We have facial recognition platform and we are planning to integrate with Nx witness server by analytics plugin. When exploring how could this be done, we came across multiple possible approaches:
- Sending raw frames from camera plugin into our remote analytics platform.
- Sending uncompressed frames from camera plugin into our remote analytics platform (we would need to implement custom decoding ?)
- Process same camera stream by our platform and synchronize timestamps when sending metadata to nx.
We are leaning towards approach 1, but we are afraid of network bandwidth issue, because each frame can have ~7MB. Do you guys have some preferable approach or can you tell how other integrators are doing it ?
Thank you for your answer.
Please sign in to leave a comment.