1 Introduction

1.1 About this document

This is the User Manual for SAFR Camera. This document describes capabilities specific to the SAFR Camera. Please refer to the SAFR Software Administration Guide for more general software capabilities that are applicable to SAFR Camera and 3^rd Party Camera Surveillance solutions as well as SAFR SCAN, the Access Control Solution.

1.2 About SAFR Camera

SAFR Camera brings NIST-leading live video face recognition and video analytics features to the edge. Embedded directly into the camera, SAFR utilizes onboard compute functionality for face detection, recognition and other analytics — eliminating the need for servers dedicated to processing video streams.

By eliminating expensive server hardware, SAFR SCAN significantly reduces the high cost and maintenance of GPU enabled server class hardware to process video and perform facial recognition. It also avoids the need for transmitting high quality video streams over the network. Finally, SAFR Camera improves reliability by bringing the watchlist and face matching to the edge.

While SAFR Camera can operate as a standalone camera with no network connection, it is also possible to take advantage of the powerful multitiered architecture of SAFR Server to provide powerful administration tools. SAFR Cameras can be managed centrally by SAFR Server or SAFR Cloud. SAFR Server or SAFR performs the following roles:

• Distributed Camera configuration (single or bulk updates)
• Firmware updates
• Centralized watchlist management
• Event aggregation for monitoring and reporting
• Event to action workflows

1.3 SAFR Architecture

SAFR Cameras support either on premises SAFR Server to keep all your data on-site or SAFR Cloud for simple serverless deployments. The graphic below depicts each of these configurations.

SAFR Cloud Deployment (Serverless Deployment)	SAFR on premises Deployment (Control your data)

With a SAFR Cloud deployment, the only hardware on premises is the SAFR Camera. No PC is required. Bandwidth requirements to SAFR Cloud are limited to face recognition events and device status.

With on premises deployments, all data remains on site.

1.3.1 Hybrid Architecture

SAFR Camera can also be used in a hybrid deployment. A hybrid deployment allows for SAFR Server to share limited data with systems outside of the local network. This may be useful for connecting multiple locations or providing a means for operators to connect to SAFR Cloud for receiving notification and viewing events.

The diagram below demonstrates a hybrid deployment. SAFR Server on premises connects to SAFR Cloud which can facilitate sharing of data between networks.

An operator can receive notifications, browse events and manage watchlists from a mobile device that is not connected directly to the same network as SAFR Server and the SAFR Cameras are hosted.

1.3.2 Advantages of 4k on-board processing

Most IP video cameras are 2 MP (1920x1080). For surveillance, this resolution of video offers a good tradeoff between cost and quality. Costs associated with video are due to storage size and processing power. As the video doubles in width, the storage and processing requirement typically increase by 4x. Thus, 4k video feeds, while only 2 times wider, take up 4 times the storage and processor power.

A 4k (8MP) video stream takes up to 15 Mbps of network bandwidth for each camera. A network with just 20 cameras would require 120 Gbps of network traffic. 50 cameras would saturate a 1 Gbps network. But when processing is moved to the edge, video need no longer be transmitted over the network or processed on a CPU. The SAFR Server is just there to aggregate a low bandwidth (50 kb per recognition event sent to server).

For software-based facial recognition , server class hardware is required with fast CPUs and expensive GPUs. A top of the line GPU costing thousands of dollars will process 12 to 15 4k video streams and require either Silver or Gold Xeon CPUs. Systems like this can cost $10,000 to $15,000. But with SAFR Camera, processing is kept on the edge and server is simply aggregating event data, avoiding the need for expensive GPUs and fast CPUs.

1.4 SAFR Camera vs. SAFR Software-based Facial Recognition

SAFR provides two products that perform facial recognition:

SAFR Software

SAFR works with 3rd party video cameras (USB or RTSP) to process video in software for the purpose of performing facial detection, facial recognition and other analytics in software. This is most often done with the help of NVidia GPUs which enhance the speed and efficiently of the software solution to scale to many cameras. In addition to analytics, SAFR Software manages watchlists, aggregates events and other management functions.

SAFR Camera

SAFR performs facial detection, facial recognition and other analytics on the camera. The SAFR Software is used to manage watchlists and aggregate events but does not have the heavy task of processing video.

SAFR offers two products for performing facial recognition. Both products perform fundamentally the same actions and have a great deal of overlap. SAFR Desktop and SAFR Server Web Console provide a single interface that works with both cameras.

