. Then collections of features are analyzed and combined into one generalized features collection, which is written to the database. The face recognition quality increases if faces are enrolled using this mode. 3. Matching. This mode performs new face image matching against face templates stored in the database. 1.1 System Requirements Windows requirements: 128 MB of RAM, 1Ghz CPU, 9MB HDD space for the installation package. Microsoft Windows 2000/XP/Vista/7. DirectX 8.1 or later. You can download DirectX upgrade from Microsoft web site. Microsoft GDI+. This library is supplied with Windows XP and Windows.NET Server family. If you are using any other modern Windows platform (Windows 98/Me and Windows NT 4.0/2000) you should download and install it from Microsoft web site. The Microsoft XML Parser (MSXML) 3.0 SP4 is required so if it is not already in the system you should download and install it from Microsoft web site. (Optionally) Video capture device (web camera). Linux requirements: PC with x86 compatible CPU Linux OS 2.6 or newer GCC-4.0.x or newer pkg-config or newer GNU Make 3.81 or newer GTK x or newer libraries and development packages libtiff-3.8.x or newer libraries and development packages libusb or newer libraries. Mac OS X: 1

5 2 IP Cameras Configuration VeriLook 5.3/MegaMatcher 4.4 Algorithm Demo can be used with most IP cameras. How to configure an IP camera read the following: Configuring on Microsoft Windows IP Cameras are configured through NdmMedia.ini configuration file in the root directory of sample distribution. The file contains commented example of the possible configuration, the format is described further. The ';' character can be used to comment out the line. The file uses usual.ini format, where certain parameters are grouped into sections: Section Description 2 [NVideoSource] Specifies parameters controlling usage of all the local video devices. [Camera] Section defines additional video input device. This can be remote stream (for example from IP camera) or "virtual" device which would provide video from media file. The first section described ([NAudioSource] should appear in the configuration file only once, as this setting control all the local devices. The other section ([Camera] define a new device and can appear as many times as needed. IP cameras can be used through two frameworks: DirectShow or Windows Media Foundation. Following flags allow to select which way is preferred and can appear in just described sections: Flag UseDirectShow Description Flag used to set whether to use DirectShow for video source. UseDirectShow=False - disables DirectShow UseDirectShow=True - enables DirectShow UseWindowsMediaFoundation Flag used to set whether to use Windows Media Foundation for video source. UseWindowsMediaFoundation=False - disables Windows Media Foundation UseWindowsMediaFoundation=True - enables Windows Media Foundation PreferDirectShow AllowDuplicateDevices Flag used to prefer DirectShow versus Windows Media Foundation. PreferDirectShow=False - prefers to use Windows Media Foundation PreferDirectShow=True - prefers to use DirectShow vs. Windows Media Foundation This flag is used to allow enumerating duplicate devices. If this flag was set to true and the same device is accessible through DirectShow and Windows Media Foundation (Windows 7) then this device will be enumerated twice. Note: Valid only for [NVideoSource]) section. Following parameters define a device and can be used in [Camera] section: Parameter Url Filename Id Description Rtsp address of media source (IP camera). For example: Url=rtsp:// :554/media/video1 Media source file name. For example: Filename=C:\test.wav Note: Only Filename or Url address can be specified at a time, not both. Camera Id (optional parameter). This Id overrides original camera Id. For example: Id=My Camera 2

6 DisplayName Display name for camera (optional parameter). This parameter overrides default camera display name. For example: DisplayName=My Camera Display name Configuring on Linux and Mac OS X IP cameras on Linux and Mac OS X are available through gstreamer framework. Rtsp stream or local media file for their content are used. IP cameras are configured through NdmMedia.conf configuration file in the root directory. The file contains commented example of the possible configuration, the format is described further. The '#' character can be used to comment out the line. The file uses.ini like format, where certain parameters are grouped into sections. Each sections defines a new device and can appear as many times as needed: Section Description [Camera] Section defines additional video input device. This can be remote stream (for example from IP camera) or "virtual" device which would provide video from media file. 2 The device is described by following parameters in each section: Parameter Url Filename Id Description Rtsp address of media source (IP camera). For example: Url=rtsp:// :554/media/video1 Media source file name. For example: Filename=/tmp/test.wav Note: Only Filename or Url address can be specified at a time, not both. Camera Id (optional parameter). This Id overrides original camera Id. For example: Id=My Camera DisplayName Display name for camera (optional parameter). This parameter overrides default camera display name. For example: DisplayName=My Camera Display name Note: Some IP cameras need special special configuration and their use is described in separate subsections below. Cisco Video Surveillance IP 4500 Cisco Video Surveillance IP 4500 camera is supported through NdmCisco plugin. The camera should be configured by editing the NdmCisco.conf file in \Cameras\NdmCisco directory of the Algorithm Demo. Multiple cameras can be specified, by using the following template (for each camera): [Camera] Ip= Username=admin Password=Admin123 ChannelId=1 ChannelName=Channel1 Id=MyId DisplayName=MyDisplayName 1. IP - defines IP address of camera channel 2. Username - camera's username. To set camera's username use Cisco software. 3. Password - camera's user password. 4. ChannelId - Camera's channel ID. 5. ChannelName - Name of camera's channel. 6. Id - NDeviceManger ID for the camera. 3

