Skip to content

Release notes - CVEDIA-RT - 2024.1.2

NX Witness Plugin for VCA


  • The face detection is now based on the new extended PVA model that include also faces. This change allows saving considerable compute resources by always running a single detector.
  • Released new improved models for vehile models, vehicle color, and cloth colors.
  • Added the ability to record local video snippets of the triggered events. To enable it, edit the base_config.json of SecuRT and configure the settings under the EventVideoRecorder object.


  • Reduced memory consumption in particular when using a high input resolution
  • Added support for recent patches on JetPack 4.6.x