Both products interoperate with the SAFR Server. Each product is not mutually exclusive. A deployment can have a mix of 3^rd party cameras and SAFR Camera. There are some features that are only available in software. They are:

• Person Body Detection – Detect a person’s body which is useful to measure true dwell time (i.e. if a person turns around, we still track as same person when their face was visible) and occupancy counting (counting people both as they enter and as they leave an area). Person body detection is on the roadmap for SAFR Camera.
• 2D Liveness – Ability to detect if a face is photo or a live person at the camera.
• VMS Integrations – SAFR developed VMS integrations are done in software and not yet available with SC800 camera. SAFR Actions based integrations exist that can be made available as sample code that integrators can deploy, but these are not part of the SAFR product line as of yet.

1.5 SAFR Camera Key Features

1.5.1 Accuracy

SAFR has the fastest and smallest facial recognition amongst submitted NIST algorithms making it possible for high accuracy facial recognition on low power solid state devices.

• 99.87% precision in LFW faces FNMR
• Accuracy with masked faces: 96.16% TPR at 0.0047% FPR on masked faces
• Face match for Live Watchlist: < 100ms Average Response Time
• Face search / Forensics analysis: 500,000 events/second

1.5.2 Offline operation

SAFR Camera are completely self-contained. Should the camera lose its network connection for any reason, it will continue to function using a locally cached watchlist. SAFR Camera stores the entire watchlist on the device. The cameras only need the network connection to receive watchlist updates, perform configuration or transmit events. During the time a camera may be offline, events are cached on the device and pushed to the server once network connectivity is restored.

1.5.3 Centralized Administration

SAFR Cameras are connected to SAFR Server or SAFR Cloud at installation time. Cameras can operate independently but this would mean managing the watchlists and camera configuration one by one. By connecting to a centralized system, cameras can be grouped into zones and managed centrally.

1.5.4 Bulk Camera Configuration

Settings can be applied to a single camera or to multiple cameras in bulk operations. Configuration can be applied by zones or to all cameras in a single operation. Once configured, watchlist are distributed to all cameras according to zone definitions.

1.5.5 Event Aggregation

As SAFR Camera matches each face, event information is sent to SAFR Server or SAFR Cloud where it can be centrally viewed and filtered by attributes such as site (e.g. building or other grouping location), source (camera), time, id class (threat, concern, stranger), age, gender and more. Reporting can be performed on historical event data.

1.5.6 Forensic Search (Search by Image)

A search of a face image can be performed against the event history or person database to find when and where the face appears. This is particularly useful when a face image was not already enrolled before an incident involving that person occurred.

1.5.7 Event to Action Triggers

Event can be used to trigger actions such as SMS or Email alerts. SAFR Actions is a tool that can be used to trigger any type of action such as:

• Running a custom script
• Make a REST API call to an external system
• Play a sound or trigger a tower light
• Generate a custom log of interesting events

1.5.8 Other Features

SAFR Camera also supports the following:

• 4k video processing (cover 2x wider area with same accuracy than 2MP cameras)
• Options for 6 to 22mm or 5 to 50mm motorized varifocal lens
  • High accuracy recognition up 100 feet (30m) or 200 feet (60m)
• Face prioritization auto-exposure for perfect lighting to the face for best recognition results
• Automated behavior with advanced control over shutter speed and aperture
• Highly secure to protect personally identifiable information
  • Data is encrypted at rest (on disk) with AES-256 and RSA-2048 ciphers
  • Data is transmitted over TLS (https) to encrypt all transactions
  • Individual user accounts controlled by an access control list (ACL)
  • Device protected by secure boot that prevents rogue OS loading
• Flexible Delivery: On premises, cloud & hybrid deployment models
• Characterization: Anonymous age, gender, gaze, occlusion and sentiment detection
• Horizontal Scalability through server clustering

2 SAFR Camera Configuration

This section describes functionality and configuration specific to SAFR Camera. The focus is on settings exposed in the SAFR SCAN Web Console. As such, this section follows the structure of SAFR SCAN Web Console.

As described above in “Camera vs. Software”, SAFR offers two products for performing facial recognition. Common functionality that is shared between both products are fully documented in SAFR Administration Guide.

2.1 Live Page

The Live tab displays the camera view of the SAFR Camera device, and is the default tab of the SAFR Camera Web Console. You can view events and enroll new faces from this page.