7 7. DisplayName - display name for the camera. Mobotix DualNight M12 Mobotix DualNight M12 IP camera is supported through NdmMobotix plugin. The camera should be configured by editing the NdmMobotix.conf file in \Cameras\NdmMobotix directory of the Algorithm Demo. Multiple cameras can be specified, by using the following template (for each camera): [Camera] Ip= Id=MobotixId DisplayName=MobotixDisplayName Channel=right 1. IP - defines IP address of camera channel 2. Id - NDeviceManger ID for the camera. 3. DisplayName - display name for the camera. 4. Channel - camera's channel. 2 Prosilica GigE Vision Prosilica GigE Vision camera is supported through NdmProsilica plugin. The plugin is saved as NdmProsilica.dll file in \Cameras\NdmProsilica directoriy of the Algorithm Demo. Note: This camera does not require any additional configuration. 4

8 3 Face Image Constraints Face recognition accuracy heavily depends on face image quality. Especially, maximum care should be taken in acquiring good quality facial images in enrollment step. During enrollment one or more (if available) images are converted into facial features template which is usually saved into database and can be further used to rapidly identify or verify the person. The following basic recommendations and requirements should be taken under consideration: Face pose and expressions 1. Use near-frontal face images for enrollment with a rotation deviation of up to 15 degrees in any direction (yaw, pitch, roll). MegaMatcher and VeriLook SDK supports roll angle deviation up to 180 degrees. 2. Use more than one image for enrollment to cover slightly different face views in yaw and pitch directions. In case of successful coverage, faces up to 25 degrees deviation from frontal can be recognized. 3. Neutral face expression is preferred in the enrollment step. During identification slight changes in facial expression do not influence face recognition accuracy. 4. Persons wearing glasses should use several images - with and without glasses - for enrollment. Thus, during identification they will be recognized in both cases. 5. Glasses with heavy frames and sunglasses will decrease face recognition accuracy. 6. Changes in face appearance because of facial hair should be accommodated in the same manner as wearing of glasses. Note: the minimum distance between eyes in the image should be 50 pixels. The recommended distance is 75 pixels. Lightning 3 1. If lighting conditions can be controlled: 1. Lighting should be equally distributed on each side of the face and from top to bottom with no significant shadows within the face region. This can be achieved by using direct frontal light or diffuse lighting. 2. Avoid sunlight or any other additional lighting. 3. Avoid lighting which can produce glares on glasses or even face skin. Cameras 1. Use cameras of similar quality for both enrollment and identification. 2. Some cameras can be configured to produce mirrored images, i.e. left eye of the persons face can become "right" eye in the face image. Make sure that face orientation is uniform across all the images used in face recognition because facial features template extracted from mirrored face images of the same person will not match with those extracted from not-mirrored ones. 5

9 4 Liveness detection Faces algorithm is capable to differentiate live faces from non live faces (e.g. photos). "Use liveness check" checkbox and "Liveness threshold" parameter in the options dialog controls the behavior of liveness check. When "Use liveness check" checkbox is marked, the liveness check is performed while matching. That is the liveness score of collected stream is calculated and checked against the liveness score threshold set in the "Liveness threshold" parameter. Using liveness check requires a stream of consecutive images. (This check is intended to be used mainly with video stream form a camera). The stream must be at least 10 frames length and the recommended length is frames. Only one person face should be visible in this stream. If the stream does not qualify as "live" and "Extraction failed" message is displayed in the log window. To maximize the liveness score of a face found in an image stream, user should move his head around a bit, tilting it, moving closer to or further from the camera while slightly changing his facial expression. (e.g. User should start with his head panned as far left as possible but still detectable by face detection and start panning it slowly right slightly changing his facial expression until he pans as far right as possible (but still detectable by face detector)). 4 6

