AxxonSoft has launched version 4.5 of the Axxon Next intelligent VMS. This version introduces new GPU acceleration capabilities and AI video analytics, self-diagnostics service, support for licensing server, enhanced web client, and new server and GUI features.

GPU Acceleration

Server Side: Axxon Next now supports hardware-accelerated video decoding on integrated Intel GPUs (Intel? Quick Sync Video technology) and NVIDIA graphics cards. The decoded video can be analyzed with any standard or AI video analytics. For AI acceleration, Axxon Next utilizes NVIDIA and Intel? GPUs and dedicated AI accelerators.

Several cards can be installed on a single server for video decoding and/or AI acceleration. Thanks to this, one server can handle multiple times more cameras when using video analytics, which reduces hardware expenses.

Client Side: Along with Intel? Quick Sync Video, you can now use NVIDIA graphics cards for hardware video decoding when the video is displayed on the client?s screen, with the possibility to install several NVIDIA cards on a single client computer.

This provides smooth playback of dozens of Full HD video feeds, including H.265, without burdening the CPU. So, you can display more cameras on the client?s screen or use client computers with lower CPU performance specs. This feature enables building huge video walls using just a few client computers of average CPU performance, which makes the video surveillance system more cost-effective.

New Neural Network for Human Detection: A new ?general? neural network for human detection has been developed, which accurately detects people in most standard situations from a regular viewing angle. For example, video surveillance in an office or a store, the camera is mounted under the ceiling, the optical axis of the lens is tilted from vertical by about 60