2.1.1 Events

Events appear on the right. Color indicates the following:

Color	Meaning	Description
Grey	Unrecognizable	Insufficient quality for a match. SAFR will not attempt to match if quality is too low. This may happen if face size is too small or too dark to be deemed acceptable quality to attempt a match.
Purple	Stranger	Face was of sufficient quality but no match was found.
Green	Known/No-concern	Face matched in watchlist and marked as “No Concern”.
Red or Yellow	Known/Threat or Concern	Face matched in watchlist and person ID Class was set to Threat or Concern. Many of SAFR’s capabilities allows for special handling of persons based on ID Class.

Strangers in the Live page will appear with an “Enroll” button. Click the enroll button to add the face to the watchlist.

2.1.1.1 SAFR Event Types

SAFR has many event types. This section documents only those event types applicable to SAFR Camera. See SAFR Administration Guide for information on all event types.

2.1.1.1.1 Person Event

Person event is generated based upon appearance of a face. Color and text will vary depending on the conditions.

Analytics also appear in person events when enabled. Following shows event with age, gender and sentiment enabled.

2.1.2 Event Face Image Selection

SAFR Camera will capture the best quality face that appears while a person is in view. For each face that appears in the camera view, an event is generated. An event is not just a single point in time but has start and and end time. When a new face appears in the camera view, an event is started. SAFR tracks the face from frame to frame. With each frame, SAFR evaluates the face image against various metrics such as face size, center pose, sharpness, contrast and occlusion. If SAFR finds a better face, the event image is updated. SAFR also updates data such as age estimate and sentiment while a face is still in view. When the face leaves view of the camera, the event is updated with an end time and no further updates are made.

In SAFR Desktop, options exist to enroll a stranger face from the event or to update an existing stored identify with the best face obtained during the event. See SAFR Software Administration Guide for more details.

2.2 People Page

The People tab allows you to manage and configure your enrolled people. You can also enroll people using saved face images.

2.2.1 Enroll a Person Using a Face Image

Do the following to enroll someone using a saved face image:

1. Click Add. You will see an empty dialog as shown below.
   
2. Click Choose File and select the face image from your local computer, tablet, or phone. The face image should be of good quality with at least 150 pixels from ear to ear.
3. Click on the Edit button in the upper right corner of the newly added person's record.
4. Enter the following information:
5. Enter the person's First name and Last name.
6. Set ID Class if desired.
7. Click Save Changes. The person will be added to the Person Database.

2.2.2 Edit Person Records

To edit the records of people already enrolled in the Person Database, click on the Edit button of their record, as indicated by the arrow below.

• Name, Person Type and ID Class fields are for filtering.
• Add and Delete buttons allow adding new and deleting persons.
• Refresh will reload. The page does not update automatically like the Live page does.
• Click anywhere except the image or click the pencil icon to edit a record.
• Click anywhere on the face image to select a record (multiple records can be selected at once for deletion).

?If faces do not appear, its likely due to setting to persist synced person face images disabled on server.

Person has many attributes. Below are just some. The Desktop software has additional properties that cannot be edited on the SAFR Camera Web Console.

2.3 Operation Settings

2.3.1 Monitoring

Monitoring mode: Specifies how SAFR Camera processes its camera view video feed. See the Operator Modes documentation in the SAFR Software Administration Guide for full descriptions of the available Operator Modes.

Below is a summary of the monitoring modes:

• All Events Monitoring (default) – Report all events including unrecognizable faces. Unrecognizable events are not reported until person is in view for at least 1.5 seconds (configurable)

• Enrolled Monitoring – Only record events for known recognized persons

• Enrolled and Stranger Monitoring – Record events for known recognized persons and strangers. Strangers are defined as a face that was of sufficient quality to attempt to recognize, but no match was found.

• Threat/Concern and Stranger Monitoring – Record events for strangers and known people marked as Threat or Concern.

• Learn and Monitor – Auto-enroll faces as they appear in front of the camera. This creates a new identify for every face as long as the quality thresholds for learning are met. This mode is most useful for various reports that learn faces such as Traversal Reports or Queue Reports.

• Note: If you want to enroll specific faces from camera, use “Enroll” option from the Events view.

Additional analytics: Enables additional analytics; age, gender and sentiment analysis. These analytics are estimated from the face image appearing in the camera view. Their accuracy will depend upon the source image quality. No biometric signature is needed to compute.

Data is reported in user interface as shown below and persisted with event so it is available in various reports or with exported event data.

Sentiment and Smile are reported as follows:

• During event (while face is in view): Display current sentiment and show smile indicator if smiling.
• After event (face left the view): Display average sentiment and % of time smile was detected.