10 5 Matching Threshold and Score VeriLook 5.3/MegaMatcher 4.4 Algorithm Demo features matching algorithm provides value of features collections score as a result. The higher is score, the higher is probability that features collections are obtained from the same person. Matching threshold is linked to false acceptance rate (FAR, different subjects mistakenly accepted as of the same) of matching algorithm. The higher is threshold, the lower is FAR and higher FRR (false rejection rate, same subjects erroneously accepted as different) and vice a versa. You can convert matching threshold to FAR (false acceptance rate) and vice versa using this table: FAR (False acceptance rate) 1 % % % % % % % 96 Matching threshold (score) or using this formula: Threshold = -12 * log10(far); where FAR is NOT percentage value. 5 7

11 6 Application VeriLook 5.3/MegaMatcher 4.4 Algorithm Demo incorporates Neurotechnology face recognition algorithm. Using this demo application face images can be enrolled from still images (from image files) or video streams (from cameras) and identification task performed. 6.1 Windows VeriLook 5.3/MegaMatcher 4.4 Algorithm Demo application in Windows OS can be started from FacesAlgorithmDemo.exe Main Window Main application window has three-pane layout, where top pane is used for displaying image or video stream and two bottom panes are used for messages logging. Menu commands and two toolbar buttons, used as shortcuts for most accessed commands, control application. Person can be enrolled using "Enroll" command. After enrolling a person's image from file the main window looks like this: 6 The main window panes are these: 8

12 1. Top face detection pane, used to display video or still images and result of face detection algorithm overlaid on image. 2. Left pane is application log, used for system information and application progress messages. 3. Right pane is matching results pane for listing id of the subject in the database, most similar to matched image. Subjects are considered similar if their score value exceeds matching FAR (False acceptance rate) set via Options (Options->Identification) dialog. Also in this pane matching score is displayed Options Dialog Options dialog window can be accessed through menu command File->Options... Options dialog window is used to configure faces detection, extraction, enrollment and identification parameters. Face Detection: Minimum IOD minimum distance between eyes. Maximum IOD maximum distance between eyes. Face confidence threshold value which controls the requirements for face detection. The greater this value is the more strict rules are applied when looking for faces in an image. Favor largest face - if checked, the largest face found in an image will be used for facial feature extraction instead of the one with the highest face confidence threshold score. Detect gender - if checked, the application will detect gender. Detect all features when extracting from file - if checked, all face features (not only these necessary for extraction) are detected when extracting from a file. Detect base feature points when extracting from file - if checked, eyes, nose tip and mouth will be detected. Max roll angle deviation - this value is used to define the maximum value in degrees of rolled face image which can be enrolled or identified. Max yaw angle deviation - this value is used to define the maximum value in degrees of yawed face image which can be enrolled or identified. Face extraction: 6 9

13 Face quality threshold controls how strict rules are applied when determining the quality of a found face for extraction. Enrollment: Save images to database - when checked saves images of enrolled faces to database. Use filename as template id - image filename is used as id for enrolled face template. Template size (enroll from image) - size of face image templates when enrolling from image. Can be used Large, Medium or Small template. It is recommended to use large template size. Template size (enroll from stream) - size of face image templates when enrolling from image. Can be used Large, 6 10

14 Medium or Small template. It is recommended to use large template size. Frame count maximum number of frames to process with face detection algorithm while enrolling subject using video camera or stream. Max record per template - maximum number of extracted face records to be saved within one face records template. Generalization template count - templates count for face features generalization (how many templates should be enrolled to generalize face features). Identification: Template size (identify from image) - default size of face images templates when performing identification from image. Can be used Large, Medium or Small template. It is recommended to use medium template size. Template size (identify from image) - default size of face images templates when performing identification from stream. Can be used Large, Medium or Small template. It is recommended to use medium template size. Disable liveness check Indicates that no liveness check ( see page 6) will be used while matching. Frame count maximum number of frames to process with face detection algorithm while identifying subject using video camera with no liveness check. Enable liveness check Indicates that the liveness check ( see page 6) will be used while matching. Liveness threshold controls the requirements for live faces. The greater this value is the more strict rules are applied to check if face in an image stream is live. (If this value is set to 0 all faces are considered to be live). Frame count maximum number of frames to process with face detection algorithm while identifying subject using video camera with liveness check enabled. Minimum value is 10. FAR - threshold that separates identical and different subjects. Matching threshold is linked to false acceptance rate (FAR, different subjects erroneously accepted as of the same) of matching algorithm. The higher is threshold, the lower is FAR and higher FRR (false rejection rate, same subjects erroneously accepted as different) and vice a versa. Speed defines identification speed. When the most accurate identification results are required it is recommended to use Low speed. But in this case identification task is performed slower. If the maximum identification speed is required it is recommended to use High identification speed. Note: Template size also affects identification speed. The highest possible identification speed is achieved when Small template and High speed is used. Misc: Flip image horizontally mirror horizontally image received from video camera. 6 11

