ONVIF Profile M is supported in XProtect.
What exactly does that mean? Is there specific support for face recognition?
I am not seeing XProtect doing anything specific with ONVIF analytics in general or Profile M specifically.
Are there specific version or updates/patches needed?
Hey Jim,
I wanted to share a helpful resource that might provide more insight into your questions.
Attached is the latest ONVIF Driver documentation, which details how Profile M can be leveraged in XProtect, including a list of the specific features supported by the ONVIF Driver for this profile.
In short, Profile M support in XProtect primarily involves handling metadata and other information sent from the camera. With any well-structured XML file, you can visualize bounding boxes and other event-related information.
Through Profile M, the ONVIF Driver enables XProtect to receive, interpret, and display metadata from compatible devices. This includes analytics such as object detection, classification, and counting, which can be used to trigger alerts and events.
Although Profile M covers a wide range of analytics and metadata, it doesn’t explicitly provide face recognition capabilities. However, if your camera or analytic device supports face recognition metadata, XProtect can receive and display this data without directly performing the recognition itself.
Additionally, Profile M metadata can be utilized to create rules and alarms within XProtect, but this functionality will depend on the analytic capabilities of the connected devices.
For further details, please refer to the attached ONVIF Driver documentation and the file in Chebel’s comment.
Kind regards,
Genadi
Section 8.13 of the Profile M spec mentions Face recognition, which is detailed in Annex G in the ONVIF analytics service spec, sections G.1 and G.2 This is what I am referring to.
Specifically, in G.2:
G.2 Face Recognition
This section describes Face Recognition rule engine that produces an event with the topic “tns1:RuleEngine/
Recognition/Face. The Face Recognition rule engine triggers the event when the identified face matches with
a reference face image, which is described using the Rule Description Language as:
<tt:RuleDescription Name="tt:FaceRecognition">
<tt:Parameters>
<tt:SimpleItemDescription Name="IncludeImage" Type=”xs:string"/>
</tt:Parameters>
<tt:Messages>
<tt:Source>
<tt:SimpleItemDescription Name=”VideoSource” Type=”tt:ReferenceToken”/>
<tt:SimpleItemDescription Name=”Rule” Type=”xs:string”/>
</tt:Source>
<tt:Data>
<tt:SimpleItemDescription Name="Likelihood" Type="xs:float"/>
<tt:SimpleItemDescription Name="Label" Type="xs:string "/>
<tt:SimpleItemDescription Name="ImageUri" Type="xs:anyURI"/>
<tt:SimpleItemDescription Name="EnrollmentID" Type="xs:string"/>
<tt:SimpleItemDescription Name="RefImageUri" Type="xs:anyURI"/>
<tt:ElementItemDescription Name="Image" Type="xs:base64Binary"/>
<tt:ElementItemDescription Name="BoundingBox" Type="tt:Rectangle"/>
</tt:Data>
<tt:ParentTopic>
tns1:RuleEngine/Recognition/Face
</tt:ParentTopic>
</tt:Messages>
</tt:RuleDescription>
Ho Jim,
XProtect’s support for ONVIF Profile M enables it to receive and display metadata from Profile M-compliant devices, including information related to analytics like face recognition.
Profile M defines a standardized event for face recognition, allowing XProtect to display and use this data. However, XProtect itself does not perform face recognition; it relies on the connected devices to provide this metadata.
For face recognition to work, you need an ONVIF Profile M-compliant device that supports face recognition and sends this metadata. XProtect can trigger rules or alerts based on this metadata, but the ability to recognize faces depends entirely on the camera or analytic device.
Events that we have and are related are Face Appearing/ Disappearing as described in the ONVIF Driver docs attached above.
If this information does not fully address your question, please feel free to clarify or specify the areas you want to be answered and we’ll be happy to provide more detailed information.
Kind regards,
Genadi
We have an ONVIF Profile M compliant device which does face recognition and sends metadata to XProtect.
The referenced document deals with face detection, not recognition. Does XProtect surface face recognition to the GUI, for example if our device sends EnrollmentID, as per the spec, is that a searchable field for the user?