2.3.2 DVR

Settings to enable recording of video to local Micro SD card and optionally to an AWS account. See Digital Video Recording (DVR) Settings section below for information on configuring DVR.

2.3.3 Camera

Camera settings offers advanced features to control the exposure and shutter speed of the camera. See Camera Settings below for information about settings in this page.

2.3.4 Image

Image settings offers features to control the picture. Image settings involve digital processing and should be reserved only when desired results cannot be achieved with Camera Settings. See Image Settings below for information about settings in this page.

2.4 System Settings

2.4.1 SAFR Server

Specifies the SAFR Server, if any, that SAFR Camera is connected & managed by.

⚠️ if SAFR Camera is managed by a SAFR Server, many of SAFR Camera’s settings won't be locally configurable anymore; they'll only be configurable via the SAFR Server.

There are four possible values for the SAFR Server field.

• None: SAFR SCAN isn't connected to a SAFR Server.
• SAFR Server: SAFR Camera is connected to a locally deployed (i.e. an on-premises) SAFR Server.
• SAFR Cloud: SAFR Camera is connected to a SAFR Server in the cloud.
• Custom: SAFR Camera is connected to a locally deployed SAFR Server that has custom SAFR service endpoints. There are two scenarios where you might want to use custom SAFR service endpoints:
  • The SAFR Server has custom domain names for each service endpoint.
  • You want to use an unsecure HTTP connection, rather than the default HTTPS connection. We do not recommend using HTTP connections, but this may be useful in debugging.

2.4.2 Behavior when Connected to SAFR Server

2.4.2.1 Persons

When you connect to SAFR Server, the local faces database on SAFR Camera will be combined with face database on SAFR Server. If SAFR finds that two face are similar, those faces will be automatically merged into a single identify. See Software SAFR Administration Guide for information on merging and unmerging faces.

2.4.2.2 Events

Events will be uploaded from SAFR Camera to SAFR Server. This occurs in real time. Events will typically appear on SAFR Server within milliseconds of being created and will be updated in real time.

If networking between SAFR Camera and SAFR Server is unavailable, events will be cached on SAFR Camera until the connection is restored. When connection is restored, cached events will be uploaded to SAFR Server and cleared from local cache. SAFR Camera is capable of storing the latest 2,000 events with images and up to 28,000 more most recent events without images.

2.4.2.3 Configuration

When connected to SAFR Server, many of the configuration can be modified on the connected SAFR Server. Use SAFR Server Web Console or SAFR Desktop to manage those settings.

2.4.3 Video Streams

Configure and manage the video stream profiles produced by the camera.

2.4.4 Network

Manually configure the device's network settings.

Note: SAFR automatically configures its network settings when the device is first turned on and contacts a DHCP server. However, you can manually configure your device's network settings here if you need to override the DHCP server.

2.4.5 Date and Time

Settings to configure NTP time synchronization server or manually set the device's date and time.

2.4.6 Update

Allows you to update both the SAFR SCAN software and the SAFR SCAN device's firmware to the latest versions. If available, SAFR will get updates directly from SAFR update server. If not, firmware can be downloaded from http://safr.real.com/firmware.

2.4.7 Reset

Allows you to either reboot the SAFR SCAN device or reset its settings to their default values.

There's no need to reboot or reset your SAFR SCAN device during the normal operation of your device, but rebooting or resetting the device can be useful if you encounter an unexpected behavior while operating the device.

2.5 Security Settings

2.5.1 System Login

The login credentials to be used to log in to SAFR Camera. This is where you can change the password.

2.5.2 Video RTSP Access

When video RTSP access is enabled, remote connections are able to view SAFR Camera’s camera feed in realtime. This is disabled by default because remote viewing isn't necessary or appropriate for most access control use cases.

2.5.3 SSL

Allows you to manually configure your SSL (Secure Sockets Layer) settings.

2.5.4802.1 x

Allows you to enable 802.1x authentication.

2.5.5 Firewall

Allows you to manually configure your firewall settings.

2.5.6 VPN

Enables you to connect the camera to a remote VPN service using OpenVPN to facilitate remove management of the camera behind NAT firewalls.

3 Digital Video Recording (DVR) Settings

Settings to enable recording of video to local Micro SD card and optionally to an AWS account.

3.1 Recording to SD Card Only

If Access Protocol is “None” video is recorded on SD Card only.