15 Flip image vertically - mirror vertically image received from video camera Menu Commands The following table explains menu commands of the Faces Algorithm Demo application. Menu command Jobs» Enroll Jobs» Enroll with generalization Jobs» Identify Tools» Clear logs Tools» Clear database Tools» Options Help» About Description Enrolls image to face database. Enrolls image to faces database with features generalization. Search for matching image in face database. Clear application log windows. Empty faces database. Display options dialog. Display information about Faces algorithm demo application Simple Usage Examples In this section simple basic scenarios of using VeriLook 5.3/MegaMatcher 4.4 Algorithm Demo application are described in a step by step fashion. Enrolling from camera 1. First, camera to be used as the capture device must be selected from "source" menu in the toolbar. After that camera video input should be visible in the upper pane of the program. 2. Faces found in the video stream are outlined in the capture image by green rectangle (the green rectangle outlines the face that best fits the matching requirements of the Neurotechnology Faces algorithm in addition this face has its eyes marked by the program). 3. To enroll a face from a video stream, "enroll" button in the toolbar can be used or option "enroll" from a system menu "jobs" can be selected. For this operation to succeed at least one face in the image must be present. Program will process a few frames and will enroll face into the database of the demo program from these frames and a dialog asking for the person to be enrolled id will be shown. Matching from camera 1. First, camera to be used as the capture device must be selected from "source" menu in the toolbar. After that camera video input should be visible in the upper pane of the program. 2. Faces found in the video stream are outlined in the capture image by colorful rectangles (the green rectangle outlines the face that best fits the matching requirements of the Neurotechnology Faces algorithm in addition this face has its eyes marked by the program, and yellow rectangles show other faces found in the image). 3. To identify a face "Identify" button must be clicked or option "Identify" must be selected from a system menu "jobs". After this the face, that best suits the matching requirements of the Neurotechnology Faces algorithm (it should be outlined by a green rectangle in the video input pane) will be matched against the database of the demo program and most probable candidate will be displayed in the bottom right pane of the program window. Enrolling from file 1. First, file input as the capture device must be selected from "source" menu in the toolbar. 2. To enroll a face from a file "enroll" button must be pressed or "enroll" option selected from the system menu "jobs". After that a file open dialog should open in which a file to be opened must be selected. The image in the file will be displayed in the upper pane of the window, with the found face outlined by a green rectangle (if such rectangle is absent it means that no face suitable for enrollment was found in the image) and a dialog asking for the person to be enrolled id will be shown. 6 12

16 The outlined face will be enrolled to the demo program database. Matching from file 1. First, file input as the capture device must be selected from "source" menu in the toolbar. 2. To identify a face from a file "match" button must be pressed or "match" option selected from the system menu "jobs". After that a file open should open in which a file to be opened must be selected. The image in the file will be displayed in the upper pane, with the found face outlined by a green rectangle (if such rectangle is absent, it means that no face, suitable for matching was found in the image). The outlined image will be matched against the demo program database and most probable candidate will be displayed in the bottom right pane of the window. 6.2 Linux VeriLook 5.3/MegaMatcher 4.4 Algorithm Demo application on Linux can be started from FacesAlgorithmDemo file. There are separate distribution for 32 bits and 64 bits operating system. See Windows ( see page 8) part to find out how to use this application. 6.3 Mac OS X VeriLook 5.3/MegaMatcher 4.4 Algorithm Demo application on Mac OS can be started from FacesAlgorithmDemo.app file. See Windows ( see page 8) part to find out how to use this application. 6 13

17 Index A Application 8 F Face Image Constraints 5 I Introduction 1 IP Cameras Configuration 2 L Linux 13 Liveness detection 6 M Mac OS X 13 Main Window 8 Matching Threshold and Score 7 Menu Commands 12 O Options Dialog 9 S Simple Usage Examples 12 System Requirements 1 W Windows 8 a