Object Recognition: AI-based additional module in Xeoma
Object Recognition in Xeoma:
AI-based Additional module
Xeoma’s Object Recognizer is a powerful and affordable tool for automated detection and recognition of various objects in a cameras’ live stream and reacting to the ones specifically selected.
Using knowledge and experience gathered in data sets, the artificial intelligence in Object Recognizer is scanning the camera image for articles that can belong to a certain type, based on specific characteristics typical for that type.
It can be used in various situations for detection or – in reverse – for ignoring one or several types of objects. For example:
• Airports: detection of airplanes taking off or preparing for landing; avoid false alarms by birds.
• Home use: react to intruders but exclude home pets or birds from triggering alarms.
• Shipping/freighting industry: detect trucks that are about to enter docks for loading or unloading; dismiss reaction to light-motor cars or buses.
• Production industry: in fish-farming for detection of birds of prey that feed on fish and thus damage the fish (supported integration with a sound cannon acoustic device).
• Business optimization: in parking businesses, detection of abandoned cars for timely evacuation of them with a tow truck.
• Entertainment industry: having automated live broadcasts in which camera always follows a sports ball.
• Naval industry: detection of boats, yachts and ships about to dock; avoid false alarms by birds.
• Military: detection of any vehicles approaching a base.
• Zookeeping: detection of a human in dangerous vicinity of wild animals (in cages, aviary, etc.)
• Police: counter terrorism measure for automated detection of unattended parked vehicles;
• Security Access Control in enterprises: detection of people or vehicles in unauthorized areas.
• Municipal: full insight of trains time of arrival and departure from stations, travelling speed.
• In various other cases, where you need the system to know what kind of object is in the camera image and react to one or several types, or to all except for one or several types.
WHY XEOMA’S OBJECT RECOGNITION IS BETTER
Saves bandwidth:
All done on your machine, no need to connect to any cloud servers.
Realtime video:
Works with real-time video streams. The more fps, the better!
Cost-effective:
One-time payment. You don’t need to pay per image.
Watch video about Object Recognizer
Object Recognizer in pictures:
Video: AI-based birds recognition in Xeoma for airports
Video: Drones detection in Xeoma. Artificial intelligence
NO SPECIAL EQUIPMENT REQUIRED
No special equipment or a powerful standalone server is required. Recognition load is on average 40% more than that of a regular camera in the program. Use our calculator here.
No special cameras with embedded recognition detectors required. Recognition processes are conducted on Xeoma’s side. High quality video stream is enough. Point the camera at an anggle that will allow computer vision clearly see objects, and make sure that the object is in the camera’s field of view for at least a few seconds. More about configuration of objects recognition
Affordable price. Depending on the required conditions and the desired result, you can choose Xeoma Pro or Xeoma Standard licenses and purchase a license for the Additional module “Object Recognizer”. Check prices for Xeoma licenses and Additional modules here
HOW IT WORKS
Using the knowledge taken from the resources that the module downloads before work, it is scanning the camera image for selected object types based on their known idiosyncrasy traits.
Found object get a green frame around them and are clearly visible to VMS operators. You can select either of the two modes of work: either to react to selected object types (the “process selected objects” option) or to ignore the selected object types while reacting to all other object types (the “not process selected objects” options). Upon “seeing” the designated object type in the camera view the module can trigger recording of video episode featuring the object, set off sound alarm, send an email alert, or run programmed reactions with the “Application Runner” module.
Tip: use the ‘Use CUDA recognition’ option to shift some part of CPU load that comes from operation of this module onto the video graphics card side. For this feature, NVidia graphics cards are supported, starting from Pascal series and up.
Xeoma’s other modules might be useful for some of the Object Recognition tasks: use the “Face Recognition” module to know which person exactly is detected, “FaceID” module for double authentication (an access card with a QR Code or chip), “Detector of Abandoned Objects” to detect any objects that are left unattended (without recognition of their type).
Please note that up to version 21.10.13 this module is available only in Xeoma for 64-bit Windows, Linux or Mac OSX operating systems. It is not available for Linux/ARM.
Starting from version 21.10.13 the module is able to work in 64-bit Linux/ARM systems as well.
HOW TO TEST
1. Launch Xeoma. Make sure your Xeoma is in a trial edition or activate a PRO edition license.
2. Add a camera or let Xeoma add a default one for you.
3. Add an Object Recognizer module to the chain.
4. Set the module up by selecting objects that it needs to detect or ignore.
See detailed setup guide here
5. Connect reactions or other filter modules for Object Recognizer to work with.
FREE TRIAL OF XEOMA
Try Xeoma for free! Fill in the fields below and you will get an email with a demo license for the Xeoma Pro edition and all Additional modules, including the AI-powered ‘Object Recognizer’.
To do that, enter your name and your email to send the license to in the fields below, and click the ‘Get Xeoma free demo licenses to email’ button.
We urge you to refrain from using emails that contain personal data, and from sending us personal data in any other way. If you still do, by submitting this form, you confirm your consent to processing of your personal data
Have any questions? Need help? Please contact us! We will be happy to help!
Do you need something else? We can develop it and add it into Xeoma as the paid development. See details