1. Go to Operation > DVR
2. Set “DVR function” to “Enabled”
3. Set Access Protocol is “None”
4. Complete the dialog as follows:
   
   Where:
   • Recording loop duration determines number of days to record until older videos are discarded. Setting this to larger than the SD card has space for will result in storing only up to the card’s capacity.
   • Clear all pre-existing SD card data results in a one time reformatting of the SD Card.
5. If successful, you should see following state:
   

3.2 Recording to SD Card and AWS account

3.2.1 Setup AWS Account

Kinesis video streaming is required to enable SAFR Camera DVR recording to cloud. See Article Create AWS Account at http://docs.real.com or http://support.real.com for information on setting up AWS Account.

3.2.2 Configure DVR

If Access Protocol is “WebRTC via AWS KVS” video is recorded on SD Card and in AWS.

1. Go to Operation > DVR
2. Set “DVR function” to “Enabled”
3. Set Access Protocol is “WebRTC via AWS KVS”
4. Complete the following dialog.
   
   Where:

• Recording loop duration determines number of days to record until older videos are discarded. Setting this to larger than the SD card has space for will result in storing only up to the card’s capacity.
• Clear all pre-existing SD card data results in a one time reformatting of the SD Card.
• Select the AWS Region closes to you or the region you prefer if proximity is not the only factor.
• Use AWS Account Access Key and Access Secret created above.
• Set AWS Channel Name to a unique value within the AWS Account.
  • This will be the Video Stream name created in AWS Kinesis Account
• Enter remaining values as desired.

5. If successful, you should see following state:
   

3.3 Accessing Recorded Video on SD Card

Recorded videos on SD Card may be access via RTSP URLs provided in the DVR Status display. Recorded video can be played directly from the camera’s SD Card. You can use tools such as ffmpeg to store the video on local disk.

3.3.1 Enable RTSP Playback from SAFR Camera

Before you can play or download recorded video from SAFR Camera, you must enable RTSP Streaming.

1. Go to Security
2. Select “Video RTSP Access” and set to “Enabled”
3. Enter a password (username is always ‘admin’)
4. If unsure, use Digest Authentication.

Playback URLs for live RTSP streams are available in System > Video Streams. Below is information non accessing the recorded video streams.

3.3.2 Viewing Recorded Video

You can view recorded video in VLC or other RTSP-capable Player as described below.

1. Go to Operation > DVR
2. Click on Get rtsp:// URL from time index
3. Enter a time index in the indicated format if desired (not required for viewing live stream)
   
4. Copy the desired RTSP URL
5. Paste that URL into your video player. VLC is a good choice. You can download VLC from https://www.videolan.org/vlc/#download.
6. VLC will prompt for username and password. Username is always ‘admin’. Password is SAFR Camera RTSP Password set as described in Enable RTSP Playback from SAFR Camera section above.

Exporting Recorded Video

To export recorded video, you will need to use a tool such as ffmpeg to download and save the video to local disk. This section document show to use ffmpeg for this purpose.

1. Get RTSP URL as described above.
2. Download a tool that will record an RTSP stream to disk. ffmpeg is a good choice. You can download ffmpeg from https://www.ffmpeg.org/download.html.
   • Download options can be confusing. The easiest option is to use the gyan.dev download site and look for a ZIP file (.zip extension) with “essential” (e.g. ffmpeg-6.1.1-essentials_build.zip)
3. Unpack the downloaded file and look for ffmpeg.exe located in the ‘bin’ directcory.
4. Place ffmpeg.exe in your command path or in the directory where you will record files to.
5. Open a DOS Command prompt and change directory to the location where your ffmpeg.exe file is
6. In DOS, use the following command to record the RTSP Stream in segments from the camera:

ffmpeg -i "RTSP_URL" -c copy -reset_timestamps 1 -metadata title=\"TITLE\" -map 0:0 -f segment -strftime 1 -segment_time 60 -segment_format mp4 "Record__%Y-%m-%d_%H-%M-%S.mp4"

Where

• RTSP_URL should be replaced with the URL obtained in Viewing Recorded Video section above. You need to add the username and password into the URL as follows:
• rtsp://admin:PASSWORD@IP_ADDRESS/PATH
• Example:
• rtsp://admin:letmein@192.168.1.139/video?localtime=2024-01-14-23-15-00
• TITLE can be any string. This will be set as the video file metadata title.
• -segment_time 60 can be changed to the duration you wish for the recorded file. If you do not wish to segment, set the value to a longer duration than needed and press keyboard shortcut control-c to stop recording when desired.
• -segment_format may be changed as follows. If name is not unique, existing file will be overwritten.

