Note, this application can process multiple streams of Full High-Definition video in real-time. It makes use of all the available CPU, GPU and RAM resource on your machine. In general, the higher specification your hardware the more video data you will be able to process and the smoother the application will run. Note, the PC running the face-recognition software should be dedicated to this task and no other major application should be run on the same PC.
We recommend the following specification of system for a 1 to 2 High-Definition camera streams.
Processor: Intel® i7 or Xeon (Quad Core) or better (CPU benchmark > 8,000)
Windows 10 64 bit
Memory: 16 GB recommended
Video Card: NVIDIA® with Kepler, Maxwell or Pascal architecture with 4GB VRAM and 640 cores (e.g. GTX 1050) with latest NVIDIA drivers installed
An SSD disk with at least 100 Gigabyte free-space for storage of results
For a system that is required to analyse 2 to 4 High-Definition camera streams we recommend the following hardware.
Processor: Modern Intel i7 or Xeon with a CPU benchmark of around 16,000 or better that supports AVX2 instructions
For example the Intel i7-8700K which has a CPU benchmark of 16,034
Or the Intel Xeon Silver 4116 with a CPU benchmark of 15,656
Windows 10 64 bit
Memory: 16 GB DDR4 2400Mhz
Video Card: NVIDIA with Pascal architecture with minimum 8GB VRam and 2,560 cores
SSD with at least 200GB free-space for storage
However, please note that in Identifier you can select the region-of-interest in the frame you wish to analyse. This can significantly reduce the processing burden.
Instructions on how to verify your hardware meets these requirements can be found here.