7. The resulting file will be saved to the same directory. Press Control-C to stop recording.

3.3.3 Creating Batch File for Saving Recordings

For convenience, the above can be put in a text file and saved as a DOS Batch file (.bat extension) or Linux shell script (.sh extension). You can even put replaceable parameters such as the following for Windows:

ffmpeg -i " rtsp://admin:%1@10.124.13.3/video?localtime=%2-%3" -c copy -reset_timestamps 1 -metadata title=\"%2 %3\" -map 0:0 -f segment -strftime 1 -segment_time 60 -segment_format mp4 "Record__%%Y-%%m-%%d_%%H-%%M-%%S.mp4"

Notice that %1, %2 and %3 were added to allow inseting command line arguments and existing “%” were replaced with “%%” to avoid them from being replaced.

Save the above as ‘record.bat’ and run as follows:

record.bat MYPASSWORD 2024-01-14 23-15-00

Where the arguments are password, date and time respectively.

4 Camera Settings

Screenshot below shows Camera Settings available from Operation > Camera.

4.1 WDR (Wide Dynamic Range)

WDR mode applies 2 different exposures to each frame and blends pixels from each image to obtain the best exposure.

WDE Mode provide a more restricted set of settings.

Backlight compensation settings are not applicable for WDR mode. The Exposure adjust speed is also not available with WDR. Finally, gain control is not available with WDR mode.

4.2 WDR Shutter Speed 1 and 2

Shutter speed can be controlled for both camera exposures. This setting sets a fixed shutter speed for each frame. If there is variation in lighting from frame to frame or at different times of day, setting a fixed value may result in suboptimal performance when lighting has changed from the levels for which it was optimized.

4.3 Shutter Speed Min and Max

Shutter Speed Min and Max settings apply limits for both of the exposure modes. This allows Auto behavior of each exposure to be applied but sets a limit on how low and how high shutter speed will vary. Behavior is similar to that described for non-WDR mode.

4.4 Night Vision Settings

You can enable the IR lights or leave them in Auto. When in Auto, IR will be enabled in low light. On will force IR lights to on always.

4.5 EXPOSURE MODE

Options of Face Prioritized and Scene Prioritized.

Face Prioritized

Camera exposure is optimized for any faces in view of the camera. When set to Face Prioritized, camera will adjust exposure to optimize for the largest face in view of the camera. This results in a dynamic exposure. This is the prefered mode for facial recognition. You will read below, how Scene Exposure Region can be used to give the camera a hint on what to set the exposure to when faces are not present

Scene Prioritized

Camera exposure is optimized for entire scene. When in this mode, camera does not adjust exposure based on presence of faces in the video. Exposure is optimized for the entire scene. While this mode is useful for creating an optimal video recording of entire scene, it will not optimize exposure to the face and thus have negative impact on face recognition.

Example with strong backlight

Scene Prioritized	Face Prioritized

4.6 SCENE EXPOSURE REGION

Determines the camera exposure when no face is present in the view of the camera. This setting allows you to give a hint to the camera to set exposure when no faces are in view. This is done by selecting a region of the video to set exposure. Possible values are:

• Top 25%
• Top 35%
• Top 50%
• Center
• Bottom 50%
• Bottom 35%
• Bottom 25%
• Custom

Setting this value to the exposure of where faces will appear in the video will reduce the amount of time I takes for the camera to adjust exposure for the face.

Examples

Incorrect Settings - Exposure zone set to region of strong backlight

Face not in view	Face in view

The problem is that the camera has to adjust from a very dark exposure to a very bright exposure to let sufficient light in to properly light the face. This takes a few hundred milliseconds. For fast moving objects, you may lose precious moments to perform recognition.

Correct Settings - Exposure zone set to region of similar lighting

Face not in view	Face in view

?Use Scene Exposure mode with face in view to chose which region offers the best exposure on the face. Adjust the region, including custom if needed, until the face is well lit. Then change exposure mode back to Face Prioritized.

Default value for Scene Exposure Region is Bottom 35%. This is most often the optimal location but may not work if floor is brightly lit from outdoor reflection.

4.7 Exposure Level

Allows the exposure to be boosted or attenuated from what the camera considers optimal. This value can be as high as 400% (4x brighter) or as low as 25% (4x darker). It is recommended to leave this at 100%, but the setting may be modified if SAFR is consistently incorrectly setting exposure.

4.8 Exposure Adjust Speed

Controls speed at which the camera adjusts exposure. Default is 72%. Values greater than 72% reduces the time it takes for the camera to adjust exposure. Decreasing this value increases the time it takes to adjust exposure.

Increasing this value too far may result in "yo yo effects". That is, due to the speed of adjustment, the camera may overshoot and oscillate the exposure back and forth.

Adjusting this value may be useful in case where faces are moving very fast through the view. If the person is moving too fast, the camera may end up adjusting exposure for a face that is no longer present in that location. Speeding up the Exposure Adjust Speed may help in this case or slowing it down may avoid adjusting exposure too quickly and keeps exposure

4.9 Backlight Compensation

SAFR Camera is designed to compensate for backlight conditions. Backlight Compensation setting provide some control over that feature. Possible values are:

• Auto
• Slight backlight ( outdoors)
• Medium backlight ( outdoors )
• Medium strong backlight (typical outdoors)
• Strong backlight (indoors / outdoors )
• Very strong backlight (typical indoors)
• Extreme backlight ( indoors)

Auto mode attempt to adjust exposure under typical conditions. Other options allow you to provide a hint to the camera as to the type of conditions it is facing which will improve its ability to handle those conditions.

Backlight Compensation primarily manages the shutter speed. Shutter speed is also controlled by the Minimum and Maximum Shutter Speed settings. SAFR Camera handles these settings by applying the setting that imposes the greatest constraint. For example, if Maximum Shutter Speed is set below what the Backlight Compensation setting would set for the current conditions, then the Maximum Shutter Speed will be applied.

4.10 Maximum Shutter Speed

Max Shutter speed is one of the tools to use to handle backlighting. Faster Shutter speeds will decrease light allowed into the camera and thus darken the view. Setting Max Shutter Sped prevents the camera from darkening the face image too much due to bright background.

4.11 Minimum Shutter Speed

Lower shutter speed results in motion blur. Setting a higher minimum shutter speed can reduce or prevent motion blur.

Normally you want the minimum shutter speed to be as low as possible to increase exposure (light on the face).

At 1/120 you will typically not get motion blur. But at this speed, you may have insufficient light.

4.12 Max Gain

When the lighting gets reduced, the camera applies more gain to the image. Higher gain adds noise to the image. This results in faces looking grainy due to the noise in the signal resulting from the gain. To control noise, you can limit the Max gain setting. But by doing this, you reduce the brightness. Gain should only be adjust once improvements from adjustment of shutter speed and iris have been exhausted.

4.13 About Zoom, Focus and Iris Controls

Zoom, focus and iris controls are located below the video preview region on the Camera settings page. All three settings have similar controls as shown below.

Controls on the left are for adjusting the setting in one direction while controls on the right adjust in the other. Any changes made are applied immediately and visible in the preview window.

Each of the buttons results in a varying degree of change as noted in the image above.

4.14 Iris

Another control that is related is Iris settings. The SAFR Camera allows control over the Iris. Opening the Iris all the way results in light diffusing and reducing focus. So increasing Iris will limit the ability to focus the camera, especially at close range. The diffusion is not as pronounced further from the camera. But increasing iris also reduces the depth of focus or the range in which a subject is in focus. Reduced range of focus means you have less time that the subject is in focus.

Reducing the Iris (making it smaller) results in a larger depth of field and you can maintain focus on the subject for a larger distance.

4.15 Zoom

Use the zoom controls to adjust the width of view in the region of interest which is usually a choke point where faces are constrained to a narrow and face tend to be looking in a single direction. Zoom should be applied so that faces are as large as possible. With a 4k camera, the width of view at the target area should be no more than 20 feet wide. Wider views will result in resolution on the face being too low for high quality recognition. If needed, use 2 or more cameras to cover wider areas (1 camera for every 20 feet).

Controls on the right side (>, >> and Tele) will increase pixels on the face. Controls on the left side (Wide, <<, and <) will decrease pixels on the face.

4.16 Focus

Auto-focus is not supported at this time. Focus should be performed after zoom adjustments.

To perform manual focus, perform the following steps

1. Start from the largest change buttons (“Near” and “Far”)
2. Press Near or Far until camera goes into focus and then out
3. Identify the end points where camera is out of focus based on least number of clicks.
4. Go to one extreme and go back one half that amount.
5. Perform steps 1-4 with the << and >> buttons
6. Finally, repeat steps 1-4 with the < and > buttons

5 Image Settings

Screenshot below shows Image Settings available from Operation > Camera.

Except for Orientation, Image Settings perform digital processing on the video frames. The Camera Settings on the other hand are applying adjustments to the physical hardware to transform the image. Image Settings should only be applied after exhausting all improvements that can be obtained through Camera Settings.

5.1 Orientation

Orientation controls the rotation of the image. Currently only landscape modes are supported by the SAFR Camera. Future firmware release will introduce corridor (portrait) mode orientation. When needed, gimbal should be rotated to facilitate a landscape image.

Following settings are available:

Setting	Effect	Example
Normal	Original image (no transformations)
Flip	180° rotation and flip on vertical
Mirror	Flip on Vertical Axis
180	180° rotation

5.2 Brightness, Contrast, Color Saturation, Sharpness

These settings apply the respective adjustments on each frame of the video. All adjustments in the Image page apply digital processing. These settings should always be use donly after making all possible adjustments on the Camera page.

5.3 Denoise

Denoise may be useful to reduce the effects of noise introduced by Gain. But denoise is digital processing and may introduce artifacts. For example, with high motion you may get ghosting effects as a result of Denoise filter. Denoise should be a last resort after adjusting settings in the Camera tab.

6 SAFR SERVER SYSTEM SETTINGS

This section describes settings that are specific to SAFR Camera when connected to SAFR Server. These settings are configured on SAFR Server only when connected to SAFR Camera. System Settings can be set from either SAFR Desktop or SAFR Server Web Console. This section describes those settings for SAFR Server Web Console. Settings in SAFR Desktop are similar but accessed from a different location.

6.1 ACCESSING SAFR WEB CONSOLE

To access SAFR Web Console, open a browser and connect to http://IPADDRESS:8090 if using SAFR On Premises server or http://safr.real.com/console if using SAFR Cloud.

6.2 Open SAFR Camera System Settings

1. Connect SAFR Camera to SAFR Server or SAFR Cloud as described above.
2. Open SAFR Web Console as described above.
3. Go to Video Feeds and locate the newly added SAFR Camera
   1. You can use Camera Name to identify your camera as shown below:
      
4. Click “Config” for the SAFR Camera
5. Go to System page. You will see the following:
   

The following table describes settings on the System page.

Setting	Description
Identity Sync – Sync Direction	Controls how person identities are synchronized between SAFR Camera and SAFR Server. Options are: Bidirectional (default) – Faces added or updated on Camera are pushed to SAFR Server and vice versa. Only from server down to processor – Only faces added or updated on Server are pushed to camera. Any changes made on camera are not pushed back to server.
Identify Sync – Persist face images	Determines if faces images are sent to camera. Enabled - Faces images on server are copied to SAFR Camera long with biometric signature and other data associated with person. This is useful if the SAFR Camera Web Console will be used frequently. It may be most useful during initial setup and troubleshooting. Disabled (default) - Only the biometric signature and person metadata is sent to the camera during sync. Face images are not transferred to the camera. This option is provided to enhance privacy by restricting transmission of data that is not required..
Limit sync to identifies with access clearance for the processor	This option limits which faces are loaded on SAFR Camera based on the Access Clearance assigned to a person. This option is most applicable to SAFR Access Control. SAFR Camera can be assigned one or more zones and each access clearance is assigned one or more zones. Person records are assigned one and only one Access Clearance. If the Access Clearance a person belongs to contains the zone in which SAFR Camera belongs, the face is loaded onto that camera.
Limit sync to identifies with zones for the processor	This option limits which faces are loaded on SAFR Camera based on the Zones assigned to a person. This option is most applicable to SAFR Camera. SAFR Camera can be assigned one or more zones and each person record can be assigned one or more zones. If the Zone a person belongs to contains a Zone in which SAFR Camera belongs, the face is loaded onto that camera.
Persist person records on disk	This option is provided to enhance privacy and security. When enabled, person records are loaded at the time the camera boots. Person records are only stored in volatile memory. If the camera loses power for any reason, person records are lost. This prevents person data from being compromised in the event the camera is stolen. Note: SAFR data is encrypted on disk. So even if this option is not enabled for multiple reasons (device contains a secure boot partition and the disk is not available and the data is encrypted with AES 256 cipher encryption keys.
Allow direct access through Processor Console	This option is provided to enhance security. When enabled, access to the SAFR Camera Web Console is not available and device can only be managed from SAFR Server. Note: A factory reset can be used to regain access to the device in the event the device cannot be connected to SAFR Server. A factory reset results in all personal data being purged from the device.

7 Troubleshooting

Symptom	Resolution
Image is completely dark with noise.	Ensure the iris was not closed completely.

1